kib (Konstantin Belousov)
User

Projects

User Details

User Since
May 16 2014, 7:35 PM (162 w, 3 d)

Recent Activity

Yesterday

kib added inline comments to D11373: Fix one more place uio_resid is truncated to int.
Mon, Jun 26, 9:27 PM
kib added inline comments to D11373: Fix one more place uio_resid is truncated to int.
Mon, Jun 26, 9:23 PM
kib committed rS320358: MFC r320121:.
MFC r320121:
Mon, Jun 26, 12:37 PM
kib committed rS320357: MFC r320125:.
MFC r320125:
Mon, Jun 26, 12:30 PM
kib committed rS320356: MFC r320108:.
MFC r320108:
Mon, Jun 26, 12:17 PM
kib accepted D11364: Fix a comment post-r320347.
Mon, Jun 26, 11:39 AM

Sun, Jun 25

kib committed rS320344: For now, allow mprotect(2) over the guards to succeed regardless of.
For now, allow mprotect(2) over the guards to succeed regardless of
Sun, Jun 25, 11:16 PM
kib accepted D11358: Another round of subr_blist.c fixes.
Sun, Jun 25, 9:38 PM
kib committed rS320339: Correctly handle small MAP_STACK requests..
Correctly handle small MAP_STACK requests.
Sun, Jun 25, 8:06 PM
kib committed rS320338: Remove stale part of the comment..
Remove stale part of the comment.
Sun, Jun 25, 7:59 PM
kib committed rS320332: Style..
Style.
Sun, Jun 25, 6:41 PM
kib added inline comments to D11358: Another round of subr_blist.c fixes.
Sun, Jun 25, 6:40 PM

Sat, Jun 24

kib committed rD50405: Document __FreeBSD_versions 1200034 and 1200035..
Document __FreeBSD_versions 1200034 and 1200035.
Sat, Jun 24, 8:09 PM
kib closed D11306: Document MAP_GUARD..
Sat, Jun 24, 5:10 PM
kib committed rS320318: Use address space guard to implement inter-segment gap..
Use address space guard to implement inter-segment gap.
Sat, Jun 24, 5:04 PM
kib committed rS320317: Implement address space guards..
Implement address space guards.
Sat, Jun 24, 5:01 PM
kib committed rS320316: Do not try to unmark MAP_ENTRY_IN_TRANSITION marked by other thread..
Do not try to unmark MAP_ENTRY_IN_TRANSITION marked by other thread.
Sat, Jun 24, 4:48 PM
kib committed rS320314: Remove the description of MAP_HASSEMAPHORE..
Remove the description of MAP_HASSEMAPHORE.
Sat, Jun 24, 4:36 PM
kib committed rS320313: Fix typo..
Fix typo.
Sat, Jun 24, 4:21 PM
kib added a comment to D11323: Fix "mount -uw /" when the filesystem type doesn't match..

I don't have a slightest idea why the special case is required. But it might be, and removing it altogether might break somebody's system. Adding additional condition is much less intrusive, and thus safer.

Sat, Jun 24, 2:53 PM
kib committed rS320312: MFC r320052:.
MFC r320052:
Sat, Jun 24, 2:45 PM
kib committed rS320311: MFC r320052:.
MFC r320052:
Sat, Jun 24, 2:42 PM
kib committed rS320309: MFC r320051:.
MFC r320051:
Sat, Jun 24, 1:53 PM
kib committed rS320308: Translate between abridged and full x87 tags for compat32.
Translate between abridged and full x87 tags for compat32
Sat, Jun 24, 11:38 AM
kib committed rS320307: Fix indent..
Fix indent.
Sat, Jun 24, 10:19 AM
kib updated the diff for D11306: Document MAP_GUARD..

More errors explanations.
Fix unrelated typo.

Sat, Jun 24, 9:54 AM

Fri, Jun 23

kib added a comment to D11197: msdosfs: expose and hide definitions/functions for makefs FAT fs support.

Honestly, I dislike the change. IMO, the msdosfs code should be copied to the userspace utility instead of making the hard-to-deal knot between kernel and userspace code.

Fri, Jun 23, 9:15 PM
kib added a comment to D11323: Fix "mount -uw /" when the filesystem type doesn't match..

Can you explain why this special case is required ?

Fri, Jun 23, 8:07 PM
kib updated the diff for D11306: Document MAP_GUARD..

Edit the description of an error from MAP_GUARD.

Fri, Jun 23, 8:04 PM
kib committed rS320293: MFC r320038:.
MFC r320038:
Fri, Jun 23, 7:04 PM
kib committed rS320290: MFC r320038:.
MFC r320038:
Fri, Jun 23, 6:58 PM
kib added a comment to D11305: Increase the pageout cluster size to 32 pages.
In D11305#234511, @alc wrote:
In D11305#234137, @kib wrote:

The rewrite of the swap pager to use radix trie for swblock tracking is still somewhere in my repo.

I'd like to find a way to reduce the amount of physical memory that will be consumed by the leaves. Otherwise, I like it.

Fri, Jun 23, 6:24 PM
kib updated the diff for D11306: Document MAP_GUARD..

Move note about MAP_STACK use of MAP_GUARD to MAP_STACK block.
Document EINVAL on invalid MAP_GUARD call, explicitely listing the requirements.

Fri, Jun 23, 6:05 PM
kib added inline comments to D11321: loader: chain load relocate data declaration is bad.
Fri, Jun 23, 10:18 AM

Thu, Jun 22

kib updated the diff for D11306: Document MAP_GUARD..

Ed' note.

Thu, Jun 22, 6:22 PM
kib accepted D11305: Increase the pageout cluster size to 32 pages.

The rewrite of the swap pager to use radix trie for swblock tracking is still somewhere in my repo.

Thu, Jun 22, 6:10 PM
kib updated the diff for D11306: Document MAP_GUARD..

I combined the feedback, selecting the parts from each. I also added half-sentence mentioning that stacks are not mapped fully in advance.

Thu, Jun 22, 6:08 PM
kib created D11306: Document MAP_GUARD..
Thu, Jun 22, 5:15 PM
kib accepted D11302: retire arm64 kernel module linker workaround.
Thu, Jun 22, 2:53 PM
kib committed rS320231: MFC r319975:.
MFC r319975:
Thu, Jun 22, 11:04 AM

Wed, Jun 21

kib committed rS320202: Call pmap_copy() only for map entries which have the backing object.
Call pmap_copy() only for map entries which have the backing object
Wed, Jun 21, 6:54 PM
kib committed rS320201: Assert that the protection of a new map entry is a subset of the max.
Assert that the protection of a new map entry is a subset of the max
Wed, Jun 21, 6:51 PM

Tue, Jun 20

kib accepted D11185: Forward compatibility for ino64..

This looks fine to me, with the note I did in the previous accept action.

Tue, Jun 20, 7:06 PM
kib added inline comments to D11185: Forward compatibility for ino64..
Tue, Jun 20, 5:04 PM
kib committed rS320147: MFC r319916:.
MFC r319916:
Tue, Jun 20, 3:03 PM
kib added inline comments to D11113: Intel SGX driver.
Tue, Jun 20, 2:01 PM
kib committed rS320144: MFC r319916:.
MFC r319916:
Tue, Jun 20, 12:40 PM
kib accepted D11185: Forward compatibility for ino64..

Overall, I have only one big issue with this patch, ignoring the small nits I noted inline.

Tue, Jun 20, 11:26 AM
kib added a comment to D11270: Correct fences for sys/refcount.h.

Are the talk(s) you referenced the ones from cppcon?

edit: looks like this is covered in the last ~15 minutes of https://www.youtube.com/watch?v=M15UKpNlpeM

Tue, Jun 20, 10:29 AM

Mon, Jun 19

kib committed rS320125: Fix batched unload for DMAR busdma in qi mode..
Fix batched unload for DMAR busdma in qi mode.
Mon, Jun 19, 9:49 PM
kib committed rS320121: Ignore the P_SYSTEM process flag, and do not request.
Ignore the P_SYSTEM process flag, and do not request
Mon, Jun 19, 8:41 PM
kib updated the diff for D11270: Correct fences for sys/refcount.h.

Provide the release fence in refcount_release(), instead of relying on user handling it correctly.

Mon, Jun 19, 3:50 PM
kib added a comment to D11270: Correct fences for sys/refcount.h.

In fact, we can add atomic_fence_rel() before the atomic_fetchadd_int() line for the same effect.

Mon, Jun 19, 3:41 PM
kib added inline comments to D11270: Correct fences for sys/refcount.h.
Mon, Jun 19, 3:30 PM
kib committed rS320108: Allow negative aio_offset only for the read and write LIO ops on.
Allow negative aio_offset only for the read and write LIO ops on
Mon, Jun 19, 3:17 PM
kib closed D11266: Allow negative aio_offset only for the read and write LIO ops on device nodes. by committing rS320108: Allow negative aio_offset only for the read and write LIO ops on.
Mon, Jun 19, 3:17 PM
kib created D11270: Correct fences for sys/refcount.h.
Mon, Jun 19, 9:52 AM
kib committed rS320094: MFC r320035:.
MFC r320035:
Mon, Jun 19, 7:15 AM
kib committed rS320092: MFC r320034:.
MFC r320034:
Mon, Jun 19, 6:48 AM

Sun, Jun 18

kib created D11266: Allow negative aio_offset only for the read and write LIO ops on device nodes..
Sun, Jun 18, 8:28 PM
kib accepted D11247: Change blist_alloc()'s allocation policy from first-fit to next-fit.
Sun, Jun 18, 2:34 PM
kib accepted D11250: add -znotext to kernel module link invocation.
Sun, Jun 18, 1:59 PM

Sat, Jun 17

kib committed rS320057: MFC r319539:.
MFC r319539:
Sat, Jun 17, 5:11 PM
kib created D11246: Make stdio cancel-safe..
Sat, Jun 17, 11:54 AM
kib committed rS320052: Do not leak syslog_mutex on cancellation..
Do not leak syslog_mutex on cancellation.
Sat, Jun 17, 11:30 AM
kib committed rS320051: Correct translations between abridged and full x87 tags..
Correct translations between abridged and full x87 tags.
Sat, Jun 17, 11:25 AM
kib committed rD50374: Document __FreeBSD_version 1200032, groff removal (bapt),.
Document __FreeBSD_version 1200032, groff removal (bapt),
Sat, Jun 17, 1:21 AM
kib committed rS320046: Add UPDATING note about kevent(2) ABI change..
Add UPDATING note about kevent(2) ABI change.
Sat, Jun 17, 1:09 AM
kib committed rS320045: Bump __FreeBSD_version for r320043, struct event 64-bit data..
Bump __FreeBSD_version for r320043, struct event 64-bit data.
Sat, Jun 17, 1:07 AM
kib committed rS320044: Regen..
Regen.
Sat, Jun 17, 12:58 AM
kib committed rS320043: Add abstime kqueue(2) timers and expand struct kevent members..
Add abstime kqueue(2) timers and expand struct kevent members.
Sat, Jun 17, 12:57 AM
kib closed D11025: Extend struct kevent fields and add absolute timers. by committing rS320043: Add abstime kqueue(2) timers and expand struct kevent members..
Sat, Jun 17, 12:57 AM

Fri, Jun 16

kib committed rS320038: Style..
Style.
Fri, Jun 16, 11:41 PM
kib committed rS320035: Move the description of kern.kq_calloutmax sysctl into a new paragraph.
Move the description of kern.kq_calloutmax sysctl into a new paragraph
Fri, Jun 16, 11:25 PM
kib committed rS320034: Start a new sentence on the new line..
Start a new sentence on the new line.
Fri, Jun 16, 11:17 PM
kib added a comment to D11025: Extend struct kevent fields and add absolute timers..

Having absolute timer that expires relative to the CLOCK_MONOTONIC solves both issues cleanly and elegantly.

Fri, Jun 16, 12:25 PM

Thu, Jun 15

kib added a comment to D11025: Extend struct kevent fields and add absolute timers..

@kib how difficult is it to also implement real NOTE_MONOTONIC, i.e. event that expires in monotonic time from the getgo? I think this would be very useful for real-world applications that care about precise timekeeping. We use CLOCK_MONOTONIC pretty much all across our code here. Yes, it can be added later on, but without it the API is somewhat incomplete IMHO.

Thu, Jun 15, 8:06 PM
kib added inline comments to D11185: Forward compatibility for ino64..
Thu, Jun 15, 4:32 PM
kib committed rS319976: MFC r319869:.
MFC r319869:
Thu, Jun 15, 3:09 PM
kib committed rS319975: Some minor improvements to vnode_pager_generic_putpages()..
Some minor improvements to vnode_pager_generic_putpages().
Thu, Jun 15, 2:34 PM

Wed, Jun 14

kib added inline comments to D11185: Forward compatibility for ino64..
Wed, Jun 14, 3:05 PM
kib added inline comments to D11185: Forward compatibility for ino64..
Wed, Jun 14, 5:28 AM
kib committed rS319937: MFC r319825:.
MFC r319825:
Wed, Jun 14, 5:12 AM

Tue, Jun 13

kib committed rS319916: Remove stray return..
Remove stray return.
Tue, Jun 13, 7:02 PM
kib accepted D11146: Improve hint maintenance on allocation.
Tue, Jun 13, 12:54 PM

Mon, Jun 12

kib added inline comments to D11025: Extend struct kevent fields and add absolute timers..
Mon, Jun 12, 9:36 PM
kib committed rS319876: Update scescx test to print syscall number and arguments..
Update scescx test to print syscall number and arguments.
Mon, Jun 12, 9:16 PM
kib committed rS319875: Add ptrace(PT_GET_SC_ARGS) command to return debuggee' current syscall.
Add ptrace(PT_GET_SC_ARGS) command to return debuggee' current syscall
Mon, Jun 12, 9:15 PM
kib closed D11080: Move syscall_args into struct thread. by committing rS319875: Add ptrace(PT_GET_SC_ARGS) command to return debuggee' current syscall.
Mon, Jun 12, 9:15 PM
kib committed rS319874: Print unimplemented syscall number to the ctty on SIGSYS, if enabled.
Print unimplemented syscall number to the ctty on SIGSYS, if enabled
Mon, Jun 12, 9:11 PM
kib committed rS319873: Move struct syscall_args syscall arguments parameters container into.
Move struct syscall_args syscall arguments parameters container into
Mon, Jun 12, 9:03 PM
kib committed rS319871: Make struct syscall_args visible to userspace compilation environment.
Make struct syscall_args visible to userspace compilation environment
Mon, Jun 12, 8:54 PM
kib committed rS319869: Decode recently added flags..
Decode recently added flags.
Mon, Jun 12, 8:42 PM
kib updated the diff for D11080: Move syscall_args into struct thread..

Update patch to the current state, including fixes to PT_GET_SC_ARGS (it passes tests) and to scescx.c test itself.

Mon, Jun 12, 8:37 PM
kib updated the diff for D11025: Extend struct kevent fields and add absolute timers..

Fix make universe. Changes are mostly in tests, in particular, NetBSD tests customizations are no longer needed because event.data becomes 64bit.

Mon, Jun 12, 6:46 PM
kib added a comment to D11146: Improve hint maintenance on allocation.
In D11146#230856, @alc wrote:

I'd like to add something else to this patch. Specifically, I'd like to remove the field "bl_free" from struct blmeta. While it is correctly maintained, it is never used. Moreover, unless the entire tree is a single leaf, the field "u.bmu_avail" at the root provides the count of free blocks in the tree.

Mon, Jun 12, 4:51 PM
kib added inline comments to D11146: Improve hint maintenance on allocation.
Mon, Jun 12, 12:04 PM

Sun, Jun 11

kib added a comment to D11113: Intel SGX driver.

This is not a review, I did not read the code for real.
Still, I noticed several somewhat big issues, with the object relock in page fault handler, with unused newbus attachment, and with headers.
I also noticed a lot of style bugs, for which I mostly pointed out one occurrence, to not make notes unreadable. Please fix them all.

Sun, Jun 11, 5:10 PM
kib updated the diff for D11080: Move syscall_args into struct thread..

Document PT_GET_SC_ARGS.

Sun, Jun 11, 2:59 PM
kib committed rS319825: More accurately handle early EFER restoration on resume..
More accurately handle early EFER restoration on resume.
Sun, Jun 11, 2:39 PM