1c33e97efSMauro Carvalho Chehab.. SPDX-License-Identifier: GPL-2.0 2c33e97efSMauro Carvalho Chehab 3c33e97efSMauro Carvalho Chehab==================== 4c33e97efSMauro Carvalho ChehabThe /proc Filesystem 5c33e97efSMauro Carvalho Chehab==================== 6c33e97efSMauro Carvalho Chehab 7c33e97efSMauro Carvalho Chehab===================== ======================================= ================ 8c33e97efSMauro Carvalho Chehab/proc/sys Terrehon Bowden <terrehon@pacbell.net>, October 7 1999 9c33e97efSMauro Carvalho Chehab Bodo Bauer <bb@ricochet.net> 10c33e97efSMauro Carvalho Chehab2.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000 11c33e97efSMauro Carvalho Chehabmove /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009 12c33e97efSMauro Carvalho Chehabfixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009 13c33e97efSMauro Carvalho Chehab===================== ======================================= ================ 14c33e97efSMauro Carvalho Chehab 15c33e97efSMauro Carvalho Chehab 16c33e97efSMauro Carvalho Chehab 17c33e97efSMauro Carvalho Chehab.. Table of Contents 18c33e97efSMauro Carvalho Chehab 19c33e97efSMauro Carvalho Chehab 0 Preface 20c33e97efSMauro Carvalho Chehab 0.1 Introduction/Credits 21c33e97efSMauro Carvalho Chehab 0.2 Legal Stuff 22c33e97efSMauro Carvalho Chehab 23c33e97efSMauro Carvalho Chehab 1 Collecting System Information 24c33e97efSMauro Carvalho Chehab 1.1 Process-Specific Subdirectories 25c33e97efSMauro Carvalho Chehab 1.2 Kernel data 26c33e97efSMauro Carvalho Chehab 1.3 IDE devices in /proc/ide 27c33e97efSMauro Carvalho Chehab 1.4 Networking info in /proc/net 28c33e97efSMauro Carvalho Chehab 1.5 SCSI info 29c33e97efSMauro Carvalho Chehab 1.6 Parallel port info in /proc/parport 30c33e97efSMauro Carvalho Chehab 1.7 TTY info in /proc/tty 31c33e97efSMauro Carvalho Chehab 1.8 Miscellaneous kernel statistics in /proc/stat 32c33e97efSMauro Carvalho Chehab 1.9 Ext4 file system parameters 33c33e97efSMauro Carvalho Chehab 34c33e97efSMauro Carvalho Chehab 2 Modifying System Parameters 35c33e97efSMauro Carvalho Chehab 36c33e97efSMauro Carvalho Chehab 3 Per-Process Parameters 37c33e97efSMauro Carvalho Chehab 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj - Adjust the oom-killer 38c33e97efSMauro Carvalho Chehab score 39c33e97efSMauro Carvalho Chehab 3.2 /proc/<pid>/oom_score - Display current oom-killer score 40c33e97efSMauro Carvalho Chehab 3.3 /proc/<pid>/io - Display the IO accounting fields 41c33e97efSMauro Carvalho Chehab 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 42c33e97efSMauro Carvalho Chehab 3.5 /proc/<pid>/mountinfo - Information about mounts 43c33e97efSMauro Carvalho Chehab 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 44c33e97efSMauro Carvalho Chehab 3.7 /proc/<pid>/task/<tid>/children - Information about task children 45c33e97efSMauro Carvalho Chehab 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 46c33e97efSMauro Carvalho Chehab 3.9 /proc/<pid>/map_files - Information about memory mapped files 47c33e97efSMauro Carvalho Chehab 3.10 /proc/<pid>/timerslack_ns - Task timerslack value 48c33e97efSMauro Carvalho Chehab 3.11 /proc/<pid>/patch_state - Livepatch patch operation state 49c33e97efSMauro Carvalho Chehab 3.12 /proc/<pid>/arch_status - Task architecture specific information 50f1f1f256SIvan Babrou 3.13 /proc/<pid>/fd - List of symlinks to open files 51c33e97efSMauro Carvalho Chehab 52c33e97efSMauro Carvalho Chehab 4 Configuring procfs 53c33e97efSMauro Carvalho Chehab 4.1 Mount options 54c33e97efSMauro Carvalho Chehab 5537e7647aSAlexey Gladkov 5 Filesystem behavior 5637e7647aSAlexey Gladkov 57c33e97efSMauro Carvalho ChehabPreface 58c33e97efSMauro Carvalho Chehab======= 59c33e97efSMauro Carvalho Chehab 60c33e97efSMauro Carvalho Chehab0.1 Introduction/Credits 61c33e97efSMauro Carvalho Chehab------------------------ 62c33e97efSMauro Carvalho Chehab 63c33e97efSMauro Carvalho ChehabThis documentation is part of a soon (or so we hope) to be released book on 64c33e97efSMauro Carvalho Chehabthe SuSE Linux distribution. As there is no complete documentation for the 65c33e97efSMauro Carvalho Chehab/proc file system and we've used many freely available sources to write these 66c33e97efSMauro Carvalho Chehabchapters, it seems only fair to give the work back to the Linux community. 67c33e97efSMauro Carvalho ChehabThis work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm 68c33e97efSMauro Carvalho Chehabafraid it's still far from complete, but we hope it will be useful. As far as 69c33e97efSMauro Carvalho Chehabwe know, it is the first 'all-in-one' document about the /proc file system. It 70c33e97efSMauro Carvalho Chehabis focused on the Intel x86 hardware, so if you are looking for PPC, ARM, 71c33e97efSMauro Carvalho ChehabSPARC, AXP, etc., features, you probably won't find what you are looking for. 72c33e97efSMauro Carvalho ChehabIt also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But 73c33e97efSMauro Carvalho Chehabadditions and patches are welcome and will be added to this document if you 74c33e97efSMauro Carvalho Chehabmail them to Bodo. 75c33e97efSMauro Carvalho Chehab 76c33e97efSMauro Carvalho ChehabWe'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of 77c33e97efSMauro Carvalho Chehabother people for help compiling this documentation. We'd also like to extend a 78c33e97efSMauro Carvalho Chehabspecial thank you to Andi Kleen for documentation, which we relied on heavily 79c33e97efSMauro Carvalho Chehabto create this document, as well as the additional information he provided. 80c33e97efSMauro Carvalho ChehabThanks to everybody else who contributed source or docs to the Linux kernel 81c33e97efSMauro Carvalho Chehaband helped create a great piece of software... :) 82c33e97efSMauro Carvalho Chehab 83c33e97efSMauro Carvalho ChehabIf you have any comments, corrections or additions, please don't hesitate to 84c33e97efSMauro Carvalho Chehabcontact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this 85c33e97efSMauro Carvalho Chehabdocument. 86c33e97efSMauro Carvalho Chehab 87c33e97efSMauro Carvalho ChehabThe latest version of this document is available online at 88c33e97efSMauro Carvalho Chehabhttp://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html 89c33e97efSMauro Carvalho Chehab 90c33e97efSMauro Carvalho ChehabIf the above direction does not works for you, you could try the kernel 91c33e97efSMauro Carvalho Chehabmailing list at linux-kernel@vger.kernel.org and/or try to reach me at 92c33e97efSMauro Carvalho Chehabcomandante@zaralinux.com. 93c33e97efSMauro Carvalho Chehab 94c33e97efSMauro Carvalho Chehab0.2 Legal Stuff 95c33e97efSMauro Carvalho Chehab--------------- 96c33e97efSMauro Carvalho Chehab 97c33e97efSMauro Carvalho ChehabWe don't guarantee the correctness of this document, and if you come to us 98c33e97efSMauro Carvalho Chehabcomplaining about how you screwed up your system because of incorrect 99c33e97efSMauro Carvalho Chehabdocumentation, we won't feel responsible... 100c33e97efSMauro Carvalho Chehab 101c33e97efSMauro Carvalho ChehabChapter 1: Collecting System Information 102c33e97efSMauro Carvalho Chehab======================================== 103c33e97efSMauro Carvalho Chehab 104c33e97efSMauro Carvalho ChehabIn This Chapter 105c33e97efSMauro Carvalho Chehab--------------- 106c33e97efSMauro Carvalho Chehab* Investigating the properties of the pseudo file system /proc and its 107c33e97efSMauro Carvalho Chehab ability to provide information on the running Linux system 108c33e97efSMauro Carvalho Chehab* Examining /proc's structure 109c33e97efSMauro Carvalho Chehab* Uncovering various information about the kernel and the processes running 110c33e97efSMauro Carvalho Chehab on the system 111c33e97efSMauro Carvalho Chehab 112c33e97efSMauro Carvalho Chehab------------------------------------------------------------------------------ 113c33e97efSMauro Carvalho Chehab 114c33e97efSMauro Carvalho ChehabThe proc file system acts as an interface to internal data structures in the 115c33e97efSMauro Carvalho Chehabkernel. It can be used to obtain information about the system and to change 116c33e97efSMauro Carvalho Chehabcertain kernel parameters at runtime (sysctl). 117c33e97efSMauro Carvalho Chehab 118c33e97efSMauro Carvalho ChehabFirst, we'll take a look at the read-only parts of /proc. In Chapter 2, we 119c33e97efSMauro Carvalho Chehabshow you how you can use /proc/sys to change settings. 120c33e97efSMauro Carvalho Chehab 121c33e97efSMauro Carvalho Chehab1.1 Process-Specific Subdirectories 122c33e97efSMauro Carvalho Chehab----------------------------------- 123c33e97efSMauro Carvalho Chehab 124c33e97efSMauro Carvalho ChehabThe directory /proc contains (among other things) one subdirectory for each 125c33e97efSMauro Carvalho Chehabprocess running on the system, which is named after the process ID (PID). 126c33e97efSMauro Carvalho Chehab 127059db434SRandy DunlapThe link 'self' points to the process reading the file system. Each process 128c33e97efSMauro Carvalho Chehabsubdirectory has the entries listed in Table 1-1. 129c33e97efSMauro Carvalho Chehab 130059db434SRandy DunlapNote that an open file descriptor to /proc/<pid> or to any of its 131c33e97efSMauro Carvalho Chehabcontained files or subdirectories does not prevent <pid> being reused 132c33e97efSMauro Carvalho Chehabfor some other process in the event that <pid> exits. Operations on 133c33e97efSMauro Carvalho Chehabopen /proc/<pid> file descriptors corresponding to dead processes 134c33e97efSMauro Carvalho Chehabnever act on any new process that the kernel may, through chance, have 135c33e97efSMauro Carvalho Chehabalso assigned the process ID <pid>. Instead, operations on these FDs 136c33e97efSMauro Carvalho Chehabusually fail with ESRCH. 137c33e97efSMauro Carvalho Chehab 138c33e97efSMauro Carvalho Chehab.. table:: Table 1-1: Process specific entries in /proc 139c33e97efSMauro Carvalho Chehab 140c33e97efSMauro Carvalho Chehab ============= =============================================================== 141c33e97efSMauro Carvalho Chehab File Content 142c33e97efSMauro Carvalho Chehab ============= =============================================================== 143c33e97efSMauro Carvalho Chehab clear_refs Clears page referenced bits shown in smaps output 144c33e97efSMauro Carvalho Chehab cmdline Command line arguments 145c33e97efSMauro Carvalho Chehab cpu Current and last cpu in which it was executed (2.4)(smp) 146c33e97efSMauro Carvalho Chehab cwd Link to the current working directory 147c33e97efSMauro Carvalho Chehab environ Values of environment variables 148c33e97efSMauro Carvalho Chehab exe Link to the executable of this process 149c33e97efSMauro Carvalho Chehab fd Directory, which contains all file descriptors 150c33e97efSMauro Carvalho Chehab maps Memory maps to executables and library files (2.4) 151c33e97efSMauro Carvalho Chehab mem Memory held by this process 152c33e97efSMauro Carvalho Chehab root Link to the root directory of this process 153c33e97efSMauro Carvalho Chehab stat Process status 154c33e97efSMauro Carvalho Chehab statm Process memory status information 155c33e97efSMauro Carvalho Chehab status Process status in human readable form 156c33e97efSMauro Carvalho Chehab wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function 157c33e97efSMauro Carvalho Chehab symbol the task is blocked in - or "0" if not blocked. 158c33e97efSMauro Carvalho Chehab pagemap Page table 159c33e97efSMauro Carvalho Chehab stack Report full stack trace, enable via CONFIG_STACKTRACE 160c33e97efSMauro Carvalho Chehab smaps An extension based on maps, showing the memory consumption of 161c33e97efSMauro Carvalho Chehab each mapping and flags associated with it 162c33e97efSMauro Carvalho Chehab smaps_rollup Accumulated smaps stats for all mappings of the process. This 163c33e97efSMauro Carvalho Chehab can be derived from smaps, but is faster and more convenient 164c33e97efSMauro Carvalho Chehab numa_maps An extension based on maps, showing the memory locality and 165c33e97efSMauro Carvalho Chehab binding policy as well as mem usage (in pages) of each mapping. 166c33e97efSMauro Carvalho Chehab ============= =============================================================== 167c33e97efSMauro Carvalho Chehab 168c33e97efSMauro Carvalho ChehabFor example, to get the status information of a process, all you have to do is 169c33e97efSMauro Carvalho Chehabread the file /proc/PID/status:: 170c33e97efSMauro Carvalho Chehab 171c33e97efSMauro Carvalho Chehab >cat /proc/self/status 172c33e97efSMauro Carvalho Chehab Name: cat 173c33e97efSMauro Carvalho Chehab State: R (running) 174c33e97efSMauro Carvalho Chehab Tgid: 5452 175c33e97efSMauro Carvalho Chehab Pid: 5452 176c33e97efSMauro Carvalho Chehab PPid: 743 177c33e97efSMauro Carvalho Chehab TracerPid: 0 (2.4) 178c33e97efSMauro Carvalho Chehab Uid: 501 501 501 501 179c33e97efSMauro Carvalho Chehab Gid: 100 100 100 100 180c33e97efSMauro Carvalho Chehab FDSize: 256 181c33e97efSMauro Carvalho Chehab Groups: 100 14 16 182c33e97efSMauro Carvalho Chehab VmPeak: 5004 kB 183c33e97efSMauro Carvalho Chehab VmSize: 5004 kB 184c33e97efSMauro Carvalho Chehab VmLck: 0 kB 185c33e97efSMauro Carvalho Chehab VmHWM: 476 kB 186c33e97efSMauro Carvalho Chehab VmRSS: 476 kB 187c33e97efSMauro Carvalho Chehab RssAnon: 352 kB 188c33e97efSMauro Carvalho Chehab RssFile: 120 kB 189c33e97efSMauro Carvalho Chehab RssShmem: 4 kB 190c33e97efSMauro Carvalho Chehab VmData: 156 kB 191c33e97efSMauro Carvalho Chehab VmStk: 88 kB 192c33e97efSMauro Carvalho Chehab VmExe: 68 kB 193c33e97efSMauro Carvalho Chehab VmLib: 1412 kB 194c33e97efSMauro Carvalho Chehab VmPTE: 20 kb 195c33e97efSMauro Carvalho Chehab VmSwap: 0 kB 196c33e97efSMauro Carvalho Chehab HugetlbPages: 0 kB 197c33e97efSMauro Carvalho Chehab CoreDumping: 0 198c33e97efSMauro Carvalho Chehab THP_enabled: 1 199c33e97efSMauro Carvalho Chehab Threads: 1 200c33e97efSMauro Carvalho Chehab SigQ: 0/28578 201c33e97efSMauro Carvalho Chehab SigPnd: 0000000000000000 202c33e97efSMauro Carvalho Chehab ShdPnd: 0000000000000000 203c33e97efSMauro Carvalho Chehab SigBlk: 0000000000000000 204c33e97efSMauro Carvalho Chehab SigIgn: 0000000000000000 205c33e97efSMauro Carvalho Chehab SigCgt: 0000000000000000 206c33e97efSMauro Carvalho Chehab CapInh: 00000000fffffeff 207c33e97efSMauro Carvalho Chehab CapPrm: 0000000000000000 208c33e97efSMauro Carvalho Chehab CapEff: 0000000000000000 209c33e97efSMauro Carvalho Chehab CapBnd: ffffffffffffffff 210c33e97efSMauro Carvalho Chehab CapAmb: 0000000000000000 211c33e97efSMauro Carvalho Chehab NoNewPrivs: 0 212c33e97efSMauro Carvalho Chehab Seccomp: 0 213c33e97efSMauro Carvalho Chehab Speculation_Store_Bypass: thread vulnerable 214fe719888SAnand K Mistry SpeculationIndirectBranch: conditional enabled 215c33e97efSMauro Carvalho Chehab voluntary_ctxt_switches: 0 216c33e97efSMauro Carvalho Chehab nonvoluntary_ctxt_switches: 1 217c33e97efSMauro Carvalho Chehab 218c33e97efSMauro Carvalho ChehabThis shows you nearly the same information you would get if you viewed it with 219c33e97efSMauro Carvalho Chehabthe ps command. In fact, ps uses the proc file system to obtain its 220c33e97efSMauro Carvalho Chehabinformation. But you get a more detailed view of the process by reading the 221c33e97efSMauro Carvalho Chehabfile /proc/PID/status. It fields are described in table 1-2. 222c33e97efSMauro Carvalho Chehab 223c33e97efSMauro Carvalho ChehabThe statm file contains more detailed information about the process 224c33e97efSMauro Carvalho Chehabmemory usage. Its seven fields are explained in Table 1-3. The stat file 225059db434SRandy Dunlapcontains detailed information about the process itself. Its fields are 226c33e97efSMauro Carvalho Chehabexplained in Table 1-4. 227c33e97efSMauro Carvalho Chehab 228c33e97efSMauro Carvalho Chehab(for SMP CONFIG users) 229c33e97efSMauro Carvalho Chehab 230c33e97efSMauro Carvalho ChehabFor making accounting scalable, RSS related information are handled in an 231c33e97efSMauro Carvalho Chehabasynchronous manner and the value may not be very precise. To see a precise 232c33e97efSMauro Carvalho Chehabsnapshot of a moment, you can see /proc/<pid>/smaps file and scan page table. 233c33e97efSMauro Carvalho ChehabIt's slow but very precise. 234c33e97efSMauro Carvalho Chehab 235c33e97efSMauro Carvalho Chehab.. table:: Table 1-2: Contents of the status files (as of 4.19) 236c33e97efSMauro Carvalho Chehab 237c33e97efSMauro Carvalho Chehab ========================== =================================================== 238c33e97efSMauro Carvalho Chehab Field Content 239c33e97efSMauro Carvalho Chehab ========================== =================================================== 240c33e97efSMauro Carvalho Chehab Name filename of the executable 241c33e97efSMauro Carvalho Chehab Umask file mode creation mask 242c33e97efSMauro Carvalho Chehab State state (R is running, S is sleeping, D is sleeping 243c33e97efSMauro Carvalho Chehab in an uninterruptible wait, Z is zombie, 244c33e97efSMauro Carvalho Chehab T is traced or stopped) 245c33e97efSMauro Carvalho Chehab Tgid thread group ID 246c33e97efSMauro Carvalho Chehab Ngid NUMA group ID (0 if none) 247c33e97efSMauro Carvalho Chehab Pid process id 248c33e97efSMauro Carvalho Chehab PPid process id of the parent process 249df61e945SChen Linxuan TracerPid PID of process tracing this process (0 if not, or 250df61e945SChen Linxuan the tracer is outside of the current pid namespace) 251c33e97efSMauro Carvalho Chehab Uid Real, effective, saved set, and file system UIDs 252c33e97efSMauro Carvalho Chehab Gid Real, effective, saved set, and file system GIDs 253c33e97efSMauro Carvalho Chehab FDSize number of file descriptor slots currently allocated 254c33e97efSMauro Carvalho Chehab Groups supplementary group list 255c33e97efSMauro Carvalho Chehab NStgid descendant namespace thread group ID hierarchy 256c33e97efSMauro Carvalho Chehab NSpid descendant namespace process ID hierarchy 257c33e97efSMauro Carvalho Chehab NSpgid descendant namespace process group ID hierarchy 258c33e97efSMauro Carvalho Chehab NSsid descendant namespace session ID hierarchy 259c33e97efSMauro Carvalho Chehab VmPeak peak virtual memory size 260c33e97efSMauro Carvalho Chehab VmSize total program size 261c33e97efSMauro Carvalho Chehab VmLck locked memory size 262c33e97efSMauro Carvalho Chehab VmPin pinned memory size 263c33e97efSMauro Carvalho Chehab VmHWM peak resident set size ("high water mark") 264c33e97efSMauro Carvalho Chehab VmRSS size of memory portions. It contains the three 265c33e97efSMauro Carvalho Chehab following parts 266c33e97efSMauro Carvalho Chehab (VmRSS = RssAnon + RssFile + RssShmem) 267c33e97efSMauro Carvalho Chehab RssAnon size of resident anonymous memory 268c33e97efSMauro Carvalho Chehab RssFile size of resident file mappings 269c33e97efSMauro Carvalho Chehab RssShmem size of resident shmem memory (includes SysV shm, 270c33e97efSMauro Carvalho Chehab mapping of tmpfs and shared anonymous mappings) 271c33e97efSMauro Carvalho Chehab VmData size of private data segments 272c33e97efSMauro Carvalho Chehab VmStk size of stack segments 273c33e97efSMauro Carvalho Chehab VmExe size of text segment 274c33e97efSMauro Carvalho Chehab VmLib size of shared library code 275c33e97efSMauro Carvalho Chehab VmPTE size of page table entries 276c33e97efSMauro Carvalho Chehab VmSwap amount of swap used by anonymous private data 277c33e97efSMauro Carvalho Chehab (shmem swap usage is not included) 278c33e97efSMauro Carvalho Chehab HugetlbPages size of hugetlb memory portions 279c33e97efSMauro Carvalho Chehab CoreDumping process's memory is currently being dumped 280c33e97efSMauro Carvalho Chehab (killing the process may lead to a corrupted core) 281c33e97efSMauro Carvalho Chehab THP_enabled process is allowed to use THP (returns 0 when 282c33e97efSMauro Carvalho Chehab PR_SET_THP_DISABLE is set on the process 283c33e97efSMauro Carvalho Chehab Threads number of threads 284c33e97efSMauro Carvalho Chehab SigQ number of signals queued/max. number for queue 285c33e97efSMauro Carvalho Chehab SigPnd bitmap of pending signals for the thread 286c33e97efSMauro Carvalho Chehab ShdPnd bitmap of shared pending signals for the process 287c33e97efSMauro Carvalho Chehab SigBlk bitmap of blocked signals 288c33e97efSMauro Carvalho Chehab SigIgn bitmap of ignored signals 289c33e97efSMauro Carvalho Chehab SigCgt bitmap of caught signals 290c33e97efSMauro Carvalho Chehab CapInh bitmap of inheritable capabilities 291c33e97efSMauro Carvalho Chehab CapPrm bitmap of permitted capabilities 292c33e97efSMauro Carvalho Chehab CapEff bitmap of effective capabilities 293c33e97efSMauro Carvalho Chehab CapBnd bitmap of capabilities bounding set 294c33e97efSMauro Carvalho Chehab CapAmb bitmap of ambient capabilities 295c33e97efSMauro Carvalho Chehab NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...) 296c33e97efSMauro Carvalho Chehab Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...) 297c33e97efSMauro Carvalho Chehab Speculation_Store_Bypass speculative store bypass mitigation status 298fe719888SAnand K Mistry SpeculationIndirectBranch indirect branch speculation mode 299c33e97efSMauro Carvalho Chehab Cpus_allowed mask of CPUs on which this process may run 300c33e97efSMauro Carvalho Chehab Cpus_allowed_list Same as previous, but in "list format" 301c33e97efSMauro Carvalho Chehab Mems_allowed mask of memory nodes allowed to this process 302c33e97efSMauro Carvalho Chehab Mems_allowed_list Same as previous, but in "list format" 303c33e97efSMauro Carvalho Chehab voluntary_ctxt_switches number of voluntary context switches 304c33e97efSMauro Carvalho Chehab nonvoluntary_ctxt_switches number of non voluntary context switches 305c33e97efSMauro Carvalho Chehab ========================== =================================================== 306c33e97efSMauro Carvalho Chehab 307c33e97efSMauro Carvalho Chehab 308c33e97efSMauro Carvalho Chehab.. table:: Table 1-3: Contents of the statm files (as of 2.6.8-rc3) 309c33e97efSMauro Carvalho Chehab 310c33e97efSMauro Carvalho Chehab ======== =============================== ============================== 311c33e97efSMauro Carvalho Chehab Field Content 312c33e97efSMauro Carvalho Chehab ======== =============================== ============================== 313c33e97efSMauro Carvalho Chehab size total program size (pages) (same as VmSize in status) 314c33e97efSMauro Carvalho Chehab resident size of memory portions (pages) (same as VmRSS in status) 315c33e97efSMauro Carvalho Chehab shared number of pages that are shared (i.e. backed by a file, same 316c33e97efSMauro Carvalho Chehab as RssFile+RssShmem in status) 317c33e97efSMauro Carvalho Chehab trs number of pages that are 'code' (not including libs; broken, 318c33e97efSMauro Carvalho Chehab includes data segment) 319c33e97efSMauro Carvalho Chehab lrs number of pages of library (always 0 on 2.6) 320c33e97efSMauro Carvalho Chehab drs number of pages of data/stack (including libs; broken, 321c33e97efSMauro Carvalho Chehab includes library text) 322c33e97efSMauro Carvalho Chehab dt number of dirty pages (always 0 on 2.6) 323c33e97efSMauro Carvalho Chehab ======== =============================== ============================== 324c33e97efSMauro Carvalho Chehab 325c33e97efSMauro Carvalho Chehab 326c33e97efSMauro Carvalho Chehab.. table:: Table 1-4: Contents of the stat files (as of 2.6.30-rc7) 327c33e97efSMauro Carvalho Chehab 328c33e97efSMauro Carvalho Chehab ============= =============================================================== 329c33e97efSMauro Carvalho Chehab Field Content 330c33e97efSMauro Carvalho Chehab ============= =============================================================== 331c33e97efSMauro Carvalho Chehab pid process id 332c33e97efSMauro Carvalho Chehab tcomm filename of the executable 333c33e97efSMauro Carvalho Chehab state state (R is running, S is sleeping, D is sleeping in an 334c33e97efSMauro Carvalho Chehab uninterruptible wait, Z is zombie, T is traced or stopped) 335c33e97efSMauro Carvalho Chehab ppid process id of the parent process 336c33e97efSMauro Carvalho Chehab pgrp pgrp of the process 337c33e97efSMauro Carvalho Chehab sid session id 338c33e97efSMauro Carvalho Chehab tty_nr tty the process uses 339c33e97efSMauro Carvalho Chehab tty_pgrp pgrp of the tty 340c33e97efSMauro Carvalho Chehab flags task flags 341c33e97efSMauro Carvalho Chehab min_flt number of minor faults 342c33e97efSMauro Carvalho Chehab cmin_flt number of minor faults with child's 343c33e97efSMauro Carvalho Chehab maj_flt number of major faults 344c33e97efSMauro Carvalho Chehab cmaj_flt number of major faults with child's 345c33e97efSMauro Carvalho Chehab utime user mode jiffies 346c33e97efSMauro Carvalho Chehab stime kernel mode jiffies 347c33e97efSMauro Carvalho Chehab cutime user mode jiffies with child's 348c33e97efSMauro Carvalho Chehab cstime kernel mode jiffies with child's 349c33e97efSMauro Carvalho Chehab priority priority level 350c33e97efSMauro Carvalho Chehab nice nice level 351c33e97efSMauro Carvalho Chehab num_threads number of threads 352c33e97efSMauro Carvalho Chehab it_real_value (obsolete, always 0) 353c33e97efSMauro Carvalho Chehab start_time time the process started after system boot 354c33e97efSMauro Carvalho Chehab vsize virtual memory size 355c33e97efSMauro Carvalho Chehab rss resident set memory size 356c33e97efSMauro Carvalho Chehab rsslim current limit in bytes on the rss 357c33e97efSMauro Carvalho Chehab start_code address above which program text can run 358c33e97efSMauro Carvalho Chehab end_code address below which program text can run 359c33e97efSMauro Carvalho Chehab start_stack address of the start of the main process stack 360c33e97efSMauro Carvalho Chehab esp current value of ESP 361c33e97efSMauro Carvalho Chehab eip current value of EIP 362c33e97efSMauro Carvalho Chehab pending bitmap of pending signals 363c33e97efSMauro Carvalho Chehab blocked bitmap of blocked signals 364c33e97efSMauro Carvalho Chehab sigign bitmap of ignored signals 365c33e97efSMauro Carvalho Chehab sigcatch bitmap of caught signals 366c33e97efSMauro Carvalho Chehab 0 (place holder, used to be the wchan address, 367c33e97efSMauro Carvalho Chehab use /proc/PID/wchan instead) 368c33e97efSMauro Carvalho Chehab 0 (place holder) 369c33e97efSMauro Carvalho Chehab 0 (place holder) 370c33e97efSMauro Carvalho Chehab exit_signal signal to send to parent thread on exit 371c33e97efSMauro Carvalho Chehab task_cpu which CPU the task is scheduled on 372c33e97efSMauro Carvalho Chehab rt_priority realtime priority 373c33e97efSMauro Carvalho Chehab policy scheduling policy (man sched_setscheduler) 374c33e97efSMauro Carvalho Chehab blkio_ticks time spent waiting for block IO 375c33e97efSMauro Carvalho Chehab gtime guest time of the task in jiffies 376c33e97efSMauro Carvalho Chehab cgtime guest time of the task children in jiffies 377c33e97efSMauro Carvalho Chehab start_data address above which program data+bss is placed 378c33e97efSMauro Carvalho Chehab end_data address below which program data+bss is placed 379c33e97efSMauro Carvalho Chehab start_brk address above which program heap can be expanded with brk() 380c33e97efSMauro Carvalho Chehab arg_start address above which program command line is placed 381c33e97efSMauro Carvalho Chehab arg_end address below which program command line is placed 382c33e97efSMauro Carvalho Chehab env_start address above which program environment is placed 383c33e97efSMauro Carvalho Chehab env_end address below which program environment is placed 384c33e97efSMauro Carvalho Chehab exit_code the thread's exit_code in the form reported by the waitpid 385c33e97efSMauro Carvalho Chehab system call 386c33e97efSMauro Carvalho Chehab ============= =============================================================== 387c33e97efSMauro Carvalho Chehab 388c33e97efSMauro Carvalho ChehabThe /proc/PID/maps file contains the currently mapped memory regions and 389c33e97efSMauro Carvalho Chehabtheir access permissions. 390c33e97efSMauro Carvalho Chehab 391c33e97efSMauro Carvalho ChehabThe format is:: 392c33e97efSMauro Carvalho Chehab 393c33e97efSMauro Carvalho Chehab address perms offset dev inode pathname 394c33e97efSMauro Carvalho Chehab 395c33e97efSMauro Carvalho Chehab 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test 396c33e97efSMauro Carvalho Chehab 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test 397c33e97efSMauro Carvalho Chehab 0804a000-0806b000 rw-p 00000000 00:00 0 [heap] 398c33e97efSMauro Carvalho Chehab a7cb1000-a7cb2000 ---p 00000000 00:00 0 399c33e97efSMauro Carvalho Chehab a7cb2000-a7eb2000 rw-p 00000000 00:00 0 400c33e97efSMauro Carvalho Chehab a7eb2000-a7eb3000 ---p 00000000 00:00 0 401c33e97efSMauro Carvalho Chehab a7eb3000-a7ed5000 rw-p 00000000 00:00 0 402c33e97efSMauro Carvalho Chehab a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6 403c33e97efSMauro Carvalho Chehab a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6 404c33e97efSMauro Carvalho Chehab a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6 405c33e97efSMauro Carvalho Chehab a800b000-a800e000 rw-p 00000000 00:00 0 406c33e97efSMauro Carvalho Chehab a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0 407c33e97efSMauro Carvalho Chehab a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0 408c33e97efSMauro Carvalho Chehab a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0 409c33e97efSMauro Carvalho Chehab a8024000-a8027000 rw-p 00000000 00:00 0 410c33e97efSMauro Carvalho Chehab a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2 411c33e97efSMauro Carvalho Chehab a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2 412c33e97efSMauro Carvalho Chehab a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2 413c33e97efSMauro Carvalho Chehab aff35000-aff4a000 rw-p 00000000 00:00 0 [stack] 414c33e97efSMauro Carvalho Chehab ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso] 415c33e97efSMauro Carvalho Chehab 416c33e97efSMauro Carvalho Chehabwhere "address" is the address space in the process that it occupies, "perms" 417c33e97efSMauro Carvalho Chehabis a set of permissions:: 418c33e97efSMauro Carvalho Chehab 419c33e97efSMauro Carvalho Chehab r = read 420c33e97efSMauro Carvalho Chehab w = write 421c33e97efSMauro Carvalho Chehab x = execute 422c33e97efSMauro Carvalho Chehab s = shared 423c33e97efSMauro Carvalho Chehab p = private (copy on write) 424c33e97efSMauro Carvalho Chehab 425c33e97efSMauro Carvalho Chehab"offset" is the offset into the mapping, "dev" is the device (major:minor), and 426c33e97efSMauro Carvalho Chehab"inode" is the inode on that device. 0 indicates that no inode is associated 427c33e97efSMauro Carvalho Chehabwith the memory region, as the case would be with BSS (uninitialized data). 428c33e97efSMauro Carvalho ChehabThe "pathname" shows the name associated file for this mapping. If the mapping 429c33e97efSMauro Carvalho Chehabis not associated with a file: 430c33e97efSMauro Carvalho Chehab 431d09e8ca6SPasha Tatashin =================== =========================================== 432c33e97efSMauro Carvalho Chehab [heap] the heap of the program 433c33e97efSMauro Carvalho Chehab [stack] the stack of the main process 434c33e97efSMauro Carvalho Chehab [vdso] the "virtual dynamic shared object", 435c33e97efSMauro Carvalho Chehab the kernel system call handler 436d09e8ca6SPasha Tatashin [anon:<name>] a private anonymous mapping that has been 4379a10064fSColin Cross named by userspace 438d09e8ca6SPasha Tatashin [anon_shmem:<name>] an anonymous shared memory mapping that has 439d09e8ca6SPasha Tatashin been named by userspace 440d09e8ca6SPasha Tatashin =================== =========================================== 441c33e97efSMauro Carvalho Chehab 442c33e97efSMauro Carvalho Chehab or if empty, the mapping is anonymous. 443c33e97efSMauro Carvalho Chehab 444c33e97efSMauro Carvalho ChehabThe /proc/PID/smaps is an extension based on maps, showing the memory 445c33e97efSMauro Carvalho Chehabconsumption for each of the process's mappings. For each mapping (aka Virtual 446c33e97efSMauro Carvalho ChehabMemory Area, or VMA) there is a series of lines such as the following:: 447c33e97efSMauro Carvalho Chehab 448c33e97efSMauro Carvalho Chehab 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash 449c33e97efSMauro Carvalho Chehab 450c33e97efSMauro Carvalho Chehab Size: 1084 kB 451c33e97efSMauro Carvalho Chehab KernelPageSize: 4 kB 452c33e97efSMauro Carvalho Chehab MMUPageSize: 4 kB 453c33e97efSMauro Carvalho Chehab Rss: 892 kB 454c33e97efSMauro Carvalho Chehab Pss: 374 kB 45530934843SVincent Whitchurch Pss_Dirty: 0 kB 456c33e97efSMauro Carvalho Chehab Shared_Clean: 892 kB 457c33e97efSMauro Carvalho Chehab Shared_Dirty: 0 kB 458c33e97efSMauro Carvalho Chehab Private_Clean: 0 kB 459c33e97efSMauro Carvalho Chehab Private_Dirty: 0 kB 460c33e97efSMauro Carvalho Chehab Referenced: 892 kB 461c33e97efSMauro Carvalho Chehab Anonymous: 0 kB 462c33e97efSMauro Carvalho Chehab LazyFree: 0 kB 463c33e97efSMauro Carvalho Chehab AnonHugePages: 0 kB 464c33e97efSMauro Carvalho Chehab ShmemPmdMapped: 0 kB 465c33e97efSMauro Carvalho Chehab Shared_Hugetlb: 0 kB 466c33e97efSMauro Carvalho Chehab Private_Hugetlb: 0 kB 467c33e97efSMauro Carvalho Chehab Swap: 0 kB 468c33e97efSMauro Carvalho Chehab SwapPss: 0 kB 469c33e97efSMauro Carvalho Chehab KernelPageSize: 4 kB 470c33e97efSMauro Carvalho Chehab MMUPageSize: 4 kB 471c33e97efSMauro Carvalho Chehab Locked: 0 kB 472c33e97efSMauro Carvalho Chehab THPeligible: 0 473c33e97efSMauro Carvalho Chehab VmFlags: rd ex mr mw me dw 474c33e97efSMauro Carvalho Chehab 475c33e97efSMauro Carvalho ChehabThe first of these lines shows the same information as is displayed for the 476c33e97efSMauro Carvalho Chehabmapping in /proc/PID/maps. Following lines show the size of the mapping 477c33e97efSMauro Carvalho Chehab(size); the size of each page allocated when backing a VMA (KernelPageSize), 478c33e97efSMauro Carvalho Chehabwhich is usually the same as the size in the page table entries; the page size 479c33e97efSMauro Carvalho Chehabused by the MMU when backing a VMA (in most cases, the same as KernelPageSize); 480c33e97efSMauro Carvalho Chehabthe amount of the mapping that is currently resident in RAM (RSS); the 481c33e97efSMauro Carvalho Chehabprocess' proportional share of this mapping (PSS); and the number of clean and 482c33e97efSMauro Carvalho Chehabdirty shared and private pages in the mapping. 483c33e97efSMauro Carvalho Chehab 484c33e97efSMauro Carvalho ChehabThe "proportional set size" (PSS) of a process is the count of pages it has 485c33e97efSMauro Carvalho Chehabin memory, where each page is divided by the number of processes sharing it. 486c33e97efSMauro Carvalho ChehabSo if a process has 1000 pages all to itself, and 1000 shared with one other 48730934843SVincent Whitchurchprocess, its PSS will be 1500. "Pss_Dirty" is the portion of PSS which 48830934843SVincent Whitchurchconsists of dirty pages. ("Pss_Clean" is not included, but it can be 48930934843SVincent Whitchurchcalculated by subtracting "Pss_Dirty" from "Pss".) 490c33e97efSMauro Carvalho Chehab 491c33e97efSMauro Carvalho ChehabNote that even a page which is part of a MAP_SHARED mapping, but has only 492c33e97efSMauro Carvalho Chehaba single pte mapped, i.e. is currently used by only one process, is accounted 493c33e97efSMauro Carvalho Chehabas private and not as shared. 494c33e97efSMauro Carvalho Chehab 495c33e97efSMauro Carvalho Chehab"Referenced" indicates the amount of memory currently marked as referenced or 496c33e97efSMauro Carvalho Chehabaccessed. 497c33e97efSMauro Carvalho Chehab 498c33e97efSMauro Carvalho Chehab"Anonymous" shows the amount of memory that does not belong to any file. Even 499c33e97efSMauro Carvalho Chehaba mapping associated with a file may contain anonymous pages: when MAP_PRIVATE 500c33e97efSMauro Carvalho Chehaband a page is modified, the file page is replaced by a private anonymous copy. 501c33e97efSMauro Carvalho Chehab 502c33e97efSMauro Carvalho Chehab"LazyFree" shows the amount of memory which is marked by madvise(MADV_FREE). 503c33e97efSMauro Carvalho ChehabThe memory isn't freed immediately with madvise(). It's freed in memory 504c33e97efSMauro Carvalho Chehabpressure if the memory is clean. Please note that the printed value might 505c33e97efSMauro Carvalho Chehabbe lower than the real value due to optimizations used in the current 506c33e97efSMauro Carvalho Chehabimplementation. If this is not desirable please file a bug report. 507c33e97efSMauro Carvalho Chehab 508c33e97efSMauro Carvalho Chehab"AnonHugePages" shows the ammount of memory backed by transparent hugepage. 509c33e97efSMauro Carvalho Chehab 510c33e97efSMauro Carvalho Chehab"ShmemPmdMapped" shows the ammount of shared (shmem/tmpfs) memory backed by 511c33e97efSMauro Carvalho Chehabhuge pages. 512c33e97efSMauro Carvalho Chehab 513c33e97efSMauro Carvalho Chehab"Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by 514c33e97efSMauro Carvalho Chehabhugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical 515c33e97efSMauro Carvalho Chehabreasons. And these are not included in {Shared,Private}_{Clean,Dirty} field. 516c33e97efSMauro Carvalho Chehab 517c33e97efSMauro Carvalho Chehab"Swap" shows how much would-be-anonymous memory is also used, but out on swap. 518c33e97efSMauro Carvalho Chehab 519c33e97efSMauro Carvalho ChehabFor shmem mappings, "Swap" includes also the size of the mapped (and not 520c33e97efSMauro Carvalho Chehabreplaced by copy-on-write) part of the underlying shmem object out on swap. 521c33e97efSMauro Carvalho Chehab"SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this 522c33e97efSMauro Carvalho Chehabdoes not take into account swapped out page of underlying shmem objects. 523c33e97efSMauro Carvalho Chehab"Locked" indicates whether the mapping is locked in memory or not. 524cb55b838SYang Shi 525c33e97efSMauro Carvalho Chehab"THPeligible" indicates whether the mapping is eligible for allocating THP 526cb55b838SYang Shipages as well as the THP is PMD mappable or not - 1 if true, 0 otherwise. 527cb55b838SYang ShiIt just shows the current status. 528c33e97efSMauro Carvalho Chehab 529c33e97efSMauro Carvalho Chehab"VmFlags" field deserves a separate description. This member represents the 530c33e97efSMauro Carvalho Chehabkernel flags associated with the particular virtual memory area in two letter 531c33e97efSMauro Carvalho Chehabencoded manner. The codes are the following: 532c33e97efSMauro Carvalho Chehab 533c33e97efSMauro Carvalho Chehab == ======================================= 534c33e97efSMauro Carvalho Chehab rd readable 535c33e97efSMauro Carvalho Chehab wr writeable 536c33e97efSMauro Carvalho Chehab ex executable 537c33e97efSMauro Carvalho Chehab sh shared 538c33e97efSMauro Carvalho Chehab mr may read 539c33e97efSMauro Carvalho Chehab mw may write 540c33e97efSMauro Carvalho Chehab me may execute 541c33e97efSMauro Carvalho Chehab ms may share 542c33e97efSMauro Carvalho Chehab gd stack segment growns down 543c33e97efSMauro Carvalho Chehab pf pure PFN range 544c33e97efSMauro Carvalho Chehab dw disabled write to the mapped file 545c33e97efSMauro Carvalho Chehab lo pages are locked in memory 546c33e97efSMauro Carvalho Chehab io memory mapped I/O area 547c33e97efSMauro Carvalho Chehab sr sequential read advise provided 548c33e97efSMauro Carvalho Chehab rr random read advise provided 549c33e97efSMauro Carvalho Chehab dc do not copy area on fork 550c33e97efSMauro Carvalho Chehab de do not expand area on remapping 551c33e97efSMauro Carvalho Chehab ac area is accountable 552c33e97efSMauro Carvalho Chehab nr swap space is not reserved for the area 553c33e97efSMauro Carvalho Chehab ht area uses huge tlb pages 5541f7faca2SPeter Xu sf synchronous page fault 555c33e97efSMauro Carvalho Chehab ar architecture specific flag 5561f7faca2SPeter Xu wf wipe on fork 557c33e97efSMauro Carvalho Chehab dd do not include area into core dump 558c33e97efSMauro Carvalho Chehab sd soft dirty flag 559c33e97efSMauro Carvalho Chehab mm mixed map area 560c33e97efSMauro Carvalho Chehab hg huge page advise flag 561c33e97efSMauro Carvalho Chehab nh no huge page advise flag 562c33e97efSMauro Carvalho Chehab mg mergable advise flag 563d5ddc6d9SMauro Carvalho Chehab bt arm64 BTI guarded page 564868770c9SSzabolcs Nagy mt arm64 MTE allocation tags are enabled 5651f7faca2SPeter Xu um userfaultfd missing tracking 5661f7faca2SPeter Xu uw userfaultfd wr-protect tracking 567c33e97efSMauro Carvalho Chehab == ======================================= 568c33e97efSMauro Carvalho Chehab 569c33e97efSMauro Carvalho ChehabNote that there is no guarantee that every flag and associated mnemonic will 570c33e97efSMauro Carvalho Chehabbe present in all further kernel releases. Things get changed, the flags may 571c33e97efSMauro Carvalho Chehabbe vanished or the reverse -- new added. Interpretation of their meaning 572c33e97efSMauro Carvalho Chehabmight change in future as well. So each consumer of these flags has to 573c33e97efSMauro Carvalho Chehabfollow each specific kernel version for the exact semantic. 574c33e97efSMauro Carvalho Chehab 575c33e97efSMauro Carvalho ChehabThis file is only present if the CONFIG_MMU kernel configuration option is 576c33e97efSMauro Carvalho Chehabenabled. 577c33e97efSMauro Carvalho Chehab 578c33e97efSMauro Carvalho ChehabNote: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent 579c33e97efSMauro Carvalho Chehaboutput can be achieved only in the single read call). 580c33e97efSMauro Carvalho Chehab 581c33e97efSMauro Carvalho ChehabThis typically manifests when doing partial reads of these files while the 582c33e97efSMauro Carvalho Chehabmemory map is being modified. Despite the races, we do provide the following 583c33e97efSMauro Carvalho Chehabguarantees: 584c33e97efSMauro Carvalho Chehab 585c33e97efSMauro Carvalho Chehab1) The mapped addresses never go backwards, which implies no two 586c33e97efSMauro Carvalho Chehab regions will ever overlap. 587c33e97efSMauro Carvalho Chehab2) If there is something at a given vaddr during the entirety of the 588c33e97efSMauro Carvalho Chehab life of the smaps/maps walk, there will be some output for it. 589c33e97efSMauro Carvalho Chehab 590c33e97efSMauro Carvalho ChehabThe /proc/PID/smaps_rollup file includes the same fields as /proc/PID/smaps, 591c33e97efSMauro Carvalho Chehabbut their values are the sums of the corresponding values for all mappings of 592c33e97efSMauro Carvalho Chehabthe process. Additionally, it contains these fields: 593c33e97efSMauro Carvalho Chehab 594c33e97efSMauro Carvalho Chehab- Pss_Anon 595c33e97efSMauro Carvalho Chehab- Pss_File 596c33e97efSMauro Carvalho Chehab- Pss_Shmem 597c33e97efSMauro Carvalho Chehab 598c33e97efSMauro Carvalho ChehabThey represent the proportional shares of anonymous, file, and shmem pages, as 599c33e97efSMauro Carvalho Chehabdescribed for smaps above. These fields are omitted in smaps since each 600c33e97efSMauro Carvalho Chehabmapping identifies the type (anon, file, or shmem) of all pages it contains. 601c33e97efSMauro Carvalho ChehabThus all information in smaps_rollup can be derived from smaps, but at a 602c33e97efSMauro Carvalho Chehabsignificantly higher cost. 603c33e97efSMauro Carvalho Chehab 604c33e97efSMauro Carvalho ChehabThe /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG 605c33e97efSMauro Carvalho Chehabbits on both physical and virtual pages associated with a process, and the 606c33e97efSMauro Carvalho Chehabsoft-dirty bit on pte (see Documentation/admin-guide/mm/soft-dirty.rst 607c33e97efSMauro Carvalho Chehabfor details). 608c33e97efSMauro Carvalho ChehabTo clear the bits for all the pages associated with the process:: 609c33e97efSMauro Carvalho Chehab 610c33e97efSMauro Carvalho Chehab > echo 1 > /proc/PID/clear_refs 611c33e97efSMauro Carvalho Chehab 612c33e97efSMauro Carvalho ChehabTo clear the bits for the anonymous pages associated with the process:: 613c33e97efSMauro Carvalho Chehab 614c33e97efSMauro Carvalho Chehab > echo 2 > /proc/PID/clear_refs 615c33e97efSMauro Carvalho Chehab 616c33e97efSMauro Carvalho ChehabTo clear the bits for the file mapped pages associated with the process:: 617c33e97efSMauro Carvalho Chehab 618c33e97efSMauro Carvalho Chehab > echo 3 > /proc/PID/clear_refs 619c33e97efSMauro Carvalho Chehab 620c33e97efSMauro Carvalho ChehabTo clear the soft-dirty bit:: 621c33e97efSMauro Carvalho Chehab 622c33e97efSMauro Carvalho Chehab > echo 4 > /proc/PID/clear_refs 623c33e97efSMauro Carvalho Chehab 624c33e97efSMauro Carvalho ChehabTo reset the peak resident set size ("high water mark") to the process's 625c33e97efSMauro Carvalho Chehabcurrent value:: 626c33e97efSMauro Carvalho Chehab 627c33e97efSMauro Carvalho Chehab > echo 5 > /proc/PID/clear_refs 628c33e97efSMauro Carvalho Chehab 629c33e97efSMauro Carvalho ChehabAny other value written to /proc/PID/clear_refs will have no effect. 630c33e97efSMauro Carvalho Chehab 631c33e97efSMauro Carvalho ChehabThe /proc/pid/pagemap gives the PFN, which can be used to find the pageflags 632c33e97efSMauro Carvalho Chehabusing /proc/kpageflags and number of times a page is mapped using 633c33e97efSMauro Carvalho Chehab/proc/kpagecount. For detailed explanation, see 634c33e97efSMauro Carvalho ChehabDocumentation/admin-guide/mm/pagemap.rst. 635c33e97efSMauro Carvalho Chehab 636c33e97efSMauro Carvalho ChehabThe /proc/pid/numa_maps is an extension based on maps, showing the memory 637c33e97efSMauro Carvalho Chehablocality and binding policy, as well as the memory usage (in pages) of 638c33e97efSMauro Carvalho Chehabeach mapping. The output follows a general format where mapping details get 639c33e97efSMauro Carvalho Chehabsummarized separated by blank spaces, one mapping per each file line:: 640c33e97efSMauro Carvalho Chehab 641c33e97efSMauro Carvalho Chehab address policy mapping details 642c33e97efSMauro Carvalho Chehab 643c33e97efSMauro Carvalho Chehab 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4 644c33e97efSMauro Carvalho Chehab 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4 645c33e97efSMauro Carvalho Chehab 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4 646c33e97efSMauro Carvalho Chehab 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 647c33e97efSMauro Carvalho Chehab 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 648c33e97efSMauro Carvalho Chehab 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4 649c33e97efSMauro Carvalho Chehab 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4 650c33e97efSMauro Carvalho Chehab 320698b000 default file=/lib64/libc-2.12.so 651c33e97efSMauro Carvalho Chehab 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4 652c33e97efSMauro Carvalho Chehab 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4 653c33e97efSMauro Carvalho Chehab 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4 654c33e97efSMauro Carvalho Chehab 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4 655c33e97efSMauro Carvalho Chehab 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4 656c33e97efSMauro Carvalho Chehab 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048 657c33e97efSMauro Carvalho Chehab 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4 658c33e97efSMauro Carvalho Chehab 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4 659c33e97efSMauro Carvalho Chehab 660c33e97efSMauro Carvalho ChehabWhere: 661c33e97efSMauro Carvalho Chehab 662c33e97efSMauro Carvalho Chehab"address" is the starting address for the mapping; 663c33e97efSMauro Carvalho Chehab 664c33e97efSMauro Carvalho Chehab"policy" reports the NUMA memory policy set for the mapping (see Documentation/admin-guide/mm/numa_memory_policy.rst); 665c33e97efSMauro Carvalho Chehab 666c33e97efSMauro Carvalho Chehab"mapping details" summarizes mapping data such as mapping type, page usage counters, 667c33e97efSMauro Carvalho Chehabnode locality page counters (N0 == node0, N1 == node1, ...) and the kernel page 668c33e97efSMauro Carvalho Chehabsize, in KB, that is backing the mapping up. 669c33e97efSMauro Carvalho Chehab 670c33e97efSMauro Carvalho Chehab1.2 Kernel data 671c33e97efSMauro Carvalho Chehab--------------- 672c33e97efSMauro Carvalho Chehab 673c33e97efSMauro Carvalho ChehabSimilar to the process entries, the kernel data files give information about 674c33e97efSMauro Carvalho Chehabthe running kernel. The files used to obtain this information are contained in 675c33e97efSMauro Carvalho Chehab/proc and are listed in Table 1-5. Not all of these will be present in your 676c33e97efSMauro Carvalho Chehabsystem. It depends on the kernel configuration and the loaded modules, which 677c33e97efSMauro Carvalho Chehabfiles are there, and which are missing. 678c33e97efSMauro Carvalho Chehab 679c33e97efSMauro Carvalho Chehab.. table:: Table 1-5: Kernel info in /proc 680c33e97efSMauro Carvalho Chehab 681c33e97efSMauro Carvalho Chehab ============ =============================================================== 682c33e97efSMauro Carvalho Chehab File Content 683c33e97efSMauro Carvalho Chehab ============ =============================================================== 684c33e97efSMauro Carvalho Chehab apm Advanced power management info 685c33e97efSMauro Carvalho Chehab buddyinfo Kernel memory allocator information (see text) (2.5) 686c33e97efSMauro Carvalho Chehab bus Directory containing bus specific information 687c33e97efSMauro Carvalho Chehab cmdline Kernel command line 688c33e97efSMauro Carvalho Chehab cpuinfo Info about the CPU 689c33e97efSMauro Carvalho Chehab devices Available devices (block and character) 690c33e97efSMauro Carvalho Chehab dma Used DMS channels 691c33e97efSMauro Carvalho Chehab filesystems Supported filesystems 692c33e97efSMauro Carvalho Chehab driver Various drivers grouped here, currently rtc (2.4) 693c33e97efSMauro Carvalho Chehab execdomains Execdomains, related to security (2.4) 694c33e97efSMauro Carvalho Chehab fb Frame Buffer devices (2.4) 695c33e97efSMauro Carvalho Chehab fs File system parameters, currently nfs/exports (2.4) 696c33e97efSMauro Carvalho Chehab ide Directory containing info about the IDE subsystem 697c33e97efSMauro Carvalho Chehab interrupts Interrupt usage 698c33e97efSMauro Carvalho Chehab iomem Memory map (2.4) 699c33e97efSMauro Carvalho Chehab ioports I/O port usage 700c33e97efSMauro Carvalho Chehab irq Masks for irq to cpu affinity (2.4)(smp?) 701c33e97efSMauro Carvalho Chehab isapnp ISA PnP (Plug&Play) Info (2.4) 702c33e97efSMauro Carvalho Chehab kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4)) 703c33e97efSMauro Carvalho Chehab kmsg Kernel messages 704c33e97efSMauro Carvalho Chehab ksyms Kernel symbol table 7054ba1d726SRandy Dunlap loadavg Load average of last 1, 5 & 15 minutes; 7064ba1d726SRandy Dunlap number of processes currently runnable (running or on ready queue); 7074ba1d726SRandy Dunlap total number of processes in system; 7084ba1d726SRandy Dunlap last pid created. 70993ea4a0bSRandy Dunlap All fields are separated by one space except "number of 71093ea4a0bSRandy Dunlap processes currently runnable" and "total number of processes 71193ea4a0bSRandy Dunlap in system", which are separated by a slash ('/'). Example: 71293ea4a0bSRandy Dunlap 0.61 0.61 0.55 3/828 22084 713c33e97efSMauro Carvalho Chehab locks Kernel locks 714c33e97efSMauro Carvalho Chehab meminfo Memory info 715c33e97efSMauro Carvalho Chehab misc Miscellaneous 716c33e97efSMauro Carvalho Chehab modules List of loaded modules 717c33e97efSMauro Carvalho Chehab mounts Mounted filesystems 718c33e97efSMauro Carvalho Chehab net Networking info (see text) 719c33e97efSMauro Carvalho Chehab pagetypeinfo Additional page allocator information (see text) (2.5) 720c33e97efSMauro Carvalho Chehab partitions Table of partitions known to the system 721c33e97efSMauro Carvalho Chehab pci Deprecated info of PCI bus (new way -> /proc/bus/pci/, 722c33e97efSMauro Carvalho Chehab decoupled by lspci (2.4) 723c33e97efSMauro Carvalho Chehab rtc Real time clock 724c33e97efSMauro Carvalho Chehab scsi SCSI info (see text) 725c33e97efSMauro Carvalho Chehab slabinfo Slab pool info 726c33e97efSMauro Carvalho Chehab softirqs softirq usage 727c33e97efSMauro Carvalho Chehab stat Overall statistics 728c33e97efSMauro Carvalho Chehab swaps Swap space utilization 729c33e97efSMauro Carvalho Chehab sys See chapter 2 730c33e97efSMauro Carvalho Chehab sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4) 731c33e97efSMauro Carvalho Chehab tty Info of tty drivers 732c33e97efSMauro Carvalho Chehab uptime Wall clock since boot, combined idle time of all cpus 733c33e97efSMauro Carvalho Chehab version Kernel version 734c33e97efSMauro Carvalho Chehab video bttv info of video resources (2.4) 735c33e97efSMauro Carvalho Chehab vmallocinfo Show vmalloced areas 736c33e97efSMauro Carvalho Chehab ============ =============================================================== 737c33e97efSMauro Carvalho Chehab 738c33e97efSMauro Carvalho ChehabYou can, for example, check which interrupts are currently in use and what 739c33e97efSMauro Carvalho Chehabthey are used for by looking in the file /proc/interrupts:: 740c33e97efSMauro Carvalho Chehab 741c33e97efSMauro Carvalho Chehab > cat /proc/interrupts 742c33e97efSMauro Carvalho Chehab CPU0 743c33e97efSMauro Carvalho Chehab 0: 8728810 XT-PIC timer 744c33e97efSMauro Carvalho Chehab 1: 895 XT-PIC keyboard 745c33e97efSMauro Carvalho Chehab 2: 0 XT-PIC cascade 746c33e97efSMauro Carvalho Chehab 3: 531695 XT-PIC aha152x 747c33e97efSMauro Carvalho Chehab 4: 2014133 XT-PIC serial 748c33e97efSMauro Carvalho Chehab 5: 44401 XT-PIC pcnet_cs 749c33e97efSMauro Carvalho Chehab 8: 2 XT-PIC rtc 750c33e97efSMauro Carvalho Chehab 11: 8 XT-PIC i82365 751c33e97efSMauro Carvalho Chehab 12: 182918 XT-PIC PS/2 Mouse 752c33e97efSMauro Carvalho Chehab 13: 1 XT-PIC fpu 753c33e97efSMauro Carvalho Chehab 14: 1232265 XT-PIC ide0 754c33e97efSMauro Carvalho Chehab 15: 7 XT-PIC ide1 755c33e97efSMauro Carvalho Chehab NMI: 0 756c33e97efSMauro Carvalho Chehab 757c33e97efSMauro Carvalho ChehabIn 2.4.* a couple of lines where added to this file LOC & ERR (this time is the 758c33e97efSMauro Carvalho Chehaboutput of a SMP machine):: 759c33e97efSMauro Carvalho Chehab 760c33e97efSMauro Carvalho Chehab > cat /proc/interrupts 761c33e97efSMauro Carvalho Chehab 762c33e97efSMauro Carvalho Chehab CPU0 CPU1 763c33e97efSMauro Carvalho Chehab 0: 1243498 1214548 IO-APIC-edge timer 764c33e97efSMauro Carvalho Chehab 1: 8949 8958 IO-APIC-edge keyboard 765c33e97efSMauro Carvalho Chehab 2: 0 0 XT-PIC cascade 766c33e97efSMauro Carvalho Chehab 5: 11286 10161 IO-APIC-edge soundblaster 767c33e97efSMauro Carvalho Chehab 8: 1 0 IO-APIC-edge rtc 768c33e97efSMauro Carvalho Chehab 9: 27422 27407 IO-APIC-edge 3c503 769c33e97efSMauro Carvalho Chehab 12: 113645 113873 IO-APIC-edge PS/2 Mouse 770c33e97efSMauro Carvalho Chehab 13: 0 0 XT-PIC fpu 771c33e97efSMauro Carvalho Chehab 14: 22491 24012 IO-APIC-edge ide0 772c33e97efSMauro Carvalho Chehab 15: 2183 2415 IO-APIC-edge ide1 773c33e97efSMauro Carvalho Chehab 17: 30564 30414 IO-APIC-level eth0 774c33e97efSMauro Carvalho Chehab 18: 177 164 IO-APIC-level bttv 775c33e97efSMauro Carvalho Chehab NMI: 2457961 2457959 776c33e97efSMauro Carvalho Chehab LOC: 2457882 2457881 777c33e97efSMauro Carvalho Chehab ERR: 2155 778c33e97efSMauro Carvalho Chehab 779c33e97efSMauro Carvalho ChehabNMI is incremented in this case because every timer interrupt generates a NMI 780c33e97efSMauro Carvalho Chehab(Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups. 781c33e97efSMauro Carvalho Chehab 782c33e97efSMauro Carvalho ChehabLOC is the local interrupt counter of the internal APIC of every CPU. 783c33e97efSMauro Carvalho Chehab 784c33e97efSMauro Carvalho ChehabERR is incremented in the case of errors in the IO-APIC bus (the bus that 785c33e97efSMauro Carvalho Chehabconnects the CPUs in a SMP system. This means that an error has been detected, 786c33e97efSMauro Carvalho Chehabthe IO-APIC automatically retry the transmission, so it should not be a big 787c33e97efSMauro Carvalho Chehabproblem, but you should read the SMP-FAQ. 788c33e97efSMauro Carvalho Chehab 789c33e97efSMauro Carvalho ChehabIn 2.6.2* /proc/interrupts was expanded again. This time the goal was for 790c33e97efSMauro Carvalho Chehab/proc/interrupts to display every IRQ vector in use by the system, not 791c33e97efSMauro Carvalho Chehabjust those considered 'most important'. The new vectors are: 792c33e97efSMauro Carvalho Chehab 793c33e97efSMauro Carvalho ChehabTHR 794c33e97efSMauro Carvalho Chehab interrupt raised when a machine check threshold counter 795c33e97efSMauro Carvalho Chehab (typically counting ECC corrected errors of memory or cache) exceeds 796c33e97efSMauro Carvalho Chehab a configurable threshold. Only available on some systems. 797c33e97efSMauro Carvalho Chehab 798c33e97efSMauro Carvalho ChehabTRM 799c33e97efSMauro Carvalho Chehab a thermal event interrupt occurs when a temperature threshold 800c33e97efSMauro Carvalho Chehab has been exceeded for the CPU. This interrupt may also be generated 801c33e97efSMauro Carvalho Chehab when the temperature drops back to normal. 802c33e97efSMauro Carvalho Chehab 803c33e97efSMauro Carvalho ChehabSPU 804c33e97efSMauro Carvalho Chehab a spurious interrupt is some interrupt that was raised then lowered 805c33e97efSMauro Carvalho Chehab by some IO device before it could be fully processed by the APIC. Hence 806c33e97efSMauro Carvalho Chehab the APIC sees the interrupt but does not know what device it came from. 807c33e97efSMauro Carvalho Chehab For this case the APIC will generate the interrupt with a IRQ vector 808c33e97efSMauro Carvalho Chehab of 0xff. This might also be generated by chipset bugs. 809c33e97efSMauro Carvalho Chehab 810059db434SRandy DunlapRES, CAL, TLB 811c33e97efSMauro Carvalho Chehab rescheduling, call and TLB flush interrupts are 812c33e97efSMauro Carvalho Chehab sent from one CPU to another per the needs of the OS. Typically, 813c33e97efSMauro Carvalho Chehab their statistics are used by kernel developers and interested users to 814c33e97efSMauro Carvalho Chehab determine the occurrence of interrupts of the given type. 815c33e97efSMauro Carvalho Chehab 816c33e97efSMauro Carvalho ChehabThe above IRQ vectors are displayed only when relevant. For example, 817c33e97efSMauro Carvalho Chehabthe threshold vector does not exist on x86_64 platforms. Others are 818c33e97efSMauro Carvalho Chehabsuppressed when the system is a uniprocessor. As of this writing, only 819c33e97efSMauro Carvalho Chehabi386 and x86_64 platforms support the new IRQ vector displays. 820c33e97efSMauro Carvalho Chehab 821c33e97efSMauro Carvalho ChehabOf some interest is the introduction of the /proc/irq directory to 2.4. 822059db434SRandy DunlapIt could be used to set IRQ to CPU affinity. This means that you can "hook" an 823c33e97efSMauro Carvalho ChehabIRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the 824c33e97efSMauro Carvalho Chehabirq subdir is one subdir for each IRQ, and two files; default_smp_affinity and 825c33e97efSMauro Carvalho Chehabprof_cpu_mask. 826c33e97efSMauro Carvalho Chehab 827c33e97efSMauro Carvalho ChehabFor example:: 828c33e97efSMauro Carvalho Chehab 829c33e97efSMauro Carvalho Chehab > ls /proc/irq/ 830c33e97efSMauro Carvalho Chehab 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask 831c33e97efSMauro Carvalho Chehab 1 11 13 15 17 19 3 5 7 9 default_smp_affinity 832c33e97efSMauro Carvalho Chehab > ls /proc/irq/0/ 833c33e97efSMauro Carvalho Chehab smp_affinity 834c33e97efSMauro Carvalho Chehab 835c33e97efSMauro Carvalho Chehabsmp_affinity is a bitmask, in which you can specify which CPUs can handle the 836059db434SRandy DunlapIRQ. You can set it by doing:: 837c33e97efSMauro Carvalho Chehab 838c33e97efSMauro Carvalho Chehab > echo 1 > /proc/irq/10/smp_affinity 839c33e97efSMauro Carvalho Chehab 840c33e97efSMauro Carvalho ChehabThis means that only the first CPU will handle the IRQ, but you can also echo 841c33e97efSMauro Carvalho Chehab5 which means that only the first and third CPU can handle the IRQ. 842c33e97efSMauro Carvalho Chehab 843c33e97efSMauro Carvalho ChehabThe contents of each smp_affinity file is the same by default:: 844c33e97efSMauro Carvalho Chehab 845c33e97efSMauro Carvalho Chehab > cat /proc/irq/0/smp_affinity 846c33e97efSMauro Carvalho Chehab ffffffff 847c33e97efSMauro Carvalho Chehab 848c33e97efSMauro Carvalho ChehabThere is an alternate interface, smp_affinity_list which allows specifying 849059db434SRandy Dunlapa CPU range instead of a bitmask:: 850c33e97efSMauro Carvalho Chehab 851c33e97efSMauro Carvalho Chehab > cat /proc/irq/0/smp_affinity_list 852c33e97efSMauro Carvalho Chehab 1024-1031 853c33e97efSMauro Carvalho Chehab 854c33e97efSMauro Carvalho ChehabThe default_smp_affinity mask applies to all non-active IRQs, which are the 855c33e97efSMauro Carvalho ChehabIRQs which have not yet been allocated/activated, and hence which lack a 856c33e97efSMauro Carvalho Chehab/proc/irq/[0-9]* directory. 857c33e97efSMauro Carvalho Chehab 858c33e97efSMauro Carvalho ChehabThe node file on an SMP system shows the node to which the device using the IRQ 859c33e97efSMauro Carvalho Chehabreports itself as being attached. This hardware locality information does not 860c33e97efSMauro Carvalho Chehabinclude information about any possible driver locality preference. 861c33e97efSMauro Carvalho Chehab 862c33e97efSMauro Carvalho Chehabprof_cpu_mask specifies which CPUs are to be profiled by the system wide 863059db434SRandy Dunlapprofiler. Default value is ffffffff (all CPUs if there are only 32 of them). 864c33e97efSMauro Carvalho Chehab 865c33e97efSMauro Carvalho ChehabThe way IRQs are routed is handled by the IO-APIC, and it's Round Robin 866c33e97efSMauro Carvalho Chehabbetween all the CPUs which are allowed to handle it. As usual the kernel has 867c33e97efSMauro Carvalho Chehabmore info than you and does a better job than you, so the defaults are the 868c33e97efSMauro Carvalho Chehabbest choice for almost everyone. [Note this applies only to those IO-APIC's 869c33e97efSMauro Carvalho Chehabthat support "Round Robin" interrupt distribution.] 870c33e97efSMauro Carvalho Chehab 871c33e97efSMauro Carvalho ChehabThere are three more important subdirectories in /proc: net, scsi, and sys. 872c33e97efSMauro Carvalho ChehabThe general rule is that the contents, or even the existence of these 873c33e97efSMauro Carvalho Chehabdirectories, depend on your kernel configuration. If SCSI is not enabled, the 874c33e97efSMauro Carvalho Chehabdirectory scsi may not exist. The same is true with the net, which is there 875c33e97efSMauro Carvalho Chehabonly when networking support is present in the running kernel. 876c33e97efSMauro Carvalho Chehab 877c33e97efSMauro Carvalho ChehabThe slabinfo file gives information about memory usage at the slab level. 878c33e97efSMauro Carvalho ChehabLinux uses slab pools for memory management above page level in version 2.2. 879c33e97efSMauro Carvalho ChehabCommonly used objects have their own slab pool (such as network buffers, 880c33e97efSMauro Carvalho Chehabdirectory cache, and so on). 881c33e97efSMauro Carvalho Chehab 882c33e97efSMauro Carvalho Chehab:: 883c33e97efSMauro Carvalho Chehab 884c33e97efSMauro Carvalho Chehab > cat /proc/buddyinfo 885c33e97efSMauro Carvalho Chehab 886c33e97efSMauro Carvalho Chehab Node 0, zone DMA 0 4 5 4 4 3 ... 887c33e97efSMauro Carvalho Chehab Node 0, zone Normal 1 0 0 1 101 8 ... 888c33e97efSMauro Carvalho Chehab Node 0, zone HighMem 2 0 0 1 1 0 ... 889c33e97efSMauro Carvalho Chehab 890c33e97efSMauro Carvalho ChehabExternal fragmentation is a problem under some workloads, and buddyinfo is a 891c33e97efSMauro Carvalho Chehabuseful tool for helping diagnose these problems. Buddyinfo will give you a 892c33e97efSMauro Carvalho Chehabclue as to how big an area you can safely allocate, or why a previous 893c33e97efSMauro Carvalho Chehaballocation failed. 894c33e97efSMauro Carvalho Chehab 895c33e97efSMauro Carvalho ChehabEach column represents the number of pages of a certain order which are 896c33e97efSMauro Carvalho Chehabavailable. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in 897c33e97efSMauro Carvalho ChehabZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE 898c33e97efSMauro Carvalho Chehabavailable in ZONE_NORMAL, etc... 899c33e97efSMauro Carvalho Chehab 900c33e97efSMauro Carvalho ChehabMore information relevant to external fragmentation can be found in 901c33e97efSMauro Carvalho Chehabpagetypeinfo:: 902c33e97efSMauro Carvalho Chehab 903c33e97efSMauro Carvalho Chehab > cat /proc/pagetypeinfo 904c33e97efSMauro Carvalho Chehab Page block order: 9 905c33e97efSMauro Carvalho Chehab Pages per block: 512 906c33e97efSMauro Carvalho Chehab 907c33e97efSMauro Carvalho Chehab Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 908c33e97efSMauro Carvalho Chehab Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0 909c33e97efSMauro Carvalho Chehab Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 910c33e97efSMauro Carvalho Chehab Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2 911c33e97efSMauro Carvalho Chehab Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0 912c33e97efSMauro Carvalho Chehab Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 913c33e97efSMauro Carvalho Chehab Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9 914c33e97efSMauro Carvalho Chehab Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0 915c33e97efSMauro Carvalho Chehab Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452 916c33e97efSMauro Carvalho Chehab Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0 917c33e97efSMauro Carvalho Chehab Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 918c33e97efSMauro Carvalho Chehab 919c33e97efSMauro Carvalho Chehab Number of blocks type Unmovable Reclaimable Movable Reserve Isolate 920c33e97efSMauro Carvalho Chehab Node 0, zone DMA 2 0 5 1 0 921c33e97efSMauro Carvalho Chehab Node 0, zone DMA32 41 6 967 2 0 922c33e97efSMauro Carvalho Chehab 923c33e97efSMauro Carvalho ChehabFragmentation avoidance in the kernel works by grouping pages of different 924c33e97efSMauro Carvalho Chehabmigrate types into the same contiguous regions of memory called page blocks. 925059db434SRandy DunlapA page block is typically the size of the default hugepage size, e.g. 2MB on 926c33e97efSMauro Carvalho ChehabX86-64. By keeping pages grouped based on their ability to move, the kernel 927c33e97efSMauro Carvalho Chehabcan reclaim pages within a page block to satisfy a high-order allocation. 928c33e97efSMauro Carvalho Chehab 929c33e97efSMauro Carvalho ChehabThe pagetypinfo begins with information on the size of a page block. It 930c33e97efSMauro Carvalho Chehabthen gives the same type of information as buddyinfo except broken down 931c33e97efSMauro Carvalho Chehabby migrate-type and finishes with details on how many page blocks of each 932c33e97efSMauro Carvalho Chehabtype exist. 933c33e97efSMauro Carvalho Chehab 934c33e97efSMauro Carvalho ChehabIf min_free_kbytes has been tuned correctly (recommendations made by hugeadm 935c33e97efSMauro Carvalho Chehabfrom libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can 936c33e97efSMauro Carvalho Chehabmake an estimate of the likely number of huge pages that can be allocated 937c33e97efSMauro Carvalho Chehabat a given point in time. All the "Movable" blocks should be allocatable 938c33e97efSMauro Carvalho Chehabunless memory has been mlock()'d. Some of the Reclaimable blocks should 939c33e97efSMauro Carvalho Chehabalso be allocatable although a lot of filesystem metadata may have to be 940c33e97efSMauro Carvalho Chehabreclaimed to achieve this. 941c33e97efSMauro Carvalho Chehab 942c33e97efSMauro Carvalho Chehab 943c33e97efSMauro Carvalho Chehabmeminfo 944c33e97efSMauro Carvalho Chehab~~~~~~~ 945c33e97efSMauro Carvalho Chehab 946c33e97efSMauro Carvalho ChehabProvides information about distribution and utilization of memory. This 9478d719afcSMike Rapoportvaries by architecture and compile options. Some of the counters reported 9488d719afcSMike Rapoporthere overlap. The memory reported by the non overlapping counters may not 9498d719afcSMike Rapoportadd up to the overall memory usage and the difference for some workloads 9508d719afcSMike Rapoportcan be substantial. In many cases there are other means to find out 9518d719afcSMike Rapoportadditional memory using subsystem specific interfaces, for instance 9528d719afcSMike Rapoport/proc/net/sockstat for TCP memory allocations. 9538d719afcSMike Rapoport 95439799b64SJohannes WeinerExample output. You may not have all of these fields. 955c33e97efSMauro Carvalho Chehab 956c33e97efSMauro Carvalho Chehab:: 957c33e97efSMauro Carvalho Chehab 958c33e97efSMauro Carvalho Chehab > cat /proc/meminfo 959c33e97efSMauro Carvalho Chehab 96039799b64SJohannes Weiner MemTotal: 32858820 kB 96139799b64SJohannes Weiner MemFree: 21001236 kB 96239799b64SJohannes Weiner MemAvailable: 27214312 kB 96339799b64SJohannes Weiner Buffers: 581092 kB 96439799b64SJohannes Weiner Cached: 5587612 kB 965c33e97efSMauro Carvalho Chehab SwapCached: 0 kB 96639799b64SJohannes Weiner Active: 3237152 kB 96739799b64SJohannes Weiner Inactive: 7586256 kB 96839799b64SJohannes Weiner Active(anon): 94064 kB 96939799b64SJohannes Weiner Inactive(anon): 4570616 kB 97039799b64SJohannes Weiner Active(file): 3143088 kB 97139799b64SJohannes Weiner Inactive(file): 3015640 kB 97239799b64SJohannes Weiner Unevictable: 0 kB 97339799b64SJohannes Weiner Mlocked: 0 kB 974c33e97efSMauro Carvalho Chehab SwapTotal: 0 kB 975c33e97efSMauro Carvalho Chehab SwapFree: 0 kB 976f6498b77SJohannes Weiner Zswap: 1904 kB 977f6498b77SJohannes Weiner Zswapped: 7792 kB 97839799b64SJohannes Weiner Dirty: 12 kB 979c33e97efSMauro Carvalho Chehab Writeback: 0 kB 98039799b64SJohannes Weiner AnonPages: 4654780 kB 98139799b64SJohannes Weiner Mapped: 266244 kB 98239799b64SJohannes Weiner Shmem: 9976 kB 98339799b64SJohannes Weiner KReclaimable: 517708 kB 98439799b64SJohannes Weiner Slab: 660044 kB 98539799b64SJohannes Weiner SReclaimable: 517708 kB 98639799b64SJohannes Weiner SUnreclaim: 142336 kB 98739799b64SJohannes Weiner KernelStack: 11168 kB 98839799b64SJohannes Weiner PageTables: 20540 kB 989ebc97a52SYosry Ahmed SecPageTables: 0 kB 990c33e97efSMauro Carvalho Chehab NFS_Unstable: 0 kB 991c33e97efSMauro Carvalho Chehab Bounce: 0 kB 992c33e97efSMauro Carvalho Chehab WritebackTmp: 0 kB 99339799b64SJohannes Weiner CommitLimit: 16429408 kB 99439799b64SJohannes Weiner Committed_AS: 7715148 kB 99539799b64SJohannes Weiner VmallocTotal: 34359738367 kB 99639799b64SJohannes Weiner VmallocUsed: 40444 kB 99739799b64SJohannes Weiner VmallocChunk: 0 kB 99839799b64SJohannes Weiner Percpu: 29312 kB 999*bd23024bSTomas Mudrunka EarlyMemtestBad: 0 kB 1000c33e97efSMauro Carvalho Chehab HardwareCorrupted: 0 kB 100139799b64SJohannes Weiner AnonHugePages: 4149248 kB 1002c33e97efSMauro Carvalho Chehab ShmemHugePages: 0 kB 1003c33e97efSMauro Carvalho Chehab ShmemPmdMapped: 0 kB 100439799b64SJohannes Weiner FileHugePages: 0 kB 100539799b64SJohannes Weiner FilePmdMapped: 0 kB 100639799b64SJohannes Weiner CmaTotal: 0 kB 100739799b64SJohannes Weiner CmaFree: 0 kB 100839799b64SJohannes Weiner HugePages_Total: 0 100939799b64SJohannes Weiner HugePages_Free: 0 101039799b64SJohannes Weiner HugePages_Rsvd: 0 101139799b64SJohannes Weiner HugePages_Surp: 0 101239799b64SJohannes Weiner Hugepagesize: 2048 kB 101339799b64SJohannes Weiner Hugetlb: 0 kB 101439799b64SJohannes Weiner DirectMap4k: 401152 kB 101539799b64SJohannes Weiner DirectMap2M: 10008576 kB 101639799b64SJohannes Weiner DirectMap1G: 24117248 kB 1017c33e97efSMauro Carvalho Chehab 1018c33e97efSMauro Carvalho ChehabMemTotal 1019059db434SRandy Dunlap Total usable RAM (i.e. physical RAM minus a few reserved 1020c33e97efSMauro Carvalho Chehab bits and the kernel binary code) 1021c33e97efSMauro Carvalho ChehabMemFree 102239799b64SJohannes Weiner Total free RAM. On highmem systems, the sum of LowFree+HighFree 1023c33e97efSMauro Carvalho ChehabMemAvailable 1024c33e97efSMauro Carvalho Chehab An estimate of how much memory is available for starting new 1025c33e97efSMauro Carvalho Chehab applications, without swapping. Calculated from MemFree, 1026c33e97efSMauro Carvalho Chehab SReclaimable, the size of the file LRU lists, and the low 1027c33e97efSMauro Carvalho Chehab watermarks in each zone. 1028c33e97efSMauro Carvalho Chehab The estimate takes into account that the system needs some 1029c33e97efSMauro Carvalho Chehab page cache to function well, and that not all reclaimable 1030c33e97efSMauro Carvalho Chehab slab will be reclaimable, due to items being in use. The 1031c33e97efSMauro Carvalho Chehab impact of those factors will vary from system to system. 1032c33e97efSMauro Carvalho ChehabBuffers 1033c33e97efSMauro Carvalho Chehab Relatively temporary storage for raw disk blocks 1034c33e97efSMauro Carvalho Chehab shouldn't get tremendously large (20MB or so) 1035c33e97efSMauro Carvalho ChehabCached 103639799b64SJohannes Weiner In-memory cache for files read from the disk (the 103739799b64SJohannes Weiner pagecache) as well as tmpfs & shmem. 103839799b64SJohannes Weiner Doesn't include SwapCached. 1039c33e97efSMauro Carvalho ChehabSwapCached 1040c33e97efSMauro Carvalho Chehab Memory that once was swapped out, is swapped back in but 1041c33e97efSMauro Carvalho Chehab still also is in the swapfile (if memory is needed it 1042c33e97efSMauro Carvalho Chehab doesn't need to be swapped out AGAIN because it is already 1043c33e97efSMauro Carvalho Chehab in the swapfile. This saves I/O) 1044c33e97efSMauro Carvalho ChehabActive 1045c33e97efSMauro Carvalho Chehab Memory that has been used more recently and usually not 1046c33e97efSMauro Carvalho Chehab reclaimed unless absolutely necessary. 1047c33e97efSMauro Carvalho ChehabInactive 1048c33e97efSMauro Carvalho Chehab Memory which has been less recently used. It is more 1049c33e97efSMauro Carvalho Chehab eligible to be reclaimed for other purposes 105039799b64SJohannes WeinerUnevictable 105139799b64SJohannes Weiner Memory allocated for userspace which cannot be reclaimed, such 105239799b64SJohannes Weiner as mlocked pages, ramfs backing pages, secret memfd pages etc. 105339799b64SJohannes WeinerMlocked 105439799b64SJohannes Weiner Memory locked with mlock(). 1055c33e97efSMauro Carvalho ChehabHighTotal, HighFree 1056059db434SRandy Dunlap Highmem is all memory above ~860MB of physical memory. 1057c33e97efSMauro Carvalho Chehab Highmem areas are for use by userspace programs, or 1058c33e97efSMauro Carvalho Chehab for the pagecache. The kernel must use tricks to access 1059c33e97efSMauro Carvalho Chehab this memory, making it slower to access than lowmem. 1060c33e97efSMauro Carvalho ChehabLowTotal, LowFree 1061c33e97efSMauro Carvalho Chehab Lowmem is memory which can be used for everything that 1062c33e97efSMauro Carvalho Chehab highmem can be used for, but it is also available for the 1063c33e97efSMauro Carvalho Chehab kernel's use for its own data structures. Among many 1064c33e97efSMauro Carvalho Chehab other things, it is where everything from the Slab is 1065c33e97efSMauro Carvalho Chehab allocated. Bad things happen when you're out of lowmem. 1066c33e97efSMauro Carvalho ChehabSwapTotal 1067c33e97efSMauro Carvalho Chehab total amount of swap space available 1068c33e97efSMauro Carvalho ChehabSwapFree 1069c33e97efSMauro Carvalho Chehab Memory which has been evicted from RAM, and is temporarily 1070c33e97efSMauro Carvalho Chehab on the disk 1071f6498b77SJohannes WeinerZswap 1072f6498b77SJohannes Weiner Memory consumed by the zswap backend (compressed size) 1073f6498b77SJohannes WeinerZswapped 1074f6498b77SJohannes Weiner Amount of anonymous memory stored in zswap (original size) 1075c33e97efSMauro Carvalho ChehabDirty 1076c33e97efSMauro Carvalho Chehab Memory which is waiting to get written back to the disk 1077c33e97efSMauro Carvalho ChehabWriteback 1078c33e97efSMauro Carvalho Chehab Memory which is actively being written back to the disk 1079c33e97efSMauro Carvalho ChehabAnonPages 1080c33e97efSMauro Carvalho Chehab Non-file backed pages mapped into userspace page tables 1081c33e97efSMauro Carvalho ChehabMapped 1082c33e97efSMauro Carvalho Chehab files which have been mmaped, such as libraries 1083c33e97efSMauro Carvalho ChehabShmem 1084c33e97efSMauro Carvalho Chehab Total memory used by shared memory (shmem) and tmpfs 1085c33e97efSMauro Carvalho ChehabKReclaimable 1086c33e97efSMauro Carvalho Chehab Kernel allocations that the kernel will attempt to reclaim 1087c33e97efSMauro Carvalho Chehab under memory pressure. Includes SReclaimable (below), and other 1088c33e97efSMauro Carvalho Chehab direct allocations with a shrinker. 1089c33e97efSMauro Carvalho ChehabSlab 1090c33e97efSMauro Carvalho Chehab in-kernel data structures cache 1091c33e97efSMauro Carvalho ChehabSReclaimable 1092c33e97efSMauro Carvalho Chehab Part of Slab, that might be reclaimed, such as caches 1093c33e97efSMauro Carvalho ChehabSUnreclaim 1094c33e97efSMauro Carvalho Chehab Part of Slab, that cannot be reclaimed on memory pressure 109539799b64SJohannes WeinerKernelStack 109639799b64SJohannes Weiner Memory consumed by the kernel stacks of all tasks 1097c33e97efSMauro Carvalho ChehabPageTables 109839799b64SJohannes Weiner Memory consumed by userspace page tables 1099ebc97a52SYosry AhmedSecPageTables 1100ebc97a52SYosry Ahmed Memory consumed by secondary page tables, this currently 1101ebc97a52SYosry Ahmed currently includes KVM mmu allocations on x86 and arm64. 1102c33e97efSMauro Carvalho ChehabNFS_Unstable 11038d92890bSNeilBrown Always zero. Previous counted pages which had been written to 11048d92890bSNeilBrown the server, but has not been committed to stable storage. 1105c33e97efSMauro Carvalho ChehabBounce 1106c33e97efSMauro Carvalho Chehab Memory used for block device "bounce buffers" 1107c33e97efSMauro Carvalho ChehabWritebackTmp 1108c33e97efSMauro Carvalho Chehab Memory used by FUSE for temporary writeback buffers 1109c33e97efSMauro Carvalho ChehabCommitLimit 1110c33e97efSMauro Carvalho Chehab Based on the overcommit ratio ('vm.overcommit_ratio'), 1111c33e97efSMauro Carvalho Chehab this is the total amount of memory currently available to 1112c33e97efSMauro Carvalho Chehab be allocated on the system. This limit is only adhered to 1113c33e97efSMauro Carvalho Chehab if strict overcommit accounting is enabled (mode 2 in 1114c33e97efSMauro Carvalho Chehab 'vm.overcommit_memory'). 1115c33e97efSMauro Carvalho Chehab 1116c33e97efSMauro Carvalho Chehab The CommitLimit is calculated with the following formula:: 1117c33e97efSMauro Carvalho Chehab 1118c33e97efSMauro Carvalho Chehab CommitLimit = ([total RAM pages] - [total huge TLB pages]) * 1119c33e97efSMauro Carvalho Chehab overcommit_ratio / 100 + [total swap pages] 1120c33e97efSMauro Carvalho Chehab 1121c33e97efSMauro Carvalho Chehab For example, on a system with 1G of physical RAM and 7G 1122c33e97efSMauro Carvalho Chehab of swap with a `vm.overcommit_ratio` of 30 it would 1123c33e97efSMauro Carvalho Chehab yield a CommitLimit of 7.3G. 1124c33e97efSMauro Carvalho Chehab 1125c33e97efSMauro Carvalho Chehab For more details, see the memory overcommit documentation 1126ee65728eSMike Rapoport in mm/overcommit-accounting. 1127c33e97efSMauro Carvalho ChehabCommitted_AS 1128c33e97efSMauro Carvalho Chehab The amount of memory presently allocated on the system. 1129c33e97efSMauro Carvalho Chehab The committed memory is a sum of all of the memory which 1130c33e97efSMauro Carvalho Chehab has been allocated by processes, even if it has not been 1131c33e97efSMauro Carvalho Chehab "used" by them as of yet. A process which malloc()'s 1G 1132c33e97efSMauro Carvalho Chehab of memory, but only touches 300M of it will show up as 1133c33e97efSMauro Carvalho Chehab using 1G. This 1G is memory which has been "committed" to 1134c33e97efSMauro Carvalho Chehab by the VM and can be used at any time by the allocating 1135c33e97efSMauro Carvalho Chehab application. With strict overcommit enabled on the system 1136c33e97efSMauro Carvalho Chehab (mode 2 in 'vm.overcommit_memory'), allocations which would 1137c33e97efSMauro Carvalho Chehab exceed the CommitLimit (detailed above) will not be permitted. 1138c33e97efSMauro Carvalho Chehab This is useful if one needs to guarantee that processes will 1139c33e97efSMauro Carvalho Chehab not fail due to lack of memory once that memory has been 1140c33e97efSMauro Carvalho Chehab successfully allocated. 1141c33e97efSMauro Carvalho ChehabVmallocTotal 114239799b64SJohannes Weiner total size of vmalloc virtual address space 1143c33e97efSMauro Carvalho ChehabVmallocUsed 1144c33e97efSMauro Carvalho Chehab amount of vmalloc area which is used 1145c33e97efSMauro Carvalho ChehabVmallocChunk 1146c33e97efSMauro Carvalho Chehab largest contiguous block of vmalloc area which is free 1147c33e97efSMauro Carvalho ChehabPercpu 1148c33e97efSMauro Carvalho Chehab Memory allocated to the percpu allocator used to back percpu 1149c33e97efSMauro Carvalho Chehab allocations. This stat excludes the cost of metadata. 1150*bd23024bSTomas MudrunkaEarlyMemtestBad 1151*bd23024bSTomas Mudrunka The amount of RAM/memory in kB, that was identified as corrupted 1152*bd23024bSTomas Mudrunka by early memtest. If memtest was not run, this field will not 1153*bd23024bSTomas Mudrunka be displayed at all. Size is never rounded down to 0 kB. 1154*bd23024bSTomas Mudrunka That means if 0 kB is reported, you can safely assume 1155*bd23024bSTomas Mudrunka there was at least one pass of memtest and none of the passes 1156*bd23024bSTomas Mudrunka found a single faulty byte of RAM. 115739799b64SJohannes WeinerHardwareCorrupted 115839799b64SJohannes Weiner The amount of RAM/memory in KB, the kernel identifies as 115939799b64SJohannes Weiner corrupted. 116039799b64SJohannes WeinerAnonHugePages 116139799b64SJohannes Weiner Non-file backed huge pages mapped into userspace page tables 116239799b64SJohannes WeinerShmemHugePages 116339799b64SJohannes Weiner Memory used by shared memory (shmem) and tmpfs allocated 116439799b64SJohannes Weiner with huge pages 116539799b64SJohannes WeinerShmemPmdMapped 116639799b64SJohannes Weiner Shared memory mapped into userspace with huge pages 116739799b64SJohannes WeinerFileHugePages 116839799b64SJohannes Weiner Memory used for filesystem data (page cache) allocated 116939799b64SJohannes Weiner with huge pages 117039799b64SJohannes WeinerFilePmdMapped 117139799b64SJohannes Weiner Page cache mapped into userspace with huge pages 117239799b64SJohannes WeinerCmaTotal 117339799b64SJohannes Weiner Memory reserved for the Contiguous Memory Allocator (CMA) 117439799b64SJohannes WeinerCmaFree 117539799b64SJohannes Weiner Free remaining memory in the CMA reserves 117639799b64SJohannes WeinerHugePages_Total, HugePages_Free, HugePages_Rsvd, HugePages_Surp, Hugepagesize, Hugetlb 117739799b64SJohannes Weiner See Documentation/admin-guide/mm/hugetlbpage.rst. 117839799b64SJohannes WeinerDirectMap4k, DirectMap2M, DirectMap1G 117939799b64SJohannes Weiner Breakdown of page table sizes used in the kernel's 118039799b64SJohannes Weiner identity mapping of RAM 1181c33e97efSMauro Carvalho Chehab 1182c33e97efSMauro Carvalho Chehabvmallocinfo 1183c33e97efSMauro Carvalho Chehab~~~~~~~~~~~ 1184c33e97efSMauro Carvalho Chehab 1185c33e97efSMauro Carvalho ChehabProvides information about vmalloced/vmaped areas. One line per area, 1186c33e97efSMauro Carvalho Chehabcontaining the virtual address range of the area, size in bytes, 1187c33e97efSMauro Carvalho Chehabcaller information of the creator, and optional information depending 1188c33e97efSMauro Carvalho Chehabon the kind of area: 1189c33e97efSMauro Carvalho Chehab 1190c33e97efSMauro Carvalho Chehab ========== =================================================== 1191c33e97efSMauro Carvalho Chehab pages=nr number of pages 1192c33e97efSMauro Carvalho Chehab phys=addr if a physical address was specified 1193c33e97efSMauro Carvalho Chehab ioremap I/O mapping (ioremap() and friends) 1194c33e97efSMauro Carvalho Chehab vmalloc vmalloc() area 1195c33e97efSMauro Carvalho Chehab vmap vmap()ed pages 1196c33e97efSMauro Carvalho Chehab user VM_USERMAP area 1197c33e97efSMauro Carvalho Chehab vpages buffer for pages pointers was vmalloced (huge area) 1198c33e97efSMauro Carvalho Chehab N<node>=nr (Only on NUMA kernels) 1199c33e97efSMauro Carvalho Chehab Number of pages allocated on memory node <node> 1200c33e97efSMauro Carvalho Chehab ========== =================================================== 1201c33e97efSMauro Carvalho Chehab 1202c33e97efSMauro Carvalho Chehab:: 1203c33e97efSMauro Carvalho Chehab 1204c33e97efSMauro Carvalho Chehab > cat /proc/vmallocinfo 1205c33e97efSMauro Carvalho Chehab 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... 1206c33e97efSMauro Carvalho Chehab /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 1207c33e97efSMauro Carvalho Chehab 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... 1208c33e97efSMauro Carvalho Chehab /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 1209c33e97efSMauro Carvalho Chehab 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... 1210c33e97efSMauro Carvalho Chehab phys=7fee8000 ioremap 1211c33e97efSMauro Carvalho Chehab 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... 1212c33e97efSMauro Carvalho Chehab phys=7fee7000 ioremap 1213c33e97efSMauro Carvalho Chehab 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 1214c33e97efSMauro Carvalho Chehab 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... 1215c33e97efSMauro Carvalho Chehab /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 1216c33e97efSMauro Carvalho Chehab 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... 1217c33e97efSMauro Carvalho Chehab pages=2 vmalloc N1=2 1218c33e97efSMauro Carvalho Chehab 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... 1219c33e97efSMauro Carvalho Chehab /0x130 [x_tables] pages=4 vmalloc N0=4 1220c33e97efSMauro Carvalho Chehab 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... 1221c33e97efSMauro Carvalho Chehab pages=14 vmalloc N2=14 1222c33e97efSMauro Carvalho Chehab 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... 1223c33e97efSMauro Carvalho Chehab pages=4 vmalloc N1=4 1224c33e97efSMauro Carvalho Chehab 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... 1225c33e97efSMauro Carvalho Chehab pages=2 vmalloc N1=2 1226c33e97efSMauro Carvalho Chehab 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... 1227c33e97efSMauro Carvalho Chehab pages=10 vmalloc N0=10 1228c33e97efSMauro Carvalho Chehab 1229c33e97efSMauro Carvalho Chehab 1230c33e97efSMauro Carvalho Chehabsoftirqs 1231c33e97efSMauro Carvalho Chehab~~~~~~~~ 1232c33e97efSMauro Carvalho Chehab 1233059db434SRandy DunlapProvides counts of softirq handlers serviced since boot time, for each CPU. 1234c33e97efSMauro Carvalho Chehab 1235c33e97efSMauro Carvalho Chehab:: 1236c33e97efSMauro Carvalho Chehab 1237c33e97efSMauro Carvalho Chehab > cat /proc/softirqs 1238c33e97efSMauro Carvalho Chehab CPU0 CPU1 CPU2 CPU3 1239c33e97efSMauro Carvalho Chehab HI: 0 0 0 0 1240c33e97efSMauro Carvalho Chehab TIMER: 27166 27120 27097 27034 1241c33e97efSMauro Carvalho Chehab NET_TX: 0 0 0 17 1242c33e97efSMauro Carvalho Chehab NET_RX: 42 0 0 39 1243c33e97efSMauro Carvalho Chehab BLOCK: 0 0 107 1121 1244c33e97efSMauro Carvalho Chehab TASKLET: 0 0 0 290 1245c33e97efSMauro Carvalho Chehab SCHED: 27035 26983 26971 26746 1246c33e97efSMauro Carvalho Chehab HRTIMER: 0 0 0 0 1247c33e97efSMauro Carvalho Chehab RCU: 1678 1769 2178 2250 1248c33e97efSMauro Carvalho Chehab 1249e24ccaafSPaul Gortmaker1.3 Networking info in /proc/net 1250c33e97efSMauro Carvalho Chehab-------------------------------- 1251c33e97efSMauro Carvalho Chehab 1252c33e97efSMauro Carvalho ChehabThe subdirectory /proc/net follows the usual pattern. Table 1-8 shows the 1253c33e97efSMauro Carvalho Chehabadditional values you get for IP version 6 if you configure the kernel to 1254c33e97efSMauro Carvalho Chehabsupport this. Table 1-9 lists the files and their meaning. 1255c33e97efSMauro Carvalho Chehab 1256c33e97efSMauro Carvalho Chehab 1257c33e97efSMauro Carvalho Chehab.. table:: Table 1-8: IPv6 info in /proc/net 1258c33e97efSMauro Carvalho Chehab 1259c33e97efSMauro Carvalho Chehab ========== ===================================================== 1260c33e97efSMauro Carvalho Chehab File Content 1261c33e97efSMauro Carvalho Chehab ========== ===================================================== 1262c33e97efSMauro Carvalho Chehab udp6 UDP sockets (IPv6) 1263c33e97efSMauro Carvalho Chehab tcp6 TCP sockets (IPv6) 1264c33e97efSMauro Carvalho Chehab raw6 Raw device statistics (IPv6) 1265c33e97efSMauro Carvalho Chehab igmp6 IP multicast addresses, which this host joined (IPv6) 1266c33e97efSMauro Carvalho Chehab if_inet6 List of IPv6 interface addresses 1267c33e97efSMauro Carvalho Chehab ipv6_route Kernel routing table for IPv6 1268c33e97efSMauro Carvalho Chehab rt6_stats Global IPv6 routing tables statistics 1269c33e97efSMauro Carvalho Chehab sockstat6 Socket statistics (IPv6) 1270c33e97efSMauro Carvalho Chehab snmp6 Snmp data (IPv6) 1271c33e97efSMauro Carvalho Chehab ========== ===================================================== 1272c33e97efSMauro Carvalho Chehab 1273c33e97efSMauro Carvalho Chehab.. table:: Table 1-9: Network info in /proc/net 1274c33e97efSMauro Carvalho Chehab 1275c33e97efSMauro Carvalho Chehab ============= ================================================================ 1276c33e97efSMauro Carvalho Chehab File Content 1277c33e97efSMauro Carvalho Chehab ============= ================================================================ 1278c33e97efSMauro Carvalho Chehab arp Kernel ARP table 1279c33e97efSMauro Carvalho Chehab dev network devices with statistics 1280c33e97efSMauro Carvalho Chehab dev_mcast the Layer2 multicast groups a device is listening too 1281c33e97efSMauro Carvalho Chehab (interface index, label, number of references, number of bound 1282c33e97efSMauro Carvalho Chehab addresses). 1283c33e97efSMauro Carvalho Chehab dev_stat network device status 1284c33e97efSMauro Carvalho Chehab ip_fwchains Firewall chain linkage 1285c33e97efSMauro Carvalho Chehab ip_fwnames Firewall chain names 1286c33e97efSMauro Carvalho Chehab ip_masq Directory containing the masquerading tables 1287c33e97efSMauro Carvalho Chehab ip_masquerade Major masquerading table 1288c33e97efSMauro Carvalho Chehab netstat Network statistics 1289c33e97efSMauro Carvalho Chehab raw raw device statistics 1290c33e97efSMauro Carvalho Chehab route Kernel routing table 1291c33e97efSMauro Carvalho Chehab rpc Directory containing rpc info 1292c33e97efSMauro Carvalho Chehab rt_cache Routing cache 1293c33e97efSMauro Carvalho Chehab snmp SNMP data 1294c33e97efSMauro Carvalho Chehab sockstat Socket statistics 12958b0a211dSYang Yang softnet_stat Per-CPU incoming packets queues statistics of online CPUs 1296c33e97efSMauro Carvalho Chehab tcp TCP sockets 1297c33e97efSMauro Carvalho Chehab udp UDP sockets 1298c33e97efSMauro Carvalho Chehab unix UNIX domain sockets 1299c33e97efSMauro Carvalho Chehab wireless Wireless interface data (Wavelan etc) 1300c33e97efSMauro Carvalho Chehab igmp IP multicast addresses, which this host joined 1301c33e97efSMauro Carvalho Chehab psched Global packet scheduler parameters. 1302c33e97efSMauro Carvalho Chehab netlink List of PF_NETLINK sockets 1303c33e97efSMauro Carvalho Chehab ip_mr_vifs List of multicast virtual interfaces 1304c33e97efSMauro Carvalho Chehab ip_mr_cache List of multicast routing cache 1305c33e97efSMauro Carvalho Chehab ============= ================================================================ 1306c33e97efSMauro Carvalho Chehab 1307c33e97efSMauro Carvalho ChehabYou can use this information to see which network devices are available in 1308c33e97efSMauro Carvalho Chehabyour system and how much traffic was routed over those devices:: 1309c33e97efSMauro Carvalho Chehab 1310c33e97efSMauro Carvalho Chehab > cat /proc/net/dev 1311c33e97efSMauro Carvalho Chehab Inter-|Receive |[... 1312c33e97efSMauro Carvalho Chehab face |bytes packets errs drop fifo frame compressed multicast|[... 1313c33e97efSMauro Carvalho Chehab lo: 908188 5596 0 0 0 0 0 0 [... 1314c33e97efSMauro Carvalho Chehab ppp0:15475140 20721 410 0 0 410 0 0 [... 1315c33e97efSMauro Carvalho Chehab eth0: 614530 7085 0 0 0 0 0 1 [... 1316c33e97efSMauro Carvalho Chehab 1317c33e97efSMauro Carvalho Chehab ...] Transmit 1318c33e97efSMauro Carvalho Chehab ...] bytes packets errs drop fifo colls carrier compressed 1319c33e97efSMauro Carvalho Chehab ...] 908188 5596 0 0 0 0 0 0 1320c33e97efSMauro Carvalho Chehab ...] 1375103 17405 0 0 0 0 0 0 1321c33e97efSMauro Carvalho Chehab ...] 1703981 5535 0 0 0 3 0 0 1322c33e97efSMauro Carvalho Chehab 1323c33e97efSMauro Carvalho ChehabIn addition, each Channel Bond interface has its own directory. For 1324c33e97efSMauro Carvalho Chehabexample, the bond0 device will have a directory called /proc/net/bond0/. 1325c33e97efSMauro Carvalho ChehabIt will contain information that is specific to that bond, such as the 1326c33e97efSMauro Carvalho Chehabcurrent slaves of the bond, the link status of the slaves, and how 1327c33e97efSMauro Carvalho Chehabmany times the slaves link has failed. 1328c33e97efSMauro Carvalho Chehab 1329e24ccaafSPaul Gortmaker1.4 SCSI info 1330c33e97efSMauro Carvalho Chehab------------- 1331c33e97efSMauro Carvalho Chehab 1332c33e97efSMauro Carvalho ChehabIf you have a SCSI host adapter in your system, you'll find a subdirectory 1333c33e97efSMauro Carvalho Chehabnamed after the driver for this adapter in /proc/scsi. You'll also see a list 1334c33e97efSMauro Carvalho Chehabof all recognized SCSI devices in /proc/scsi:: 1335c33e97efSMauro Carvalho Chehab 1336c33e97efSMauro Carvalho Chehab >cat /proc/scsi/scsi 1337c33e97efSMauro Carvalho Chehab Attached devices: 1338c33e97efSMauro Carvalho Chehab Host: scsi0 Channel: 00 Id: 00 Lun: 00 1339c33e97efSMauro Carvalho Chehab Vendor: IBM Model: DGHS09U Rev: 03E0 1340c33e97efSMauro Carvalho Chehab Type: Direct-Access ANSI SCSI revision: 03 1341c33e97efSMauro Carvalho Chehab Host: scsi0 Channel: 00 Id: 06 Lun: 00 1342c33e97efSMauro Carvalho Chehab Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04 1343c33e97efSMauro Carvalho Chehab Type: CD-ROM ANSI SCSI revision: 02 1344c33e97efSMauro Carvalho Chehab 1345c33e97efSMauro Carvalho Chehab 1346c33e97efSMauro Carvalho ChehabThe directory named after the driver has one file for each adapter found in 1347c33e97efSMauro Carvalho Chehabthe system. These files contain information about the controller, including 1348c33e97efSMauro Carvalho Chehabthe used IRQ and the IO address range. The amount of information shown is 1349c33e97efSMauro Carvalho Chehabdependent on the adapter you use. The example shows the output for an Adaptec 1350c33e97efSMauro Carvalho ChehabAHA-2940 SCSI adapter:: 1351c33e97efSMauro Carvalho Chehab 1352c33e97efSMauro Carvalho Chehab > cat /proc/scsi/aic7xxx/0 1353c33e97efSMauro Carvalho Chehab 1354c33e97efSMauro Carvalho Chehab Adaptec AIC7xxx driver version: 5.1.19/3.2.4 1355c33e97efSMauro Carvalho Chehab Compile Options: 1356c33e97efSMauro Carvalho Chehab TCQ Enabled By Default : Disabled 1357c33e97efSMauro Carvalho Chehab AIC7XXX_PROC_STATS : Disabled 1358c33e97efSMauro Carvalho Chehab AIC7XXX_RESET_DELAY : 5 1359c33e97efSMauro Carvalho Chehab Adapter Configuration: 1360c33e97efSMauro Carvalho Chehab SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter 1361c33e97efSMauro Carvalho Chehab Ultra Wide Controller 1362c33e97efSMauro Carvalho Chehab PCI MMAPed I/O Base: 0xeb001000 1363c33e97efSMauro Carvalho Chehab Adapter SEEPROM Config: SEEPROM found and used. 1364c33e97efSMauro Carvalho Chehab Adaptec SCSI BIOS: Enabled 1365c33e97efSMauro Carvalho Chehab IRQ: 10 1366c33e97efSMauro Carvalho Chehab SCBs: Active 0, Max Active 2, 1367c33e97efSMauro Carvalho Chehab Allocated 15, HW 16, Page 255 1368c33e97efSMauro Carvalho Chehab Interrupts: 160328 1369c33e97efSMauro Carvalho Chehab BIOS Control Word: 0x18b6 1370c33e97efSMauro Carvalho Chehab Adapter Control Word: 0x005b 1371c33e97efSMauro Carvalho Chehab Extended Translation: Enabled 1372c33e97efSMauro Carvalho Chehab Disconnect Enable Flags: 0xffff 1373c33e97efSMauro Carvalho Chehab Ultra Enable Flags: 0x0001 1374c33e97efSMauro Carvalho Chehab Tag Queue Enable Flags: 0x0000 1375c33e97efSMauro Carvalho Chehab Ordered Queue Tag Flags: 0x0000 1376c33e97efSMauro Carvalho Chehab Default Tag Queue Depth: 8 1377c33e97efSMauro Carvalho Chehab Tagged Queue By Device array for aic7xxx host instance 0: 1378c33e97efSMauro Carvalho Chehab {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255} 1379c33e97efSMauro Carvalho Chehab Actual queue depth per device for aic7xxx host instance 0: 1380c33e97efSMauro Carvalho Chehab {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1} 1381c33e97efSMauro Carvalho Chehab Statistics: 1382c33e97efSMauro Carvalho Chehab (scsi0:0:0:0) 1383c33e97efSMauro Carvalho Chehab Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8 1384c33e97efSMauro Carvalho Chehab Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0) 1385c33e97efSMauro Carvalho Chehab Total transfers 160151 (74577 reads and 85574 writes) 1386c33e97efSMauro Carvalho Chehab (scsi0:0:6:0) 1387c33e97efSMauro Carvalho Chehab Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15 1388c33e97efSMauro Carvalho Chehab Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0) 1389c33e97efSMauro Carvalho Chehab Total transfers 0 (0 reads and 0 writes) 1390c33e97efSMauro Carvalho Chehab 1391c33e97efSMauro Carvalho Chehab 1392e24ccaafSPaul Gortmaker1.5 Parallel port info in /proc/parport 1393c33e97efSMauro Carvalho Chehab--------------------------------------- 1394c33e97efSMauro Carvalho Chehab 1395c33e97efSMauro Carvalho ChehabThe directory /proc/parport contains information about the parallel ports of 1396c33e97efSMauro Carvalho Chehabyour system. It has one subdirectory for each port, named after the port 1397c33e97efSMauro Carvalho Chehabnumber (0,1,2,...). 1398c33e97efSMauro Carvalho Chehab 1399c33e97efSMauro Carvalho ChehabThese directories contain the four files shown in Table 1-10. 1400c33e97efSMauro Carvalho Chehab 1401c33e97efSMauro Carvalho Chehab 1402c33e97efSMauro Carvalho Chehab.. table:: Table 1-10: Files in /proc/parport 1403c33e97efSMauro Carvalho Chehab 1404c33e97efSMauro Carvalho Chehab ========= ==================================================================== 1405c33e97efSMauro Carvalho Chehab File Content 1406c33e97efSMauro Carvalho Chehab ========= ==================================================================== 1407c33e97efSMauro Carvalho Chehab autoprobe Any IEEE-1284 device ID information that has been acquired. 1408c33e97efSMauro Carvalho Chehab devices list of the device drivers using that port. A + will appear by the 1409c33e97efSMauro Carvalho Chehab name of the device currently using the port (it might not appear 1410c33e97efSMauro Carvalho Chehab against any). 1411c33e97efSMauro Carvalho Chehab hardware Parallel port's base address, IRQ line and DMA channel. 1412c33e97efSMauro Carvalho Chehab irq IRQ that parport is using for that port. This is in a separate 1413c33e97efSMauro Carvalho Chehab file to allow you to alter it by writing a new value in (IRQ 1414c33e97efSMauro Carvalho Chehab number or none). 1415c33e97efSMauro Carvalho Chehab ========= ==================================================================== 1416c33e97efSMauro Carvalho Chehab 1417e24ccaafSPaul Gortmaker1.6 TTY info in /proc/tty 1418c33e97efSMauro Carvalho Chehab------------------------- 1419c33e97efSMauro Carvalho Chehab 1420c33e97efSMauro Carvalho ChehabInformation about the available and actually used tty's can be found in the 1421c33e97efSMauro Carvalho Chehabdirectory /proc/tty. You'll find entries for drivers and line disciplines in 1422c33e97efSMauro Carvalho Chehabthis directory, as shown in Table 1-11. 1423c33e97efSMauro Carvalho Chehab 1424c33e97efSMauro Carvalho Chehab 1425c33e97efSMauro Carvalho Chehab.. table:: Table 1-11: Files in /proc/tty 1426c33e97efSMauro Carvalho Chehab 1427c33e97efSMauro Carvalho Chehab ============= ============================================== 1428c33e97efSMauro Carvalho Chehab File Content 1429c33e97efSMauro Carvalho Chehab ============= ============================================== 1430c33e97efSMauro Carvalho Chehab drivers list of drivers and their usage 1431c33e97efSMauro Carvalho Chehab ldiscs registered line disciplines 1432c33e97efSMauro Carvalho Chehab driver/serial usage statistic and status of single tty lines 1433c33e97efSMauro Carvalho Chehab ============= ============================================== 1434c33e97efSMauro Carvalho Chehab 1435c33e97efSMauro Carvalho ChehabTo see which tty's are currently in use, you can simply look into the file 1436c33e97efSMauro Carvalho Chehab/proc/tty/drivers:: 1437c33e97efSMauro Carvalho Chehab 1438c33e97efSMauro Carvalho Chehab > cat /proc/tty/drivers 1439c33e97efSMauro Carvalho Chehab pty_slave /dev/pts 136 0-255 pty:slave 1440c33e97efSMauro Carvalho Chehab pty_master /dev/ptm 128 0-255 pty:master 1441c33e97efSMauro Carvalho Chehab pty_slave /dev/ttyp 3 0-255 pty:slave 1442c33e97efSMauro Carvalho Chehab pty_master /dev/pty 2 0-255 pty:master 1443c33e97efSMauro Carvalho Chehab serial /dev/cua 5 64-67 serial:callout 1444c33e97efSMauro Carvalho Chehab serial /dev/ttyS 4 64-67 serial 1445c33e97efSMauro Carvalho Chehab /dev/tty0 /dev/tty0 4 0 system:vtmaster 1446c33e97efSMauro Carvalho Chehab /dev/ptmx /dev/ptmx 5 2 system 1447c33e97efSMauro Carvalho Chehab /dev/console /dev/console 5 1 system:console 1448c33e97efSMauro Carvalho Chehab /dev/tty /dev/tty 5 0 system:/dev/tty 1449c33e97efSMauro Carvalho Chehab unknown /dev/tty 4 1-63 console 1450c33e97efSMauro Carvalho Chehab 1451c33e97efSMauro Carvalho Chehab 1452e24ccaafSPaul Gortmaker1.7 Miscellaneous kernel statistics in /proc/stat 1453c33e97efSMauro Carvalho Chehab------------------------------------------------- 1454c33e97efSMauro Carvalho Chehab 1455c33e97efSMauro Carvalho ChehabVarious pieces of information about kernel activity are available in the 1456c33e97efSMauro Carvalho Chehab/proc/stat file. All of the numbers reported in this file are aggregates 1457c33e97efSMauro Carvalho Chehabsince the system first booted. For a quick look, simply cat the file:: 1458c33e97efSMauro Carvalho Chehab 1459c33e97efSMauro Carvalho Chehab > cat /proc/stat 1460c33e97efSMauro Carvalho Chehab cpu 2255 34 2290 22625563 6290 127 456 0 0 0 1461c33e97efSMauro Carvalho Chehab cpu0 1132 34 1441 11311718 3675 127 438 0 0 0 1462c33e97efSMauro Carvalho Chehab cpu1 1123 0 849 11313845 2614 0 18 0 0 0 1463c33e97efSMauro Carvalho Chehab intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...] 1464c33e97efSMauro Carvalho Chehab ctxt 1990473 1465c33e97efSMauro Carvalho Chehab btime 1062191376 1466c33e97efSMauro Carvalho Chehab processes 2915 1467c33e97efSMauro Carvalho Chehab procs_running 1 1468c33e97efSMauro Carvalho Chehab procs_blocked 0 1469c33e97efSMauro Carvalho Chehab softirq 183433 0 21755 12 39 1137 231 21459 2263 1470c33e97efSMauro Carvalho Chehab 1471c33e97efSMauro Carvalho ChehabThe very first "cpu" line aggregates the numbers in all of the other "cpuN" 1472c33e97efSMauro Carvalho Chehablines. These numbers identify the amount of time the CPU has spent performing 1473c33e97efSMauro Carvalho Chehabdifferent kinds of work. Time units are in USER_HZ (typically hundredths of a 1474c33e97efSMauro Carvalho Chehabsecond). The meanings of the columns are as follows, from left to right: 1475c33e97efSMauro Carvalho Chehab 1476c33e97efSMauro Carvalho Chehab- user: normal processes executing in user mode 1477c33e97efSMauro Carvalho Chehab- nice: niced processes executing in user mode 1478c33e97efSMauro Carvalho Chehab- system: processes executing in kernel mode 1479c33e97efSMauro Carvalho Chehab- idle: twiddling thumbs 1480c33e97efSMauro Carvalho Chehab- iowait: In a word, iowait stands for waiting for I/O to complete. But there 1481c33e97efSMauro Carvalho Chehab are several problems: 1482c33e97efSMauro Carvalho Chehab 1483059db434SRandy Dunlap 1. CPU will not wait for I/O to complete, iowait is the time that a task is 1484059db434SRandy Dunlap waiting for I/O to complete. When CPU goes into idle state for 1485059db434SRandy Dunlap outstanding task I/O, another task will be scheduled on this CPU. 1486c33e97efSMauro Carvalho Chehab 2. In a multi-core CPU, the task waiting for I/O to complete is not running 1487c33e97efSMauro Carvalho Chehab on any CPU, so the iowait of each CPU is difficult to calculate. 1488c33e97efSMauro Carvalho Chehab 3. The value of iowait field in /proc/stat will decrease in certain 1489c33e97efSMauro Carvalho Chehab conditions. 1490c33e97efSMauro Carvalho Chehab 1491c33e97efSMauro Carvalho Chehab So, the iowait is not reliable by reading from /proc/stat. 1492c33e97efSMauro Carvalho Chehab- irq: servicing interrupts 1493c33e97efSMauro Carvalho Chehab- softirq: servicing softirqs 1494c33e97efSMauro Carvalho Chehab- steal: involuntary wait 1495c33e97efSMauro Carvalho Chehab- guest: running a normal guest 1496c33e97efSMauro Carvalho Chehab- guest_nice: running a niced guest 1497c33e97efSMauro Carvalho Chehab 1498c33e97efSMauro Carvalho ChehabThe "intr" line gives counts of interrupts serviced since boot time, for each 1499c33e97efSMauro Carvalho Chehabof the possible system interrupts. The first column is the total of all 1500c33e97efSMauro Carvalho Chehabinterrupts serviced including unnumbered architecture specific interrupts; 1501c33e97efSMauro Carvalho Chehabeach subsequent column is the total for that particular numbered interrupt. 1502c33e97efSMauro Carvalho ChehabUnnumbered interrupts are not shown, only summed into the total. 1503c33e97efSMauro Carvalho Chehab 1504c33e97efSMauro Carvalho ChehabThe "ctxt" line gives the total number of context switches across all CPUs. 1505c33e97efSMauro Carvalho Chehab 1506c33e97efSMauro Carvalho ChehabThe "btime" line gives the time at which the system booted, in seconds since 1507c33e97efSMauro Carvalho Chehabthe Unix epoch. 1508c33e97efSMauro Carvalho Chehab 1509c33e97efSMauro Carvalho ChehabThe "processes" line gives the number of processes and threads created, which 1510c33e97efSMauro Carvalho Chehabincludes (but is not limited to) those created by calls to the fork() and 1511c33e97efSMauro Carvalho Chehabclone() system calls. 1512c33e97efSMauro Carvalho Chehab 1513c33e97efSMauro Carvalho ChehabThe "procs_running" line gives the total number of threads that are 1514c33e97efSMauro Carvalho Chehabrunning or ready to run (i.e., the total number of runnable threads). 1515c33e97efSMauro Carvalho Chehab 1516c33e97efSMauro Carvalho ChehabThe "procs_blocked" line gives the number of processes currently blocked, 1517c33e97efSMauro Carvalho Chehabwaiting for I/O to complete. 1518c33e97efSMauro Carvalho Chehab 1519c33e97efSMauro Carvalho ChehabThe "softirq" line gives counts of softirqs serviced since boot time, for each 1520c33e97efSMauro Carvalho Chehabof the possible system softirqs. The first column is the total of all 1521c33e97efSMauro Carvalho Chehabsoftirqs serviced; each subsequent column is the total for that particular 1522c33e97efSMauro Carvalho Chehabsoftirq. 1523c33e97efSMauro Carvalho Chehab 1524c33e97efSMauro Carvalho Chehab 1525e24ccaafSPaul Gortmaker1.8 Ext4 file system parameters 1526c33e97efSMauro Carvalho Chehab------------------------------- 1527c33e97efSMauro Carvalho Chehab 1528c33e97efSMauro Carvalho ChehabInformation about mounted ext4 file systems can be found in 1529c33e97efSMauro Carvalho Chehab/proc/fs/ext4. Each mounted filesystem will have a directory in 1530c33e97efSMauro Carvalho Chehab/proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or 1531c33e97efSMauro Carvalho Chehab/proc/fs/ext4/dm-0). The files in each per-device directory are shown 1532c33e97efSMauro Carvalho Chehabin Table 1-12, below. 1533c33e97efSMauro Carvalho Chehab 1534c33e97efSMauro Carvalho Chehab.. table:: Table 1-12: Files in /proc/fs/ext4/<devname> 1535c33e97efSMauro Carvalho Chehab 1536c33e97efSMauro Carvalho Chehab ============== ========================================================== 1537c33e97efSMauro Carvalho Chehab File Content 1538c33e97efSMauro Carvalho Chehab mb_groups details of multiblock allocator buddy cache of free blocks 1539c33e97efSMauro Carvalho Chehab ============== ========================================================== 1540c33e97efSMauro Carvalho Chehab 1541e24ccaafSPaul Gortmaker1.9 /proc/consoles 1542059db434SRandy Dunlap------------------- 1543c33e97efSMauro Carvalho ChehabShows registered system console lines. 1544c33e97efSMauro Carvalho Chehab 1545c33e97efSMauro Carvalho ChehabTo see which character device lines are currently used for the system console 1546c33e97efSMauro Carvalho Chehab/dev/console, you may simply look into the file /proc/consoles:: 1547c33e97efSMauro Carvalho Chehab 1548c33e97efSMauro Carvalho Chehab > cat /proc/consoles 1549c33e97efSMauro Carvalho Chehab tty0 -WU (ECp) 4:7 1550c33e97efSMauro Carvalho Chehab ttyS0 -W- (Ep) 4:64 1551c33e97efSMauro Carvalho Chehab 1552c33e97efSMauro Carvalho ChehabThe columns are: 1553c33e97efSMauro Carvalho Chehab 1554c33e97efSMauro Carvalho Chehab+--------------------+-------------------------------------------------------+ 1555c33e97efSMauro Carvalho Chehab| device | name of the device | 1556c33e97efSMauro Carvalho Chehab+====================+=======================================================+ 1557c33e97efSMauro Carvalho Chehab| operations | * R = can do read operations | 1558c33e97efSMauro Carvalho Chehab| | * W = can do write operations | 1559c33e97efSMauro Carvalho Chehab| | * U = can do unblank | 1560c33e97efSMauro Carvalho Chehab+--------------------+-------------------------------------------------------+ 1561c33e97efSMauro Carvalho Chehab| flags | * E = it is enabled | 1562c33e97efSMauro Carvalho Chehab| | * C = it is preferred console | 1563c33e97efSMauro Carvalho Chehab| | * B = it is primary boot console | 1564c33e97efSMauro Carvalho Chehab| | * p = it is used for printk buffer | 1565c33e97efSMauro Carvalho Chehab| | * b = it is not a TTY but a Braille device | 1566c33e97efSMauro Carvalho Chehab| | * a = it is safe to use when cpu is offline | 1567c33e97efSMauro Carvalho Chehab+--------------------+-------------------------------------------------------+ 1568c33e97efSMauro Carvalho Chehab| major:minor | major and minor number of the device separated by a | 1569c33e97efSMauro Carvalho Chehab| | colon | 1570c33e97efSMauro Carvalho Chehab+--------------------+-------------------------------------------------------+ 1571c33e97efSMauro Carvalho Chehab 1572c33e97efSMauro Carvalho ChehabSummary 1573c33e97efSMauro Carvalho Chehab------- 1574c33e97efSMauro Carvalho Chehab 1575c33e97efSMauro Carvalho ChehabThe /proc file system serves information about the running system. It not only 1576c33e97efSMauro Carvalho Chehaballows access to process data but also allows you to request the kernel status 1577c33e97efSMauro Carvalho Chehabby reading files in the hierarchy. 1578c33e97efSMauro Carvalho Chehab 1579c33e97efSMauro Carvalho ChehabThe directory structure of /proc reflects the types of information and makes 1580c33e97efSMauro Carvalho Chehabit easy, if not obvious, where to look for specific data. 1581c33e97efSMauro Carvalho Chehab 1582c33e97efSMauro Carvalho ChehabChapter 2: Modifying System Parameters 1583c33e97efSMauro Carvalho Chehab====================================== 1584c33e97efSMauro Carvalho Chehab 1585c33e97efSMauro Carvalho ChehabIn This Chapter 1586c33e97efSMauro Carvalho Chehab--------------- 1587c33e97efSMauro Carvalho Chehab 1588c33e97efSMauro Carvalho Chehab* Modifying kernel parameters by writing into files found in /proc/sys 1589c33e97efSMauro Carvalho Chehab* Exploring the files which modify certain parameters 1590c33e97efSMauro Carvalho Chehab* Review of the /proc/sys file tree 1591c33e97efSMauro Carvalho Chehab 1592c33e97efSMauro Carvalho Chehab------------------------------------------------------------------------------ 1593c33e97efSMauro Carvalho Chehab 1594c33e97efSMauro Carvalho ChehabA very interesting part of /proc is the directory /proc/sys. This is not only 1595c33e97efSMauro Carvalho Chehaba source of information, it also allows you to change parameters within the 1596c33e97efSMauro Carvalho Chehabkernel. Be very careful when attempting this. You can optimize your system, 1597c33e97efSMauro Carvalho Chehabbut you can also cause it to crash. Never alter kernel parameters on a 1598c33e97efSMauro Carvalho Chehabproduction system. Set up a development machine and test to make sure that 1599c33e97efSMauro Carvalho Chehabeverything works the way you want it to. You may have no alternative but to 1600c33e97efSMauro Carvalho Chehabreboot the machine once an error has been made. 1601c33e97efSMauro Carvalho Chehab 1602059db434SRandy DunlapTo change a value, simply echo the new value into the file. 1603059db434SRandy DunlapYou need to be root to do this. You can create your own boot script 1604059db434SRandy Dunlapto perform this every time your system boots. 1605c33e97efSMauro Carvalho Chehab 1606c33e97efSMauro Carvalho ChehabThe files in /proc/sys can be used to fine tune and monitor miscellaneous and 1607c33e97efSMauro Carvalho Chehabgeneral things in the operation of the Linux kernel. Since some of the files 1608c33e97efSMauro Carvalho Chehabcan inadvertently disrupt your system, it is advisable to read both 1609c33e97efSMauro Carvalho Chehabdocumentation and source before actually making adjustments. In any case, be 1610c33e97efSMauro Carvalho Chehabvery careful when writing to any of these files. The entries in /proc may 1611c33e97efSMauro Carvalho Chehabchange slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt 1612c33e97efSMauro Carvalho Chehabreview the kernel documentation in the directory /usr/src/linux/Documentation. 1613c33e97efSMauro Carvalho ChehabThis chapter is heavily based on the documentation included in the pre 2.2 1614c33e97efSMauro Carvalho Chehabkernels, and became part of it in version 2.2.1 of the Linux kernel. 1615c33e97efSMauro Carvalho Chehab 1616c33e97efSMauro Carvalho ChehabPlease see: Documentation/admin-guide/sysctl/ directory for descriptions of these 1617c33e97efSMauro Carvalho Chehabentries. 1618c33e97efSMauro Carvalho Chehab 1619c33e97efSMauro Carvalho ChehabSummary 1620c33e97efSMauro Carvalho Chehab------- 1621c33e97efSMauro Carvalho Chehab 1622c33e97efSMauro Carvalho ChehabCertain aspects of kernel behavior can be modified at runtime, without the 1623c33e97efSMauro Carvalho Chehabneed to recompile the kernel, or even to reboot the system. The files in the 1624c33e97efSMauro Carvalho Chehab/proc/sys tree can not only be read, but also modified. You can use the echo 1625c33e97efSMauro Carvalho Chehabcommand to write value into these files, thereby changing the default settings 1626c33e97efSMauro Carvalho Chehabof the kernel. 1627c33e97efSMauro Carvalho Chehab 1628c33e97efSMauro Carvalho Chehab 1629c33e97efSMauro Carvalho ChehabChapter 3: Per-process Parameters 1630c33e97efSMauro Carvalho Chehab================================= 1631c33e97efSMauro Carvalho Chehab 1632c33e97efSMauro Carvalho Chehab3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score 1633c33e97efSMauro Carvalho Chehab-------------------------------------------------------------------------------- 1634c33e97efSMauro Carvalho Chehab 1635059db434SRandy DunlapThese files can be used to adjust the badness heuristic used to select which 1636059db434SRandy Dunlapprocess gets killed in out of memory (oom) conditions. 1637c33e97efSMauro Carvalho Chehab 1638c33e97efSMauro Carvalho ChehabThe badness heuristic assigns a value to each candidate task ranging from 0 1639c33e97efSMauro Carvalho Chehab(never kill) to 1000 (always kill) to determine which process is targeted. The 1640c33e97efSMauro Carvalho Chehabunits are roughly a proportion along that range of allowed memory the process 1641c33e97efSMauro Carvalho Chehabmay allocate from based on an estimation of its current memory and swap use. 1642c33e97efSMauro Carvalho ChehabFor example, if a task is using all allowed memory, its badness score will be 1643c33e97efSMauro Carvalho Chehab1000. If it is using half of its allowed memory, its score will be 500. 1644c33e97efSMauro Carvalho Chehab 1645c33e97efSMauro Carvalho ChehabThe amount of "allowed" memory depends on the context in which the oom killer 1646c33e97efSMauro Carvalho Chehabwas called. If it is due to the memory assigned to the allocating task's cpuset 1647c33e97efSMauro Carvalho Chehabbeing exhausted, the allowed memory represents the set of mems assigned to that 1648c33e97efSMauro Carvalho Chehabcpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed 1649c33e97efSMauro Carvalho Chehabmemory represents the set of mempolicy nodes. If it is due to a memory 1650c33e97efSMauro Carvalho Chehablimit (or swap limit) being reached, the allowed memory is that configured 1651c33e97efSMauro Carvalho Chehablimit. Finally, if it is due to the entire system being out of memory, the 1652c33e97efSMauro Carvalho Chehaballowed memory represents all allocatable resources. 1653c33e97efSMauro Carvalho Chehab 1654c33e97efSMauro Carvalho ChehabThe value of /proc/<pid>/oom_score_adj is added to the badness score before it 1655c33e97efSMauro Carvalho Chehabis used to determine which task to kill. Acceptable values range from -1000 1656c33e97efSMauro Carvalho Chehab(OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to 1657c33e97efSMauro Carvalho Chehabpolarize the preference for oom killing either by always preferring a certain 1658c33e97efSMauro Carvalho Chehabtask or completely disabling it. The lowest possible value, -1000, is 1659c33e97efSMauro Carvalho Chehabequivalent to disabling oom killing entirely for that task since it will always 1660c33e97efSMauro Carvalho Chehabreport a badness score of 0. 1661c33e97efSMauro Carvalho Chehab 1662c33e97efSMauro Carvalho ChehabConsequently, it is very simple for userspace to define the amount of memory to 1663c33e97efSMauro Carvalho Chehabconsider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for 1664c33e97efSMauro Carvalho Chehabexample, is roughly equivalent to allowing the remainder of tasks sharing the 1665c33e97efSMauro Carvalho Chehabsame system, cpuset, mempolicy, or memory controller resources to use at least 1666c33e97efSMauro Carvalho Chehab50% more memory. A value of -500, on the other hand, would be roughly 1667c33e97efSMauro Carvalho Chehabequivalent to discounting 50% of the task's allowed memory from being considered 1668c33e97efSMauro Carvalho Chehabas scoring against the task. 1669c33e97efSMauro Carvalho Chehab 1670c33e97efSMauro Carvalho ChehabFor backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also 1671c33e97efSMauro Carvalho Chehabbe used to tune the badness score. Its acceptable values range from -16 1672c33e97efSMauro Carvalho Chehab(OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17 1673c33e97efSMauro Carvalho Chehab(OOM_DISABLE) to disable oom killing entirely for that task. Its value is 1674c33e97efSMauro Carvalho Chehabscaled linearly with /proc/<pid>/oom_score_adj. 1675c33e97efSMauro Carvalho Chehab 1676c33e97efSMauro Carvalho ChehabThe value of /proc/<pid>/oom_score_adj may be reduced no lower than the last 1677c33e97efSMauro Carvalho Chehabvalue set by a CAP_SYS_RESOURCE process. To reduce the value any lower 1678c33e97efSMauro Carvalho Chehabrequires CAP_SYS_RESOURCE. 1679c33e97efSMauro Carvalho Chehab 1680c33e97efSMauro Carvalho Chehab 1681c33e97efSMauro Carvalho Chehab3.2 /proc/<pid>/oom_score - Display current oom-killer score 1682c33e97efSMauro Carvalho Chehab------------------------------------------------------------- 1683c33e97efSMauro Carvalho Chehab 1684059db434SRandy DunlapThis file can be used to check the current score used by the oom-killer for 1685c33e97efSMauro Carvalho Chehabany given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which 1686c33e97efSMauro Carvalho Chehabprocess should be killed in an out-of-memory situation. 1687c33e97efSMauro Carvalho Chehab 1688b1aa7c93SMichal HockoPlease note that the exported value includes oom_score_adj so it is 1689b1aa7c93SMichal Hockoeffectively in range [0,2000]. 1690b1aa7c93SMichal Hocko 1691c33e97efSMauro Carvalho Chehab 1692c33e97efSMauro Carvalho Chehab3.3 /proc/<pid>/io - Display the IO accounting fields 1693c33e97efSMauro Carvalho Chehab------------------------------------------------------- 1694c33e97efSMauro Carvalho Chehab 1695059db434SRandy DunlapThis file contains IO statistics for each running process. 1696c33e97efSMauro Carvalho Chehab 1697c33e97efSMauro Carvalho ChehabExample 1698c33e97efSMauro Carvalho Chehab~~~~~~~ 1699c33e97efSMauro Carvalho Chehab 1700c33e97efSMauro Carvalho Chehab:: 1701c33e97efSMauro Carvalho Chehab 1702c33e97efSMauro Carvalho Chehab test:/tmp # dd if=/dev/zero of=/tmp/test.dat & 1703c33e97efSMauro Carvalho Chehab [1] 3828 1704c33e97efSMauro Carvalho Chehab 1705c33e97efSMauro Carvalho Chehab test:/tmp # cat /proc/3828/io 1706c33e97efSMauro Carvalho Chehab rchar: 323934931 1707c33e97efSMauro Carvalho Chehab wchar: 323929600 1708c33e97efSMauro Carvalho Chehab syscr: 632687 1709c33e97efSMauro Carvalho Chehab syscw: 632675 1710c33e97efSMauro Carvalho Chehab read_bytes: 0 1711c33e97efSMauro Carvalho Chehab write_bytes: 323932160 1712c33e97efSMauro Carvalho Chehab cancelled_write_bytes: 0 1713c33e97efSMauro Carvalho Chehab 1714c33e97efSMauro Carvalho Chehab 1715c33e97efSMauro Carvalho ChehabDescription 1716c33e97efSMauro Carvalho Chehab~~~~~~~~~~~ 1717c33e97efSMauro Carvalho Chehab 1718c33e97efSMauro Carvalho Chehabrchar 1719c33e97efSMauro Carvalho Chehab^^^^^ 1720c33e97efSMauro Carvalho Chehab 1721c33e97efSMauro Carvalho ChehabI/O counter: chars read 1722c33e97efSMauro Carvalho ChehabThe number of bytes which this task has caused to be read from storage. This 1723c33e97efSMauro Carvalho Chehabis simply the sum of bytes which this process passed to read() and pread(). 1724c33e97efSMauro Carvalho ChehabIt includes things like tty IO and it is unaffected by whether or not actual 1725c33e97efSMauro Carvalho Chehabphysical disk IO was required (the read might have been satisfied from 1726059db434SRandy Dunlappagecache). 1727c33e97efSMauro Carvalho Chehab 1728c33e97efSMauro Carvalho Chehab 1729c33e97efSMauro Carvalho Chehabwchar 1730c33e97efSMauro Carvalho Chehab^^^^^ 1731c33e97efSMauro Carvalho Chehab 1732c33e97efSMauro Carvalho ChehabI/O counter: chars written 1733c33e97efSMauro Carvalho ChehabThe number of bytes which this task has caused, or shall cause to be written 1734c33e97efSMauro Carvalho Chehabto disk. Similar caveats apply here as with rchar. 1735c33e97efSMauro Carvalho Chehab 1736c33e97efSMauro Carvalho Chehab 1737c33e97efSMauro Carvalho Chehabsyscr 1738c33e97efSMauro Carvalho Chehab^^^^^ 1739c33e97efSMauro Carvalho Chehab 1740c33e97efSMauro Carvalho ChehabI/O counter: read syscalls 1741c33e97efSMauro Carvalho ChehabAttempt to count the number of read I/O operations, i.e. syscalls like read() 1742c33e97efSMauro Carvalho Chehaband pread(). 1743c33e97efSMauro Carvalho Chehab 1744c33e97efSMauro Carvalho Chehab 1745c33e97efSMauro Carvalho Chehabsyscw 1746c33e97efSMauro Carvalho Chehab^^^^^ 1747c33e97efSMauro Carvalho Chehab 1748c33e97efSMauro Carvalho ChehabI/O counter: write syscalls 1749c33e97efSMauro Carvalho ChehabAttempt to count the number of write I/O operations, i.e. syscalls like 1750c33e97efSMauro Carvalho Chehabwrite() and pwrite(). 1751c33e97efSMauro Carvalho Chehab 1752c33e97efSMauro Carvalho Chehab 1753c33e97efSMauro Carvalho Chehabread_bytes 1754c33e97efSMauro Carvalho Chehab^^^^^^^^^^ 1755c33e97efSMauro Carvalho Chehab 1756c33e97efSMauro Carvalho ChehabI/O counter: bytes read 1757c33e97efSMauro Carvalho ChehabAttempt to count the number of bytes which this process really did cause to 1758c33e97efSMauro Carvalho Chehabbe fetched from the storage layer. Done at the submit_bio() level, so it is 1759c33e97efSMauro Carvalho Chehabaccurate for block-backed filesystems. <please add status regarding NFS and 1760c33e97efSMauro Carvalho ChehabCIFS at a later time> 1761c33e97efSMauro Carvalho Chehab 1762c33e97efSMauro Carvalho Chehab 1763c33e97efSMauro Carvalho Chehabwrite_bytes 1764c33e97efSMauro Carvalho Chehab^^^^^^^^^^^ 1765c33e97efSMauro Carvalho Chehab 1766c33e97efSMauro Carvalho ChehabI/O counter: bytes written 1767c33e97efSMauro Carvalho ChehabAttempt to count the number of bytes which this process caused to be sent to 1768c33e97efSMauro Carvalho Chehabthe storage layer. This is done at page-dirtying time. 1769c33e97efSMauro Carvalho Chehab 1770c33e97efSMauro Carvalho Chehab 1771c33e97efSMauro Carvalho Chehabcancelled_write_bytes 1772c33e97efSMauro Carvalho Chehab^^^^^^^^^^^^^^^^^^^^^ 1773c33e97efSMauro Carvalho Chehab 1774c33e97efSMauro Carvalho ChehabThe big inaccuracy here is truncate. If a process writes 1MB to a file and 1775c33e97efSMauro Carvalho Chehabthen deletes the file, it will in fact perform no writeout. But it will have 1776c33e97efSMauro Carvalho Chehabbeen accounted as having caused 1MB of write. 1777c33e97efSMauro Carvalho ChehabIn other words: The number of bytes which this process caused to not happen, 1778c33e97efSMauro Carvalho Chehabby truncating pagecache. A task can cause "negative" IO too. If this task 1779c33e97efSMauro Carvalho Chehabtruncates some dirty pagecache, some IO which another task has been accounted 1780c33e97efSMauro Carvalho Chehabfor (in its write_bytes) will not be happening. We _could_ just subtract that 1781c33e97efSMauro Carvalho Chehabfrom the truncating task's write_bytes, but there is information loss in doing 1782c33e97efSMauro Carvalho Chehabthat. 1783c33e97efSMauro Carvalho Chehab 1784c33e97efSMauro Carvalho Chehab 1785c33e97efSMauro Carvalho Chehab.. Note:: 1786c33e97efSMauro Carvalho Chehab 1787c33e97efSMauro Carvalho Chehab At its current implementation state, this is a bit racy on 32-bit machines: 1788c33e97efSMauro Carvalho Chehab if process A reads process B's /proc/pid/io while process B is updating one 1789c33e97efSMauro Carvalho Chehab of those 64-bit counters, process A could see an intermediate result. 1790c33e97efSMauro Carvalho Chehab 1791c33e97efSMauro Carvalho Chehab 1792c33e97efSMauro Carvalho ChehabMore information about this can be found within the taskstats documentation in 1793c33e97efSMauro Carvalho ChehabDocumentation/accounting. 1794c33e97efSMauro Carvalho Chehab 1795c33e97efSMauro Carvalho Chehab3.4 /proc/<pid>/coredump_filter - Core dump filtering settings 1796c33e97efSMauro Carvalho Chehab--------------------------------------------------------------- 1797c33e97efSMauro Carvalho ChehabWhen a process is dumped, all anonymous memory is written to a core file as 1798c33e97efSMauro Carvalho Chehablong as the size of the core file isn't limited. But sometimes we don't want 1799c33e97efSMauro Carvalho Chehabto dump some memory segments, for example, huge shared memory or DAX. 1800c33e97efSMauro Carvalho ChehabConversely, sometimes we want to save file-backed memory segments into a core 1801c33e97efSMauro Carvalho Chehabfile, not only the individual files. 1802c33e97efSMauro Carvalho Chehab 1803c33e97efSMauro Carvalho Chehab/proc/<pid>/coredump_filter allows you to customize which memory segments 1804c33e97efSMauro Carvalho Chehabwill be dumped when the <pid> process is dumped. coredump_filter is a bitmask 1805c33e97efSMauro Carvalho Chehabof memory types. If a bit of the bitmask is set, memory segments of the 1806c33e97efSMauro Carvalho Chehabcorresponding memory type are dumped, otherwise they are not dumped. 1807c33e97efSMauro Carvalho Chehab 1808c33e97efSMauro Carvalho ChehabThe following 9 memory types are supported: 1809c33e97efSMauro Carvalho Chehab 1810c33e97efSMauro Carvalho Chehab - (bit 0) anonymous private memory 1811c33e97efSMauro Carvalho Chehab - (bit 1) anonymous shared memory 1812c33e97efSMauro Carvalho Chehab - (bit 2) file-backed private memory 1813c33e97efSMauro Carvalho Chehab - (bit 3) file-backed shared memory 1814c33e97efSMauro Carvalho Chehab - (bit 4) ELF header pages in file-backed private memory areas (it is 1815c33e97efSMauro Carvalho Chehab effective only if the bit 2 is cleared) 1816c33e97efSMauro Carvalho Chehab - (bit 5) hugetlb private memory 1817c33e97efSMauro Carvalho Chehab - (bit 6) hugetlb shared memory 1818c33e97efSMauro Carvalho Chehab - (bit 7) DAX private memory 1819c33e97efSMauro Carvalho Chehab - (bit 8) DAX shared memory 1820c33e97efSMauro Carvalho Chehab 1821c33e97efSMauro Carvalho Chehab Note that MMIO pages such as frame buffer are never dumped and vDSO pages 1822c33e97efSMauro Carvalho Chehab are always dumped regardless of the bitmask status. 1823c33e97efSMauro Carvalho Chehab 1824c33e97efSMauro Carvalho Chehab Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is 1825c33e97efSMauro Carvalho Chehab only affected by bit 5-6, and DAX is only affected by bits 7-8. 1826c33e97efSMauro Carvalho Chehab 1827c33e97efSMauro Carvalho ChehabThe default value of coredump_filter is 0x33; this means all anonymous memory 1828c33e97efSMauro Carvalho Chehabsegments, ELF header pages and hugetlb private memory are dumped. 1829c33e97efSMauro Carvalho Chehab 1830c33e97efSMauro Carvalho ChehabIf you don't want to dump all shared memory segments attached to pid 1234, 1831c33e97efSMauro Carvalho Chehabwrite 0x31 to the process's proc file:: 1832c33e97efSMauro Carvalho Chehab 1833c33e97efSMauro Carvalho Chehab $ echo 0x31 > /proc/1234/coredump_filter 1834c33e97efSMauro Carvalho Chehab 1835c33e97efSMauro Carvalho ChehabWhen a new process is created, the process inherits the bitmask status from its 1836c33e97efSMauro Carvalho Chehabparent. It is useful to set up coredump_filter before the program runs. 1837c33e97efSMauro Carvalho ChehabFor example:: 1838c33e97efSMauro Carvalho Chehab 1839c33e97efSMauro Carvalho Chehab $ echo 0x7 > /proc/self/coredump_filter 1840c33e97efSMauro Carvalho Chehab $ ./some_program 1841c33e97efSMauro Carvalho Chehab 1842c33e97efSMauro Carvalho Chehab3.5 /proc/<pid>/mountinfo - Information about mounts 1843c33e97efSMauro Carvalho Chehab-------------------------------------------------------- 1844c33e97efSMauro Carvalho Chehab 1845c33e97efSMauro Carvalho ChehabThis file contains lines of the form:: 1846c33e97efSMauro Carvalho Chehab 1847c33e97efSMauro Carvalho Chehab 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue 1848ff9c3d43SChristoph Anton Mitterer (1)(2)(3) (4) (5) (6) (n…m) (m+1)(m+2) (m+3) (m+4) 1849c33e97efSMauro Carvalho Chehab 1850c33e97efSMauro Carvalho Chehab (1) mount ID: unique identifier of the mount (may be reused after umount) 1851c33e97efSMauro Carvalho Chehab (2) parent ID: ID of parent (or of self for the top of the mount tree) 1852c33e97efSMauro Carvalho Chehab (3) major:minor: value of st_dev for files on filesystem 1853c33e97efSMauro Carvalho Chehab (4) root: root of the mount within the filesystem 1854c33e97efSMauro Carvalho Chehab (5) mount point: mount point relative to the process's root 1855c33e97efSMauro Carvalho Chehab (6) mount options: per mount options 1856ff9c3d43SChristoph Anton Mitterer (n…m) optional fields: zero or more fields of the form "tag[:value]" 1857ff9c3d43SChristoph Anton Mitterer (m+1) separator: marks the end of the optional fields 1858ff9c3d43SChristoph Anton Mitterer (m+2) filesystem type: name of filesystem of the form "type[.subtype]" 1859ff9c3d43SChristoph Anton Mitterer (m+3) mount source: filesystem specific information or "none" 1860ff9c3d43SChristoph Anton Mitterer (m+4) super options: per super block options 1861c33e97efSMauro Carvalho Chehab 1862c33e97efSMauro Carvalho ChehabParsers should ignore all unrecognised optional fields. Currently the 1863c33e97efSMauro Carvalho Chehabpossible optional fields are: 1864c33e97efSMauro Carvalho Chehab 1865c33e97efSMauro Carvalho Chehab================ ============================================================== 1866c33e97efSMauro Carvalho Chehabshared:X mount is shared in peer group X 1867c33e97efSMauro Carvalho Chehabmaster:X mount is slave to peer group X 1868c33e97efSMauro Carvalho Chehabpropagate_from:X mount is slave and receives propagation from peer group X [#]_ 1869c33e97efSMauro Carvalho Chehabunbindable mount is unbindable 1870c33e97efSMauro Carvalho Chehab================ ============================================================== 1871c33e97efSMauro Carvalho Chehab 1872c33e97efSMauro Carvalho Chehab.. [#] X is the closest dominant peer group under the process's root. If 1873c33e97efSMauro Carvalho Chehab X is the immediate master of the mount, or if there's no dominant peer 1874c33e97efSMauro Carvalho Chehab group under the same root, then only the "master:X" field is present 1875c33e97efSMauro Carvalho Chehab and not the "propagate_from:X" field. 1876c33e97efSMauro Carvalho Chehab 1877c33e97efSMauro Carvalho ChehabFor more information on mount propagation see: 1878c33e97efSMauro Carvalho Chehab 1879cf06612cSMauro Carvalho Chehab Documentation/filesystems/sharedsubtree.rst 1880c33e97efSMauro Carvalho Chehab 1881c33e97efSMauro Carvalho Chehab 1882c33e97efSMauro Carvalho Chehab3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm 1883c33e97efSMauro Carvalho Chehab-------------------------------------------------------- 1884059db434SRandy DunlapThese files provide a method to access a task's comm value. It also allows for 1885c33e97efSMauro Carvalho Chehaba task to set its own or one of its thread siblings comm value. The comm value 1886c33e97efSMauro Carvalho Chehabis limited in size compared to the cmdline value, so writing anything longer 1887c33e97efSMauro Carvalho Chehabthen the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated 1888c33e97efSMauro Carvalho Chehabcomm value. 1889c33e97efSMauro Carvalho Chehab 1890c33e97efSMauro Carvalho Chehab 1891c33e97efSMauro Carvalho Chehab3.7 /proc/<pid>/task/<tid>/children - Information about task children 1892c33e97efSMauro Carvalho Chehab------------------------------------------------------------------------- 1893c33e97efSMauro Carvalho ChehabThis file provides a fast way to retrieve first level children pids 1894c33e97efSMauro Carvalho Chehabof a task pointed by <pid>/<tid> pair. The format is a space separated 1895c33e97efSMauro Carvalho Chehabstream of pids. 1896c33e97efSMauro Carvalho Chehab 1897059db434SRandy DunlapNote the "first level" here -- if a child has its own children they will 1898059db434SRandy Dunlapnot be listed here; one needs to read /proc/<children-pid>/task/<tid>/children 1899c33e97efSMauro Carvalho Chehabto obtain the descendants. 1900c33e97efSMauro Carvalho Chehab 1901c33e97efSMauro Carvalho ChehabSince this interface is intended to be fast and cheap it doesn't 1902c33e97efSMauro Carvalho Chehabguarantee to provide precise results and some children might be 1903c33e97efSMauro Carvalho Chehabskipped, especially if they've exited right after we printed their 1904059db434SRandy Dunlappids, so one needs to either stop or freeze processes being inspected 1905c33e97efSMauro Carvalho Chehabif precise results are needed. 1906c33e97efSMauro Carvalho Chehab 1907c33e97efSMauro Carvalho Chehab 1908c33e97efSMauro Carvalho Chehab3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file 1909c33e97efSMauro Carvalho Chehab--------------------------------------------------------------- 1910c33e97efSMauro Carvalho ChehabThis file provides information associated with an opened file. The regular 19113845f256SKalesh Singhfiles have at least four fields -- 'pos', 'flags', 'mnt_id' and 'ino'. 19123845f256SKalesh SinghThe 'pos' represents the current offset of the opened file in decimal 19133845f256SKalesh Singhform [see lseek(2) for details], 'flags' denotes the octal O_xxx mask the 19143845f256SKalesh Singhfile has been created with [see open(2) for details] and 'mnt_id' represents 19153845f256SKalesh Singhmount ID of the file system containing the opened file [see 3.5 19163845f256SKalesh Singh/proc/<pid>/mountinfo for details]. 'ino' represents the inode number of 19173845f256SKalesh Singhthe file. 1918c33e97efSMauro Carvalho Chehab 1919c33e97efSMauro Carvalho ChehabA typical output is:: 1920c33e97efSMauro Carvalho Chehab 1921c33e97efSMauro Carvalho Chehab pos: 0 1922c33e97efSMauro Carvalho Chehab flags: 0100002 1923c33e97efSMauro Carvalho Chehab mnt_id: 19 19243845f256SKalesh Singh ino: 63107 1925c33e97efSMauro Carvalho Chehab 1926c33e97efSMauro Carvalho ChehabAll locks associated with a file descriptor are shown in its fdinfo too:: 1927c33e97efSMauro Carvalho Chehab 1928c33e97efSMauro Carvalho Chehab lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF 1929c33e97efSMauro Carvalho Chehab 1930c33e97efSMauro Carvalho ChehabThe files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags 1931c33e97efSMauro Carvalho Chehabpair provide additional information particular to the objects they represent. 1932c33e97efSMauro Carvalho Chehab 1933c33e97efSMauro Carvalho ChehabEventfd files 1934c33e97efSMauro Carvalho Chehab~~~~~~~~~~~~~ 1935c33e97efSMauro Carvalho Chehab 1936c33e97efSMauro Carvalho Chehab:: 1937c33e97efSMauro Carvalho Chehab 1938c33e97efSMauro Carvalho Chehab pos: 0 1939c33e97efSMauro Carvalho Chehab flags: 04002 1940c33e97efSMauro Carvalho Chehab mnt_id: 9 19413845f256SKalesh Singh ino: 63107 1942c33e97efSMauro Carvalho Chehab eventfd-count: 5a 1943c33e97efSMauro Carvalho Chehab 1944c33e97efSMauro Carvalho Chehabwhere 'eventfd-count' is hex value of a counter. 1945c33e97efSMauro Carvalho Chehab 1946c33e97efSMauro Carvalho ChehabSignalfd files 1947c33e97efSMauro Carvalho Chehab~~~~~~~~~~~~~~ 1948c33e97efSMauro Carvalho Chehab 1949c33e97efSMauro Carvalho Chehab:: 1950c33e97efSMauro Carvalho Chehab 1951c33e97efSMauro Carvalho Chehab pos: 0 1952c33e97efSMauro Carvalho Chehab flags: 04002 1953c33e97efSMauro Carvalho Chehab mnt_id: 9 19543845f256SKalesh Singh ino: 63107 1955c33e97efSMauro Carvalho Chehab sigmask: 0000000000000200 1956c33e97efSMauro Carvalho Chehab 1957c33e97efSMauro Carvalho Chehabwhere 'sigmask' is hex value of the signal mask associated 1958c33e97efSMauro Carvalho Chehabwith a file. 1959c33e97efSMauro Carvalho Chehab 1960c33e97efSMauro Carvalho ChehabEpoll files 1961c33e97efSMauro Carvalho Chehab~~~~~~~~~~~ 1962c33e97efSMauro Carvalho Chehab 1963c33e97efSMauro Carvalho Chehab:: 1964c33e97efSMauro Carvalho Chehab 1965c33e97efSMauro Carvalho Chehab pos: 0 1966c33e97efSMauro Carvalho Chehab flags: 02 1967c33e97efSMauro Carvalho Chehab mnt_id: 9 19683845f256SKalesh Singh ino: 63107 1969c33e97efSMauro Carvalho Chehab tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 1970c33e97efSMauro Carvalho Chehab 1971c33e97efSMauro Carvalho Chehabwhere 'tfd' is a target file descriptor number in decimal form, 1972c33e97efSMauro Carvalho Chehab'events' is events mask being watched and the 'data' is data 1973c33e97efSMauro Carvalho Chehabassociated with a target [see epoll(7) for more details]. 1974c33e97efSMauro Carvalho Chehab 1975c33e97efSMauro Carvalho ChehabThe 'pos' is current offset of the target file in decimal form 1976c33e97efSMauro Carvalho Chehab[see lseek(2)], 'ino' and 'sdev' are inode and device numbers 1977c33e97efSMauro Carvalho Chehabwhere target file resides, all in hex format. 1978c33e97efSMauro Carvalho Chehab 1979c33e97efSMauro Carvalho ChehabFsnotify files 1980c33e97efSMauro Carvalho Chehab~~~~~~~~~~~~~~ 1981c33e97efSMauro Carvalho ChehabFor inotify files the format is the following:: 1982c33e97efSMauro Carvalho Chehab 1983c33e97efSMauro Carvalho Chehab pos: 0 1984c33e97efSMauro Carvalho Chehab flags: 02000000 19853845f256SKalesh Singh mnt_id: 9 19863845f256SKalesh Singh ino: 63107 1987c33e97efSMauro Carvalho Chehab inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d 1988c33e97efSMauro Carvalho Chehab 1989059db434SRandy Dunlapwhere 'wd' is a watch descriptor in decimal form, i.e. a target file 1990c33e97efSMauro Carvalho Chehabdescriptor number, 'ino' and 'sdev' are inode and device where the 1991c33e97efSMauro Carvalho Chehabtarget file resides and the 'mask' is the mask of events, all in hex 1992c33e97efSMauro Carvalho Chehabform [see inotify(7) for more details]. 1993c33e97efSMauro Carvalho Chehab 1994c33e97efSMauro Carvalho ChehabIf the kernel was built with exportfs support, the path to the target 1995c33e97efSMauro Carvalho Chehabfile is encoded as a file handle. The file handle is provided by three 1996c33e97efSMauro Carvalho Chehabfields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex 1997c33e97efSMauro Carvalho Chehabformat. 1998c33e97efSMauro Carvalho Chehab 1999c33e97efSMauro Carvalho ChehabIf the kernel is built without exportfs support the file handle won't be 2000c33e97efSMauro Carvalho Chehabprinted out. 2001c33e97efSMauro Carvalho Chehab 2002c33e97efSMauro Carvalho ChehabIf there is no inotify mark attached yet the 'inotify' line will be omitted. 2003c33e97efSMauro Carvalho Chehab 2004c33e97efSMauro Carvalho ChehabFor fanotify files the format is:: 2005c33e97efSMauro Carvalho Chehab 2006c33e97efSMauro Carvalho Chehab pos: 0 2007c33e97efSMauro Carvalho Chehab flags: 02 2008c33e97efSMauro Carvalho Chehab mnt_id: 9 20093845f256SKalesh Singh ino: 63107 2010c33e97efSMauro Carvalho Chehab fanotify flags:10 event-flags:0 2011c33e97efSMauro Carvalho Chehab fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 2012c33e97efSMauro Carvalho Chehab fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 2013c33e97efSMauro Carvalho Chehab 2014c33e97efSMauro Carvalho Chehabwhere fanotify 'flags' and 'event-flags' are values used in fanotify_init 2015c33e97efSMauro Carvalho Chehabcall, 'mnt_id' is the mount point identifier, 'mflags' is the value of 2016c33e97efSMauro Carvalho Chehabflags associated with mark which are tracked separately from events 2017059db434SRandy Dunlapmask. 'ino' and 'sdev' are target inode and device, 'mask' is the events 2018c33e97efSMauro Carvalho Chehabmask and 'ignored_mask' is the mask of events which are to be ignored. 2019059db434SRandy DunlapAll are in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask' 2020059db434SRandy Dunlapprovide information about flags and mask used in fanotify_mark 2021c33e97efSMauro Carvalho Chehabcall [see fsnotify manpage for details]. 2022c33e97efSMauro Carvalho Chehab 2023c33e97efSMauro Carvalho ChehabWhile the first three lines are mandatory and always printed, the rest is 2024c33e97efSMauro Carvalho Chehaboptional and may be omitted if no marks created yet. 2025c33e97efSMauro Carvalho Chehab 2026c33e97efSMauro Carvalho ChehabTimerfd files 2027c33e97efSMauro Carvalho Chehab~~~~~~~~~~~~~ 2028c33e97efSMauro Carvalho Chehab 2029c33e97efSMauro Carvalho Chehab:: 2030c33e97efSMauro Carvalho Chehab 2031c33e97efSMauro Carvalho Chehab pos: 0 2032c33e97efSMauro Carvalho Chehab flags: 02 2033c33e97efSMauro Carvalho Chehab mnt_id: 9 20343845f256SKalesh Singh ino: 63107 2035c33e97efSMauro Carvalho Chehab clockid: 0 2036c33e97efSMauro Carvalho Chehab ticks: 0 2037c33e97efSMauro Carvalho Chehab settime flags: 01 2038c33e97efSMauro Carvalho Chehab it_value: (0, 49406829) 2039c33e97efSMauro Carvalho Chehab it_interval: (1, 0) 2040c33e97efSMauro Carvalho Chehab 2041c33e97efSMauro Carvalho Chehabwhere 'clockid' is the clock type and 'ticks' is the number of the timer expirations 2042c33e97efSMauro Carvalho Chehabthat have occurred [see timerfd_create(2) for details]. 'settime flags' are 2043c33e97efSMauro Carvalho Chehabflags in octal form been used to setup the timer [see timerfd_settime(2) for 2044059db434SRandy Dunlapdetails]. 'it_value' is remaining time until the timer expiration. 2045c33e97efSMauro Carvalho Chehab'it_interval' is the interval for the timer. Note the timer might be set up 2046c33e97efSMauro Carvalho Chehabwith TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value' 2047c33e97efSMauro Carvalho Chehabstill exhibits timer's remaining time. 2048c33e97efSMauro Carvalho Chehab 20493845f256SKalesh SinghDMA Buffer files 20503845f256SKalesh Singh~~~~~~~~~~~~~~~~ 20513845f256SKalesh Singh 20523845f256SKalesh Singh:: 20533845f256SKalesh Singh 20543845f256SKalesh Singh pos: 0 20553845f256SKalesh Singh flags: 04002 20563845f256SKalesh Singh mnt_id: 9 20573845f256SKalesh Singh ino: 63107 20583845f256SKalesh Singh size: 32768 20593845f256SKalesh Singh count: 2 20603845f256SKalesh Singh exp_name: system-heap 20613845f256SKalesh Singh 20623845f256SKalesh Singhwhere 'size' is the size of the DMA buffer in bytes. 'count' is the file count of 20633845f256SKalesh Singhthe DMA buffer file. 'exp_name' is the name of the DMA buffer exporter. 20643845f256SKalesh Singh 2065c33e97efSMauro Carvalho Chehab3.9 /proc/<pid>/map_files - Information about memory mapped files 2066c33e97efSMauro Carvalho Chehab--------------------------------------------------------------------- 2067c33e97efSMauro Carvalho ChehabThis directory contains symbolic links which represent memory mapped files 2068c33e97efSMauro Carvalho Chehabthe process is maintaining. Example output:: 2069c33e97efSMauro Carvalho Chehab 2070c33e97efSMauro Carvalho Chehab | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so 2071c33e97efSMauro Carvalho Chehab | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so 2072c33e97efSMauro Carvalho Chehab | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so 2073c33e97efSMauro Carvalho Chehab | ... 2074c33e97efSMauro Carvalho Chehab | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1 2075c33e97efSMauro Carvalho Chehab | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls 2076c33e97efSMauro Carvalho Chehab 2077c33e97efSMauro Carvalho ChehabThe name of a link represents the virtual memory bounds of a mapping, i.e. 2078c33e97efSMauro Carvalho Chehabvm_area_struct::vm_start-vm_area_struct::vm_end. 2079c33e97efSMauro Carvalho Chehab 2080c33e97efSMauro Carvalho ChehabThe main purpose of the map_files is to retrieve a set of memory mapped 2081c33e97efSMauro Carvalho Chehabfiles in a fast way instead of parsing /proc/<pid>/maps or 2082c33e97efSMauro Carvalho Chehab/proc/<pid>/smaps, both of which contain many more records. At the same 2083c33e97efSMauro Carvalho Chehabtime one can open(2) mappings from the listings of two processes and 2084c33e97efSMauro Carvalho Chehabcomparing their inode numbers to figure out which anonymous memory areas 2085c33e97efSMauro Carvalho Chehabare actually shared. 2086c33e97efSMauro Carvalho Chehab 2087c33e97efSMauro Carvalho Chehab3.10 /proc/<pid>/timerslack_ns - Task timerslack value 2088c33e97efSMauro Carvalho Chehab--------------------------------------------------------- 2089c33e97efSMauro Carvalho ChehabThis file provides the value of the task's timerslack value in nanoseconds. 2090059db434SRandy DunlapThis value specifies an amount of time that normal timers may be deferred 2091c33e97efSMauro Carvalho Chehabin order to coalesce timers and avoid unnecessary wakeups. 2092c33e97efSMauro Carvalho Chehab 2093c33e97efSMauro Carvalho ChehabThis allows a task's interactivity vs power consumption tradeoff to be 2094c33e97efSMauro Carvalho Chehabadjusted. 2095c33e97efSMauro Carvalho Chehab 2096059db434SRandy DunlapWriting 0 to the file will set the task's timerslack to the default value. 2097c33e97efSMauro Carvalho Chehab 2098c33e97efSMauro Carvalho ChehabValid values are from 0 - ULLONG_MAX 2099c33e97efSMauro Carvalho Chehab 2100c33e97efSMauro Carvalho ChehabAn application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level 2101c33e97efSMauro Carvalho Chehabpermissions on the task specified to change its timerslack_ns value. 2102c33e97efSMauro Carvalho Chehab 2103c33e97efSMauro Carvalho Chehab3.11 /proc/<pid>/patch_state - Livepatch patch operation state 2104c33e97efSMauro Carvalho Chehab----------------------------------------------------------------- 2105c33e97efSMauro Carvalho ChehabWhen CONFIG_LIVEPATCH is enabled, this file displays the value of the 2106c33e97efSMauro Carvalho Chehabpatch state for the task. 2107c33e97efSMauro Carvalho Chehab 2108c33e97efSMauro Carvalho ChehabA value of '-1' indicates that no patch is in transition. 2109c33e97efSMauro Carvalho Chehab 2110c33e97efSMauro Carvalho ChehabA value of '0' indicates that a patch is in transition and the task is 2111c33e97efSMauro Carvalho Chehabunpatched. If the patch is being enabled, then the task hasn't been 2112c33e97efSMauro Carvalho Chehabpatched yet. If the patch is being disabled, then the task has already 2113c33e97efSMauro Carvalho Chehabbeen unpatched. 2114c33e97efSMauro Carvalho Chehab 2115c33e97efSMauro Carvalho ChehabA value of '1' indicates that a patch is in transition and the task is 2116c33e97efSMauro Carvalho Chehabpatched. If the patch is being enabled, then the task has already been 2117c33e97efSMauro Carvalho Chehabpatched. If the patch is being disabled, then the task hasn't been 2118c33e97efSMauro Carvalho Chehabunpatched yet. 2119c33e97efSMauro Carvalho Chehab 2120c33e97efSMauro Carvalho Chehab3.12 /proc/<pid>/arch_status - task architecture specific status 2121c33e97efSMauro Carvalho Chehab------------------------------------------------------------------- 2122c33e97efSMauro Carvalho ChehabWhen CONFIG_PROC_PID_ARCH_STATUS is enabled, this file displays the 2123c33e97efSMauro Carvalho Chehabarchitecture specific status of the task. 2124c33e97efSMauro Carvalho Chehab 2125c33e97efSMauro Carvalho ChehabExample 2126c33e97efSMauro Carvalho Chehab~~~~~~~ 2127c33e97efSMauro Carvalho Chehab 2128c33e97efSMauro Carvalho Chehab:: 2129c33e97efSMauro Carvalho Chehab 2130c33e97efSMauro Carvalho Chehab $ cat /proc/6753/arch_status 2131c33e97efSMauro Carvalho Chehab AVX512_elapsed_ms: 8 2132c33e97efSMauro Carvalho Chehab 2133c33e97efSMauro Carvalho ChehabDescription 2134c33e97efSMauro Carvalho Chehab~~~~~~~~~~~ 2135c33e97efSMauro Carvalho Chehab 2136059db434SRandy Dunlapx86 specific entries 2137c33e97efSMauro Carvalho Chehab~~~~~~~~~~~~~~~~~~~~~ 2138c33e97efSMauro Carvalho Chehab 2139059db434SRandy DunlapAVX512_elapsed_ms 2140c33e97efSMauro Carvalho Chehab^^^^^^^^^^^^^^^^^^ 2141c33e97efSMauro Carvalho Chehab 2142c33e97efSMauro Carvalho Chehab If AVX512 is supported on the machine, this entry shows the milliseconds 2143c33e97efSMauro Carvalho Chehab elapsed since the last time AVX512 usage was recorded. The recording 2144c33e97efSMauro Carvalho Chehab happens on a best effort basis when a task is scheduled out. This means 2145c33e97efSMauro Carvalho Chehab that the value depends on two factors: 2146c33e97efSMauro Carvalho Chehab 2147c33e97efSMauro Carvalho Chehab 1) The time which the task spent on the CPU without being scheduled 2148c33e97efSMauro Carvalho Chehab out. With CPU isolation and a single runnable task this can take 2149c33e97efSMauro Carvalho Chehab several seconds. 2150c33e97efSMauro Carvalho Chehab 2151c33e97efSMauro Carvalho Chehab 2) The time since the task was scheduled out last. Depending on the 2152c33e97efSMauro Carvalho Chehab reason for being scheduled out (time slice exhausted, syscall ...) 2153c33e97efSMauro Carvalho Chehab this can be arbitrary long time. 2154c33e97efSMauro Carvalho Chehab 2155c33e97efSMauro Carvalho Chehab As a consequence the value cannot be considered precise and authoritative 2156c33e97efSMauro Carvalho Chehab information. The application which uses this information has to be aware 2157c33e97efSMauro Carvalho Chehab of the overall scenario on the system in order to determine whether a 2158c33e97efSMauro Carvalho Chehab task is a real AVX512 user or not. Precise information can be obtained 2159c33e97efSMauro Carvalho Chehab with performance counters. 2160c33e97efSMauro Carvalho Chehab 2161c33e97efSMauro Carvalho Chehab A special value of '-1' indicates that no AVX512 usage was recorded, thus 2162c33e97efSMauro Carvalho Chehab the task is unlikely an AVX512 user, but depends on the workload and the 2163c33e97efSMauro Carvalho Chehab scheduling scenario, it also could be a false negative mentioned above. 2164c33e97efSMauro Carvalho Chehab 2165f1f1f256SIvan Babrou3.13 /proc/<pid>/fd - List of symlinks to open files 2166f1f1f256SIvan Babrou------------------------------------------------------- 2167f1f1f256SIvan BabrouThis directory contains symbolic links which represent open files 2168f1f1f256SIvan Babrouthe process is maintaining. Example output:: 2169f1f1f256SIvan Babrou 2170f1f1f256SIvan Babrou lr-x------ 1 root root 64 Sep 20 17:53 0 -> /dev/null 2171f1f1f256SIvan Babrou l-wx------ 1 root root 64 Sep 20 17:53 1 -> /dev/null 2172f1f1f256SIvan Babrou lrwx------ 1 root root 64 Sep 20 17:53 10 -> 'socket:[12539]' 2173f1f1f256SIvan Babrou lrwx------ 1 root root 64 Sep 20 17:53 11 -> 'socket:[12540]' 2174f1f1f256SIvan Babrou lrwx------ 1 root root 64 Sep 20 17:53 12 -> 'socket:[12542]' 2175f1f1f256SIvan Babrou 2176f1f1f256SIvan BabrouThe number of open files for the process is stored in 'size' member 2177f1f1f256SIvan Babrouof stat() output for /proc/<pid>/fd for fast access. 2178f1f1f256SIvan Babrou------------------------------------------------------- 2179f1f1f256SIvan Babrou 2180f1f1f256SIvan Babrou 2181059db434SRandy DunlapChapter 4: Configuring procfs 2182059db434SRandy Dunlap============================= 2183c33e97efSMauro Carvalho Chehab 2184c33e97efSMauro Carvalho Chehab4.1 Mount options 2185c33e97efSMauro Carvalho Chehab--------------------- 2186c33e97efSMauro Carvalho Chehab 2187c33e97efSMauro Carvalho ChehabThe following mount options are supported: 2188c33e97efSMauro Carvalho Chehab 2189c33e97efSMauro Carvalho Chehab ========= ======================================================== 2190c33e97efSMauro Carvalho Chehab hidepid= Set /proc/<pid>/ access mode. 2191c33e97efSMauro Carvalho Chehab gid= Set the group authorized to learn processes information. 219237e7647aSAlexey Gladkov subset= Show only the specified subset of procfs. 2193c33e97efSMauro Carvalho Chehab ========= ======================================================== 2194c33e97efSMauro Carvalho Chehab 21951c6c4d11SAlexey Gladkovhidepid=off or hidepid=0 means classic mode - everybody may access all 21961c6c4d11SAlexey Gladkov/proc/<pid>/ directories (default). 2197c33e97efSMauro Carvalho Chehab 21981c6c4d11SAlexey Gladkovhidepid=noaccess or hidepid=1 means users may not access any /proc/<pid>/ 21991c6c4d11SAlexey Gladkovdirectories but their own. Sensitive files like cmdline, sched*, status are now 22001c6c4d11SAlexey Gladkovprotected against other users. This makes it impossible to learn whether any 22011c6c4d11SAlexey Gladkovuser runs specific program (given the program doesn't reveal itself by its 22021c6c4d11SAlexey Gladkovbehaviour). As an additional bonus, as /proc/<pid>/cmdline is unaccessible for 22031c6c4d11SAlexey Gladkovother users, poorly written programs passing sensitive information via program 22041c6c4d11SAlexey Gladkovarguments are now protected against local eavesdroppers. 2205c33e97efSMauro Carvalho Chehab 22061c6c4d11SAlexey Gladkovhidepid=invisible or hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be 22071c6c4d11SAlexey Gladkovfully invisible to other users. It doesn't mean that it hides a fact whether a 22081c6c4d11SAlexey Gladkovprocess with a specific pid value exists (it can be learned by other means, e.g. 22091c6c4d11SAlexey Gladkovby "kill -0 $PID"), but it hides process' uid and gid, which may be learned by 22101c6c4d11SAlexey Gladkovstat()'ing /proc/<pid>/ otherwise. It greatly complicates an intruder's task of 22111c6c4d11SAlexey Gladkovgathering information about running processes, whether some daemon runs with 22121c6c4d11SAlexey Gladkovelevated privileges, whether other user runs some sensitive program, whether 22131c6c4d11SAlexey Gladkovother users run any program at all, etc. 2214c33e97efSMauro Carvalho Chehab 22151c6c4d11SAlexey Gladkovhidepid=ptraceable or hidepid=4 means that procfs should only contain 22161c6c4d11SAlexey Gladkov/proc/<pid>/ directories that the caller can ptrace. 2217c33e97efSMauro Carvalho Chehab 2218c33e97efSMauro Carvalho Chehabgid= defines a group authorized to learn processes information otherwise 2219c33e97efSMauro Carvalho Chehabprohibited by hidepid=. If you use some daemon like identd which needs to learn 2220c33e97efSMauro Carvalho Chehabinformation about processes information, just add identd to this group. 222137e7647aSAlexey Gladkov 222237e7647aSAlexey Gladkovsubset=pid hides all top level files and directories in the procfs that 222337e7647aSAlexey Gladkovare not related to tasks. 222437e7647aSAlexey Gladkov 2225059db434SRandy DunlapChapter 5: Filesystem behavior 2226059db434SRandy Dunlap============================== 222737e7647aSAlexey Gladkov 222837e7647aSAlexey GladkovOriginally, before the advent of pid namepsace, procfs was a global file 222937e7647aSAlexey Gladkovsystem. It means that there was only one procfs instance in the system. 223037e7647aSAlexey Gladkov 223137e7647aSAlexey GladkovWhen pid namespace was added, a separate procfs instance was mounted in 223237e7647aSAlexey Gladkoveach pid namespace. So, procfs mount options are global among all 2233565dbe72SMauro Carvalho Chehabmountpoints within the same namespace:: 223437e7647aSAlexey Gladkov 223537e7647aSAlexey Gladkov # grep ^proc /proc/mounts 223637e7647aSAlexey Gladkov proc /proc proc rw,relatime,hidepid=2 0 0 223737e7647aSAlexey Gladkov 223837e7647aSAlexey Gladkov # strace -e mount mount -o hidepid=1 -t proc proc /tmp/proc 223937e7647aSAlexey Gladkov mount("proc", "/tmp/proc", "proc", 0, "hidepid=1") = 0 224037e7647aSAlexey Gladkov +++ exited with 0 +++ 224137e7647aSAlexey Gladkov 224237e7647aSAlexey Gladkov # grep ^proc /proc/mounts 224337e7647aSAlexey Gladkov proc /proc proc rw,relatime,hidepid=2 0 0 224437e7647aSAlexey Gladkov proc /tmp/proc proc rw,relatime,hidepid=2 0 0 224537e7647aSAlexey Gladkov 224637e7647aSAlexey Gladkovand only after remounting procfs mount options will change at all 2247565dbe72SMauro Carvalho Chehabmountpoints:: 224837e7647aSAlexey Gladkov 224937e7647aSAlexey Gladkov # mount -o remount,hidepid=1 -t proc proc /tmp/proc 225037e7647aSAlexey Gladkov 225137e7647aSAlexey Gladkov # grep ^proc /proc/mounts 225237e7647aSAlexey Gladkov proc /proc proc rw,relatime,hidepid=1 0 0 225337e7647aSAlexey Gladkov proc /tmp/proc proc rw,relatime,hidepid=1 0 0 225437e7647aSAlexey Gladkov 225537e7647aSAlexey GladkovThis behavior is different from the behavior of other filesystems. 225637e7647aSAlexey Gladkov 225737e7647aSAlexey GladkovThe new procfs behavior is more like other filesystems. Each procfs mount 225837e7647aSAlexey Gladkovcreates a new procfs instance. Mount options affect own procfs instance. 225937e7647aSAlexey GladkovIt means that it became possible to have several procfs instances 2260565dbe72SMauro Carvalho Chehabdisplaying tasks with different filtering options in one pid namespace:: 226137e7647aSAlexey Gladkov 22621c6c4d11SAlexey Gladkov # mount -o hidepid=invisible -t proc proc /proc 22631c6c4d11SAlexey Gladkov # mount -o hidepid=noaccess -t proc proc /tmp/proc 226437e7647aSAlexey Gladkov # grep ^proc /proc/mounts 22651c6c4d11SAlexey Gladkov proc /proc proc rw,relatime,hidepid=invisible 0 0 22661c6c4d11SAlexey Gladkov proc /tmp/proc proc rw,relatime,hidepid=noaccess 0 0 2267