linux(4): Eliminate unnecessary error initialization.

MFC after: 2 weeks

linux(4): Eliminate unnecessary head initialization.

MFC after: 2 weeks

linux(4): style, wrap too long line.

MFC after: 2 weeks

linux(4): Eliminating remnants of futex sdt.

MFC after: 2 weeks

linux(4): Handle special case for regular futex in handle_futex_death().

Handle some races in handle_futex_death() which can prevents a wakeup of
potential waiters which can cause these waiters to b

linux(4): Handle special case for regular futex in handle_futex_death().

Handle some races in handle_futex_death() which can prevents a wakeup of
potential waiters which can cause these waiters to block forever.

Differential Revision: https://reviews.freebsd.org/D31280
MFC after: 2 weeks

show more ...

linux(4): Futex address must be 32-bit aligned.

Linux futex documentation explicitly states that EINVAL is returned if
the futex is not 4-byte aligned. Check futex alignment as a Linux do
and return

linux(4): Futex address must be 32-bit aligned.

Linux futex documentation explicitly states that EINVAL is returned if
the futex is not 4-byte aligned. Check futex alignment as a Linux do
and return EINVAL.

Differential Revision: https://reviews.freebsd.org/D31279
MFC after: 2 weeks

show more ...

linux(4): Finish cf8d74e3fe63.

Add forgotten val3_compare initialization in case of time64 futex.

MFC after: 2 weeks

linux(4): Replace casuword32 by casueword32.

Follow the r349951 (30b3018d), add check to react to stops and requests
to terminate between retries.

Reviewed by: kib
Differential Revision: https://r

linux(4): Replace casuword32 by casueword32.

Follow the r349951 (30b3018d), add check to react to stops and requests
to terminate between retries.

Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D31254
MFC after: 2 weeks

show more ...

linux(4): Implement pi futexes using umtx.

Differential Revision: https://reviews.freebsd.org/D31240
MFC after: 2 weeks

linux(4): Replace copyin() by fueword32() in handle_futex_death().

According to fetch(9) fueword facility designed to fetch atomically
small amount of data from user space.

Reviewed by: kib
Differ

linux(4): Replace copyin() by fueword32() in handle_futex_death().

According to fetch(9) fueword facility designed to fetch atomically
small amount of data from user space.

Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D31239
MFC after: 2 weeks

show more ...

linux(4): Reimplement futexes using umtx.

Differential Revision: https://reviews.freebsd.org/D31236
MFC after: 2 weeks

umtx: Split umtx.h on two counterparts.

To prevent umtx.h polluting by future changes split it on two headers:
umtx.h - ABI header for userspace;
umtxvar.h - the kernel staff.

While here fix umtx_k

umtx: Split umtx.h on two counterparts.

To prevent umtx.h polluting by future changes split it on two headers:
umtx.h - ABI header for userspace;
umtxvar.h - the kernel staff.

While here fix umtx_key_match style.

Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D31248
MFC after: 2 weeks

show more ...

linux(4): Factor out the futex_wait() op into linux_futex_wait().

MFC after: 2 weeks

linux(4): Prevent an endless loop.

In the futex_atomic_op() the encoded_op is a user-supplied parameter.
If the user specifies an incorrect value for this parameter paired with a valid
*uaddr parame

linux(4): Prevent an endless loop.

In the futex_atomic_op() the encoded_op is a user-supplied parameter.
If the user specifies an incorrect value for this parameter paired with a valid
*uaddr parameter the caller will go into the endless loop. To prevent this check
futex_atomic_op() result and break the loop in case of ENOSYS.

MFC after: 2 weeks

show more ...

linux(4): Eliminate bogus comment.

For the caller is no need for access checking here, as the caller must take care
of EFAULT handling. Moreover, this check would be superfluous, since EFAULT is
ext

linux(4): Eliminate bogus comment.

For the caller is no need for access checking here, as the caller must take care
of EFAULT handling. Moreover, this check would be superfluous, since EFAULT is
extremily rare, and we prefer the fast path.

MFC after: 2 weeks

show more ...

linux(4): Allow musl brand to use FUTEX_REQUEUE op.

Initial patch from submitter was adapted by me to prevent unconditional
FUTEX_REQUEUE use.

PR: 255947
Submitted by: Philippe Michaud-Boudreaul

linux(4): Allow musl brand to use FUTEX_REQUEUE op.

Initial patch from submitter was adapted by me to prevent unconditional
FUTEX_REQUEUE use.

PR: 255947
Submitted by: Philippe Michaud-Boudreault
Differential Revision: https://reviews.freebsd.org/D30332

show more ...

linux(4): Factor out the FUTEX_WAKE_OP op into linux_futex_wakeop().

MFC after: 2 weeks

linux(4): Factor out the FUTEX_CMP_REQUEUE op into linux_futex_requeue().

MFC after: 2 weeks

linux(4): Factor out the FUTEX_WAKE op into linux_futex_wake().

MFC after: 2 weeks

linux(4): Factor out the FUTEX_WAIT op into linux_futex_wait().

MFC after: 2 weeks

linux(4): Refactor the struct linux_futex_args.

Move flags and rtclock to the struct linux_futex_args. This will be used when
I split linux_futex() into separate futex op functions.

MFC after: 2 w

linux(4): Refactor the struct linux_futex_args.

Move flags and rtclock to the struct linux_futex_args. This will be used when
I split linux_futex() into separate futex op functions.

MFC after: 2 weeks

show more ...

Drop "All rights reserved" from my copyright statements.

Add email and fixup years while here.

Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D30912
MFC after: 2 weeks

linux(4): Implement futex_time64 system call.

MFC after: 2 weeks

linux(4): Prevent integer overflow in futex_requeue.

To prevent a signed integer overflow in futex_requeue add a sanity check
to catch negative values of nrwake or nrrequeue.

MFC after: 2 weeks

linux: remove redundant SDT tracepoints

Remove all the 'entry' and 'return' probes; they clutter up the source
and are redundant to FBT.

Reviewed By: dchagin
Sponsored By: EPSRC
Differential Revisi

linux: remove redundant SDT tracepoints

Remove all the 'entry' and 'return' probes; they clutter up the source
and are redundant to FBT.

Reviewed By: dchagin
Sponsored By: EPSRC
Differential Revision: https://reviews.freebsd.org/D30040

show more ...