kib retitled D58626: autofs_lookup(): busy the mount point around autofs_trigger() from autofs: busy the mount point around autofs_trigger() to autofs_lookup(): busy the mount point around autofs_trigger().

Mon, Aug 3, 5:56 PM

kib requested review of D58626: autofs_lookup(): busy the mount point around autofs_trigger().

Mon, Aug 3, 5:45 PM

kib added a comment to D58624: pci: Add tunable to reserve bridge bus numbers.

If this is a solution, then IMO it should be activated by default. How would user know to tweak this knob if SR-IOV fails?

Mon, Aug 3, 4:47 PM

kib accepted D58623: rtld: Remove a stale #ifdef PIC.

Mon, Aug 3, 3:12 PM

kib added a comment to D58337: umtx: add FAIR_LINUX_FUTEX_HASHING for a distribution-fair chain hash.

I do not see how this could be an option. Either there is no (bad) impact on the native umtxes, and then we should just change the constant, perhaps adding a comment explaining its origin and reasoning behind. Or there is the impact, but then you could check that the curproc ABI is linux, and use the proposed constant for Linux processes.

Mon, Aug 3, 6:06 AM

kib committed rG0aa72376d73b: geom_zero(4): reset the uio vector on each uiomove() (authored by kib).

geom_zero(4): reset the uio vector on each uiomove()

Mon, Aug 3, 12:26 AM

kib committed rG7d243813ca63: tests/libpthread: add pthread_cond_clockwait(3) tests (authored by kib).

tests/libpthread: add pthread_cond_clockwait(3) tests

Mon, Aug 3, 12:26 AM

kib committed rGd62a413e76bd: pthread_cond_timedwait.3: document pthread_cond_clockwait(3) (authored by kib).

pthread_cond_timedwait.3: document pthread_cond_clockwait(3)

Mon, Aug 3, 12:26 AM

kib committed rGff36771c13e9: pthread_cond_timedwait.3: use .Fo/.Fc for long arguments list (authored by kib).

pthread_cond_timedwait.3: use .Fo/.Fc for long arguments list

Mon, Aug 3, 12:26 AM

kib committed rG4dcfa6d003e0: libthr: implement pthread_cond_clockwait(3) (authored by kib).

libthr: implement pthread_cond_clockwait(3)

Mon, Aug 3, 12:26 AM

kib committed rGd32bf064e951: libthr/thread/thr_cond.c: some style (authored by kib).

libthr/thread/thr_cond.c: some style

Mon, Aug 3, 12:26 AM

kib committed rG0737ff2da5f4: libthr/thread/thr_umtx.c: style _thr_ucond_wait() (authored by kib).

libthr/thread/thr_umtx.c: style _thr_ucond_wait()

Mon, Aug 3, 12:26 AM

kib committed rG16015bf932bd: umtx_op(2): add the CVWAIT_UMTX_TIME flag for the UMTX_OP_CV_WAIT_UC request (authored by kib).

umtx_op(2): add the CVWAIT_UMTX_TIME flag for the UMTX_OP_CV_WAIT_UC request

Mon, Aug 3, 12:26 AM

kib committed rG671e2cb6a237: _umtx_op.2: document the CVWAIT_UMTX_TIME flag for the UMTX_OP_CV_WAIT_UC… (authored by kib).

_umtx_op.2: document the CVWAIT_UMTX_TIME flag for the UMTX_OP_CV_WAIT_UC…

Mon, Aug 3, 12:26 AM

Sun, Aug 2

kib committed rG4c58eef12d30: stat.2: enhance the description of st_blocks (authored by kib).

stat.2: enhance the description of st_blocks

Sun, Aug 2, 10:18 PM

kib closed D58592: stat.2: enhance the description of st_blocks.

Sun, Aug 2, 10:18 PM

kib updated the diff for D58592: stat.2: enhance the description of st_blocks.

Fore
segments
mention 'non-synthetic' for filesystems

Sun, Aug 2, 10:05 PM

kib updated the diff for D58567: dounmount(9): temporarily enable recursion for the covered vnode lock.

Add comment.
Do not unconditionally disable recursion for coveredvnode on error.

Sun, Aug 2, 7:32 PM

kib added a comment to D58592: stat.2: enhance the description of st_blocks.

In D58592#1344688, @markmi_dsl-only.net wrote:
Does:
. . .
STANDARDS
The stat() and fstat() system calls are expected to conform to IEEE Std
1003.1-1990 (“POSIX.1”).  The fstatat() system call follows The Open
Group Extended API Set 2 specification.
. . .

need any related wording about extensions or aspects not covered by POSIX for the likes of fstat?

Sun, Aug 2, 5:41 PM

kib accepted D58511: ofed: fix GID table reference leak in roce_gid_update_addr_callback().

Sun, Aug 2, 3:34 PM

kib accepted D58510: ofed: don't stop removing stale RoCE GIDs at the first hole.

Sun, Aug 2, 3:34 PM

kib added a comment to D58593: Bump __FreeBSD_version to 1600020 after the amd64 FRED KBI change.

What are the hypothetical KBI incompatibilities caused by the FRED commit?

Sun, Aug 2, 3:30 PM

kib added a comment to D38545: linsysfs: Reimplement bus scan code..

The patch requires a lot of work. It is not complicated, but quite time consuming. Also, it requires access to dual-bool the machine with Linux and FreeBSD, to compare the outcome on the same system.

Sun, Aug 2, 3:01 PM

kib requested review of D58592: stat.2: enhance the description of st_blocks.

Sun, Aug 2, 2:59 PM

Sat, Aug 1

kib added a comment to D58586: Make ptrace(2) operating on process descriptors.

This patch adds a capability for ptrace, and allows to do ptrace(req, procdesc, addr, data) instead of ptrace(req, pid, addr, data).

Sat, Aug 1, 12:32 PM

kib requested review of D58586: Make ptrace(2) operating on process descriptors.

Sat, Aug 1, 12:30 PM

kib committed rG77d6c45afdca: tests/sys/capsicum: adjust tests for the new reaping behavior (authored by kib).

tests/sys/capsicum: adjust tests for the new reaping behavior

Sat, Aug 1, 12:20 PM

kib committed rGa96f285b5953: tests/sys/kern: adjust tests for the new reaping behavior (authored by kib).

tests/sys/kern: adjust tests for the new reaping behavior

Sat, Aug 1, 12:20 PM

kib committed rGc49198534a9b: pdwait(2), wait(2): document interaction between pdopenpid() and waitpid() (authored by kib).

pdwait(2), wait(2): document interaction between pdopenpid() and waitpid()

Sat, Aug 1, 12:20 PM

kib committed rGbcdb6ba94d08: processes: add zombie references, each of them prevents reap (authored by kib).

processes: add zombie references, each of them prevents reap

Sat, Aug 1, 12:20 PM

kib committed rG7da4edd30098: pdwait(2): change handling of the exited processes (authored by kib).

pdwait(2): change handling of the exited processes

Sat, Aug 1, 12:20 PM

kib closed D58569: tests/sys/capsicum: adjust tests for the new reaping behavior.

Sat, Aug 1, 12:20 PM

kib closed D58530: tests/sys/kern: adjust tests for new reaping behavior.

Sat, Aug 1, 12:20 PM

kib committed rGa24674ecab24: kern/kern_exit.c: make wait_fill_siginfo/wrusage global (authored by kib).

kern/kern_exit.c: make wait_fill_siginfo/wrusage global

Sat, Aug 1, 12:20 PM

kib closed D58264: processes: add zombie references, each of them prevents reap.

Sat, Aug 1, 12:19 PM

kib closed D58407: pdwait(2): change handling of the exited processes.

Sat, Aug 1, 12:19 PM

kib added a comment to D58264: processes: add zombie references, each of them prevents reap.

In D58264#1344458, @pho wrote:

I am, although it seems strange that I was not able to reproduce Mark's zombie issue.

Sat, Aug 1, 11:38 AM

kib added a comment to D58264: processes: add zombie references, each of them prevents reap.

In D58264#1344442, @kib wrote:
In D58264#1344441, @pho wrote:
In D58264#1343336, @kib wrote:
In D58264#1343328, @markj wrote:
I ran the full regression test suite with the three patches applied, and got an OOM somewhere in the network tests. I can see a zombie leak somewhere, running the test again I see:
6441 threads:  19 running, 834 sleeping, 4960 zombie, 628 waiting
and the zombie count is growing slowly.
Could you please print out p_zombierefs for several of such zombies?
BTW one of the changes that I upload shortly might fix it, but I am not sure.

I will ask pho to help with the reproduction.
I ran most of the stress2 tests plus the tests in /usr/tests, without seeing any zombie leaks. This on ufs-dd72f46f512f.
Sorry, do you mean the ufs branch from the deviant3? It does not have these changes.

Could you please give the pdrfork branch some run time?

Sat, Aug 1, 10:40 AM

kib added a comment to D58264: processes: add zombie references, each of them prevents reap.

In D58264#1344441, @pho wrote:
In D58264#1343336, @kib wrote:
In D58264#1343328, @markj wrote:
I ran the full regression test suite with the three patches applied, and got an OOM somewhere in the network tests. I can see a zombie leak somewhere, running the test again I see:
6441 threads:  19 running, 834 sleeping, 4960 zombie, 628 waiting
and the zombie count is growing slowly.
Could you please print out p_zombierefs for several of such zombies?
BTW one of the changes that I upload shortly might fix it, but I am not sure.

I will ask pho to help with the reproduction.
I ran most of the stress2 tests plus the tests in /usr/tests, without seeing any zombie leaks. This on ufs-dd72f46f512f.

Sat, Aug 1, 10:15 AM

Fri, Jul 31

kib accepted D58580: atomic: Implement atomic_{set,clear}_8 in _atomic_subword.h.

I wonder if we should start provide the definitive list of MI aromics in atomic(9).

Fri, Jul 31, 11:16 PM

kib committed rG833bdae6c58b: proc_realparent(): do not mark the child as orphan when reparenting to p_opptr… (authored by kib).

proc_realparent(): do not mark the child as orphan when reparenting to p_opptr…

Fri, Jul 31, 3:58 PM

kib closed D58566: proc_realparent(): do not mark the child as orphan when reparenting to p_opptr pid.

Fri, Jul 31, 3:58 PM

kib accepted D58577: ptrace: Propagate errors from set_fpregs().

Fri, Jul 31, 3:47 PM

kib accepted D58577: ptrace: Propagate errors from set_fpregs().

Fri, Jul 31, 3:41 PM

kib added a comment to D58264: processes: add zombie references, each of them prevents reap.

In D58264#1344203, @markj wrote:

I think this is ok. The result is not quite the same as what I proposed originally: I believe, with your patches, that if a process creates a child with pdfork(), then it cannot use pdwait() to reap the child, it still must use waitpid(). It would be nicer to make pdwait() reap the child in this case, IMO.

Fri, Jul 31, 3:35 PM

kib added inline comments to D58320: LinuxKPI: Add PAGE_IS_LKPI_PAGE for Linux struct page instead of vm_page.

Fri, Jul 31, 1:04 PM

kib added a comment to D58569: tests/sys/capsicum: adjust tests for the new reaping behavior.

I handled the SIGCHLD delivery check in the destructor somewhat rude, but the tests are not structured to make it easy to handle properly.
Also I did not tried that under Linux, I think it is not needed.

Fri, Jul 31, 9:53 AM

kib requested review of D58569: tests/sys/capsicum: adjust tests for the new reaping behavior.

Fri, Jul 31, 9:51 AM

kib added inline comments to D58562: amd64: Implement per-thread kernel stack canaries.

Fri, Jul 31, 6:13 AM

kib accepted D58549: vfs_mountroot: unmute console in interactive prompt.

Fri, Jul 31, 4:24 AM

kib requested review of D58567: dounmount(9): temporarily enable recursion for the covered vnode lock.

Fri, Jul 31, 4:21 AM

kib updated the summary of D58566: proc_realparent(): do not mark the child as orphan when reparenting to p_opptr pid.

Fri, Jul 31, 2:07 AM

kib requested review of D58566: proc_realparent(): do not mark the child as orphan when reparenting to p_opptr pid.

Fri, Jul 31, 2:06 AM

Thu, Jul 30

kib committed rG72952bf6a307: amd64: try to fix the build with old clang that does not know about FRED (authored by kib).

amd64: try to fix the build with old clang that does not know about FRED

Thu, Jul 30, 5:09 PM

kib committed rGcef05c5a62ba: amd64: do not allow to set reserved bits in MXCSR for ptrace(PT_SETFPREGS) (authored by kib).

amd64: do not allow to set reserved bits in MXCSR for ptrace(PT_SETFPREGS)

Thu, Jul 30, 5:09 PM

kib closed D58550: amd64: try to fix the build with old clang that does not know about FRED.

Thu, Jul 30, 5:09 PM

kib closed D58548: amd64: do not allow to set reserved bits in MXCSR for ptrace(PT_SETFPREGS).

Thu, Jul 30, 5:09 PM

kib added a comment to D58548: amd64: do not allow to set reserved bits in MXCSR for ptrace(PT_SETFPREGS).

In D58548#1343721, @jhb wrote:

Have you had a chance to test this under GDB and/or LLDB? I can't imagine they'd be setting any reserved bits so this should be fine.

Thu, Jul 30, 4:47 PM

kib updated the diff for D58550: amd64: try to fix the build with old clang that does not know about FRED.

More tabs.

Thu, Jul 30, 11:43 AM

kib added inline comments to D58550: amd64: try to fix the build with old clang that does not know about FRED.

Thu, Jul 30, 11:42 AM

kib updated the diff for D58550: amd64: try to fix the build with old clang that does not know about FRED.

Use tabs after .if/.err

Thu, Jul 30, 11:21 AM

kib updated the diff for D58550: amd64: try to fix the build with old clang that does not know about FRED.

tabs not spaces

Thu, Jul 30, 11:04 AM

kib updated the diff for D58550: amd64: try to fix the build with old clang that does not know about FRED.

clang <= 18.
.err for non-ax.

Thu, Jul 30, 10:55 AM

kib added a reviewer for D58550: amd64: try to fix the build with old clang that does not know about FRED: jrtc27.

Thu, Jul 30, 10:22 AM

kib requested review of D58550: amd64: try to fix the build with old clang that does not know about FRED.

Thu, Jul 30, 9:48 AM

kib accepted D58540: stand: Improve error handling when loading ELF files.

Thu, Jul 30, 9:17 AM

kib accepted D58543: stand: Reject ELF files with PT_LOAD segments where filesz > memsz.

Thu, Jul 30, 8:13 AM

kib accepted D58542: kld: Reject kernel modules with PT_LOAD segments where filesz > memsz.

Thu, Jul 30, 8:12 AM

kib updated the diff for D58530: tests/sys/kern: adjust tests for new reaping behavior.

More tests updates

Thu, Jul 30, 6:15 AM

kib accepted D58539: stand: Remove a pointless goto.

Thu, Jul 30, 5:53 AM

kib accepted D58541: rtld: Reject ELF files with PT_LOAD or PT_TLS segments where filesz > memsz.

I do not object, but I also note that rtld trusts the format of the objects to be correct. There is a lot of missed bounds checks.

Thu, Jul 30, 4:24 AM

kib requested review of D58548: amd64: do not allow to set reserved bits in MXCSR for ptrace(PT_SETFPREGS).

Thu, Jul 30, 4:22 AM

kib committed rG930f2e4da964: kern_execve(): avoid storing non-VDIR into p_textdvp (authored by kib).

kern_execve(): avoid storing non-VDIR into p_textdvp

Thu, Jul 30, 2:57 AM

kib closed D58506: kern_execve(): avoid storing non-VDIR into p_textdvp.

Thu, Jul 30, 2:57 AM

kib committed rGbf0ffcaa5649: mknod.2: properly document root requirements (authored by kib).

mknod.2: properly document root requirements

Thu, Jul 30, 2:54 AM

kib committed rG26cad57094d1: pthread_cond_wait.3: describe spurious wakeups (authored by kib).

pthread_cond_wait.3: describe spurious wakeups

Thu, Jul 30, 2:54 AM

kib committed rG361ee195955f: mknod.2: update the man page (authored by kib).

mknod.2: update the man page

Thu, Jul 30, 2:54 AM

Wed, Jul 29

kib updated the diff for D58264: processes: add zombie references, each of them prevents reap.

Clear PZOMBIEREF_PROCDESC when the last procdesc reference is gone.
Man page editing.

Wed, Jul 29, 4:21 PM

kib added a comment to D58264: processes: add zombie references, each of them prevents reap.

In D58264#1343328, @markj wrote:
I ran the full regression test suite with the three patches applied, and got an OOM somewhere in the network tests. I can see a zombie leak somewhere, running the test again I see:
6441 threads:  19 running, 834 sleeping, 4960 zombie, 628 waiting
and the zombie count is growing slowly.

Wed, Jul 29, 4:20 PM

kib added a comment to D58530: tests/sys/kern: adjust tests for new reaping behavior.

The changes required to the test suite are very illustrative.

Wed, Jul 29, 6:13 AM

kib (Konstantin Belousov)User

Projects

User Details

Recent ActivityView All

Today

Yesterday

Sun, Aug 2

Sat, Aug 1

Fri, Jul 31

Thu, Jul 30

Wed, Jul 29

kib (Konstantin Belousov)
User

Recent Activity
View All