linuxkpi: Add `woken_wake_function()` and `wait_woken()`

They are used by the i915 DRM driver starting with Linux 6.7.

`(struct wait_queue)->flags` is no longer always zero. I wonder if some
code r

linuxkpi: Add `woken_wake_function()` and `wait_woken()`

They are used by the i915 DRM driver starting with Linux 6.7.

`(struct wait_queue)->flags` is no longer always zero. I wonder if some
code relied on this...

Reviewed by: markj
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D48755

show more ...

linuxkpi: Fix up jiffies handling

A few issues found by code inspection while hunting for bugzilla PR
286512:
- The "expires" field in struct delayed_work should be unsigned.
- In linux_timer_jiffie

linuxkpi: Fix up jiffies handling

A few issues found by code inspection while hunting for bugzilla PR
286512:
- The "expires" field in struct delayed_work should be unsigned.
- In linux_timer_jiffies_until(), clamp the return value to INT_MAX:
this return value is used as a ticks count, not a jiffies count, so we
should avoid returning too large a value, lest it get truncated. It's
unlikely we are dealing with values that large, but we should be
careful anyway.
- In linux_add_to_sleepqueue(), truncate the timeout to INT_MAX, as this
value is passed to sleepq_set_timeout() as a ticks value. Typically
it's multiplied by ticks_sbt to get an sbintime, and we should make
sure the multiplication doesn't overflow. In drm-kmod, there is at
least one call mod_delayed_work(... MAX_SCHEDULE_TIMEOUT).

Fixes: 325aa4dbd10d ("linuxkpi: Introduce a properly typed jiffies")
Reviewed by: olce, bz, dumbbell, kib
Tested by: dumbbell, bz
Differential Revision: https://reviews.freebsd.org/D50192

show more ...

linuxkpi: Introduce a properly typed jiffies

Now that we have a long-sized tick counter, we can migrate to using
properly typed timeout parameters in various bits of the LinuxKPI. This
fixes a subt

linuxkpi: Introduce a properly typed jiffies

Now that we have a long-sized tick counter, we can migrate to using
properly typed timeout parameters in various bits of the LinuxKPI. This
fixes a subtle incompatibility that is otherwise difficult to paper over
and leads to bugs when ticks values are sign-extended.

- Introduce a "jiffies" symbol in subr_ticks.S, declared only in the
LinuxKPI as an unsigned long.
- Remove all references to "ticks" from the LinuxKPI.
- Convert interfaces to match Linux's type signatures where it makes
sense.

Reviewed by: manu
Tested by: bz
Differential Revision: https://reviews.freebsd.org/D48523

show more ...

sleepqueue: Remove kernel stack swapping support, part 10

- Remove kick_proc0().
- Make the return type of sleepq_broadcast(), sleepq_signal(), etc.,
void.
- Fix up callers.

Tested by: pho
Review

sleepqueue: Remove kernel stack swapping support, part 10

- Remove kick_proc0().
- Make the return type of sleepq_broadcast(), sleepq_signal(), etc.,
void.
- Fix up callers.

Tested by: pho
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D46128

show more ...

proc: Remove kernel stack swapping support, part 7

Remove some uses of PHOLD which were there only to prevent the process'
threads from being swapped out.

Tested by: pho
Reviewed by: imp, kib
Diffe

proc: Remove kernel stack swapping support, part 7

Remove some uses of PHOLD which were there only to prevent the process'
threads from being swapped out.

Tested by: pho
Reviewed by: imp, kib
Differential Revision: https://reviews.freebsd.org/D46118

show more ...

sys: Automated cleanup of cdefs and other formatting

Apply the following automated changes to try to eliminate
no-longer-needed sys/cdefs.h includes as well as now-empty
blank lines in a row.

Remov

sys: Automated cleanup of cdefs and other formatting

Apply the following automated changes to try to eliminate
no-longer-needed sys/cdefs.h includes as well as now-empty
blank lines in a row.

Remove /^#if.*\n#endif.*\n#include\s+<sys/cdefs.h>.*\n/
Remove /\n+#include\s+<sys/cdefs.h>.*\n+#if.*\n#endif.*\n+/
Remove /\n+#if.*\n#endif.*\n+/
Remove /^#if.*\n#endif.*\n/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/types.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/param.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/capsicum.h>/

Sponsored by: Netflix

show more ...

sys: Remove $FreeBSD$: one-line .c pattern

Remove /^[\s*]*__FBSDID\("\$FreeBSD\$"\);?\s*\n/

Revert "LinuxKPI: Allow wake_up to be executed within a critical section"

This change was based on currently reverted commit 7dea0c9e6eba.

This reverts commit 89889ab470b9f1a1cd36913dd219b78efbf484

Revert "LinuxKPI: Allow wake_up to be executed within a critical section"

This change was based on currently reverted commit 7dea0c9e6eba.

This reverts commit 89889ab470b9f1a1cd36913dd219b78efbf484df.

show more ...

LinuxKPI: Allow wake_up to be executed within a critical section

by replaceing of spin_lock() call with spin_lock_irqsave()

This fixes following panic in drm-kmod:

panic: mi_switch: switch in a cr

LinuxKPI: Allow wake_up to be executed within a critical section

by replaceing of spin_lock() call with spin_lock_irqsave()

This fixes following panic in drm-kmod:

panic: mi_switch: switch in a critical section
cpuid = 2
time = 1636939794
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b
vpanic() at vpanic+0x187
panic() at panic+0x43
mi_switch() at mi_switch+0x198
__mtx_lock_sleep() at __mtx_lock_sleep+0x1c9
__mtx_lock_flags() at __mtx_lock_flags+0xa2
linux_wake_up() at linux_wake_up+0x38
__active_retire() at __active_retire+0xb7
dma_fence_signal() at dma_fence_signal+0x100
dma_resv_add_shared_fence() at dma_resv_add_shared_fence+0x96
i915_gem_do_execbuffer() at i915_gem_do_execbuffer+0x11d0
i915_gem_execbuffer2_ioctl() at i915_gem_execbuffer2_ioctl+0x19a
drm_ioctl_kernel() at drm_ioctl_kernel+0x72
drm_ioctl() at drm_ioctl+0x2c4
linux_file_ioctl() at linux_file_ioctl+0x297
kern_ioctl() at kern_ioctl+0x1dc
sys_ioctl() at sys_ioctl+0x124
amd64_syscall() at amd64_syscall+0x124
fast_syscall_common() at fast_syscall_common+0xf8
--- syscall (54, FreeBSD ELF64, sys_ioctl)

MFC after: 1 week
Reviewed by: manu
Reported by: Graham Perrin <grahamperrin_AT_gmail_DOT_com>
PR: 261166
Differential Revision: https://reviews.freebsd.org/D33888

show more ...

Implement the init_wait_entry() function macro in the LinuxKPI.

Submitted by: Johannes Lundberg <johalun0@gmail.com>
MFC after: 1 week
Sponsored by: Mellanox Technologies
Sponsored by: Limelight Net

Implement the init_wait_entry() function macro in the LinuxKPI.

Submitted by: Johannes Lundberg <johalun0@gmail.com>
MFC after: 1 week
Sponsored by: Mellanox Technologies
Sponsored by: Limelight Networks

show more ...

Allow TASK_PARKED bit being set when going to sleep in the LinuxKPI.

Found by: Johannes Lundberg <johalun0@gmail.com>
MFC after: 1 week
Sponsored by: Mellanox Technologies

Wrap long lines.

MFC after: 3 days

Optimise use of Giant in the LinuxKPI.

- Make sure Giant is locked when calling PCI device methods.
Newbus currently requires this.

- Avoid unlocking Giant right before aquiring the sleepqueue lock

Optimise use of Giant in the LinuxKPI.

- Make sure Giant is locked when calling PCI device methods.
Newbus currently requires this.

- Avoid unlocking Giant right before aquiring the sleepqueue lock.
This can save a task switch.

MFC after: 1 week
Sponsored by: Mellanox Technologies

show more ...

Implement DEFINE_WAIT_FUNC() function macro and default_wake_function()
in the LinuxKPI.

MFC after: 1 week
Submitted by: Johannes Lundberg <johalun0@gmail.com>
Sponsored by: Mellanox Technologies
Sp

Implement DEFINE_WAIT_FUNC() function macro and default_wake_function()
in the LinuxKPI.

MFC after: 1 week
Submitted by: Johannes Lundberg <johalun0@gmail.com>
Sponsored by: Mellanox Technologies
Sponsored by: Limelight Networks

show more ...

Use mstosbt() instead of SBT_1MS in the LinuxKPI to get the last few bits
of precision.

MFC after: 1 week
Suggested by: ian@
Sponsored by: Mellanox Technologies

Implement msleep_interruptible() in the LinuxKPI. While at it use pause_sbt()
instead of pause() in the msleep() function to avoid rounding errors when
converting delay values forth and back. Add a g

Implement msleep_interruptible() in the LinuxKPI. While at it use pause_sbt()
instead of pause() in the msleep() function to avoid rounding errors when
converting delay values forth and back. Add a guard for a delay value
of zero milliseconds which is undefined.

MFC after: 1 week
Requested by: Johannes Lundberg <johalun0@gmail.com>
Sponsored by: Mellanox Technologies

show more ...

Return correct error code to user-space when a system call receives a
signal in the LinuxKPI.

The read(), write() and mmap() system calls can return either EINTR or
ERESTART upon receiving a signal.

Return correct error code to user-space when a system call receives a
signal in the LinuxKPI.

The read(), write() and mmap() system calls can return either EINTR or
ERESTART upon receiving a signal. Add code to figure out the correct
return value by temporarily storing the return code from the relevant
FreeBSD kernel APIs in the Linux task structure.

MFC after: 3 days
Sponsored by: Mellanox Technologies

show more ...

Merge ^/head r325663 through r325841.

Remove release and acquire semantics when accessing the "state" field of the
LinuxKPI task struct. Change type of "state" variable from "int" to
"atomic_t" to simplify code and avoid unneccessary cas

Remove release and acquire semantics when accessing the "state" field of the
LinuxKPI task struct. Change type of "state" variable from "int" to
"atomic_t" to simplify code and avoid unneccessary casting.

MFC after: 1 week
Sponsored by: Mellanox Technologies

show more ...

Merge ^/head r323559 through r325504.

MFhead@r324837

While here, diff reduce some of the changes in sys/boot by moving
MK_COVERAGE=no to sys/boot/Makefile.inc .

Make the PHOLD in linux_wait_event_common() unconditional.

After some in-progress work is committed, this would otherwise be the only
instance of #if(n)def NO_SWAPPING in the tree. Moreover, the req

Make the PHOLD in linux_wait_event_common() unconditional.

After some in-progress work is committed, this would otherwise be the only
instance of #if(n)def NO_SWAPPING in the tree. Moreover, the requisite
opt_vm.h include was missing, so the PHOLD/PRELE calls were always being
compiled in anyway.

MFC after: 1 week

show more ...

MFhead@r322451

Merge ^/head r321383 through r322397.

Make sure the linux_wait_event_common() function in the LinuxKPI properly
handles a timeout value of MAX_SCHEDULE_TIMEOUT which basically means there
is no timeout. This is a regression issue after r

Make sure the linux_wait_event_common() function in the LinuxKPI properly
handles a timeout value of MAX_SCHEDULE_TIMEOUT which basically means there
is no timeout. This is a regression issue after r319757.

While at it change the type of returned variable from "long" to "int" to
match the actual return type.

MFC after: 1 week
Sponsored by: Mellanox Technologies

show more ...