kib (Konstantin Belousov)
User

Projects

User Details

User Since
May 16 2014, 7:35 PM (153 w, 3 d)

Recent Activity

Today

kib committed rS317408: MFC r316698:.
MFC r316698:
Tue, Apr 25, 1:17 PM
kib committed rS317407: MFC r316698:.
MFC r316698:
Tue, Apr 25, 1:07 PM

Yesterday

kib added a comment to D10469: The c11 appendix K string functions.
In D10469#217188, @ed wrote:

That sounds reasonable. A couple of random remarks from my side:

  1. If implemented as a separate library, maybe it would make more sense to maintain it outside the base system? What about making it a separate package and placing it in the ports tree? That way it can also be packaged easily by other operating systems.

I do not have strong opinion there. Probably the argument pro inclusion into the base is the same as for libstdthreads, which was added solely (?) because it is in C standard, and which specification was criticized.

Mon, Apr 24, 8:41 AM
kib committed rS317354: MFC r316679:.
MFC r316679:
Mon, Apr 24, 7:52 AM
kib added a comment to D10469: The c11 appendix K string functions.

This adds significant portion of the functions from the appendix. So there are two points I have in mind about the stuff:

Mon, Apr 24, 7:44 AM

Sun, Apr 23

kib committed rS317342: MFC r316213:.
MFC r316213:
Sun, Apr 23, 8:33 PM

Sat, Apr 22

kib accepted D10449: Add a new GDB_LIBEXEC option to install gdb and kgdb to /usr/libexec..

I am fine with /usr/libexec/{,k}gdb, and I agree with the changes in whole. I cannot provide useful comments on the details of the make infrastructure changes, except what Ed noted.

Sat, Apr 22, 5:21 AM

Fri, Apr 21

kib added a comment to D10439: ino64.

I deliberately refuse to make any gratuitous style changes in the copied files. They have to be kept in sync with the master copies.

Fri, Apr 21, 2:11 PM
kib committed rS317250: MFC r316852:.
MFC r316852:
Fri, Apr 21, 10:16 AM
kib committed rS317249: MFC r316852:.
MFC r316852:
Fri, Apr 21, 10:13 AM
kib committed rS317248: MFC r316851:.
MFC r316851:
Fri, Apr 21, 10:11 AM

Thu, Apr 20

kib created D10439: ino64.
Thu, Apr 20, 5:55 PM
kib committed rS317196: Write-combine framebuffer writes through user-space mappings, if possible..
Write-combine framebuffer writes through user-space mappings, if possible.
Thu, Apr 20, 3:18 PM
kib added inline comments to D10435: x86: Add dynamic interrupt rebalancing.
Thu, Apr 20, 10:04 AM

Wed, Apr 19

kib committed rS317134: MFC r316532:.
MFC r316532:
Wed, Apr 19, 11:13 AM
kib committed rS317133: MFC r316531:.
MFC r316531:
Wed, Apr 19, 11:12 AM
kib committed rS317132: MFC r316529:.
MFC r316529:
Wed, Apr 19, 11:10 AM
kib committed rS317131: MFC r316739:.
MFC r316739:
Wed, Apr 19, 10:58 AM
kib committed rS317130: MFC r316739:.
MFC r316739:
Wed, Apr 19, 10:54 AM

Tue, Apr 18

kib added a comment to D10170: Capsicumize cpuset_*.

compat32 capabilities.conf requires the same changes, and corresponding tables must be regenerated as well, in the follow-up commit.

Tue, Apr 18, 11:46 AM

Fri, Apr 14

kib committed rS316852: In fsck_ffs pass1, prevent the inosused variable from wrapping..
In fsck_ffs pass1, prevent the inosused variable from wrapping.
Fri, Apr 14, 3:22 PM
kib committed rS316851: Correct calculation of the entry->free_down in the invariants-checking.
Correct calculation of the entry->free_down in the invariants-checking
Fri, Apr 14, 3:16 PM

Thu, Apr 13

kib closed D10339: Do not loose dirty bits for removing PROT_WRITE on arm64..
Thu, Apr 13, 6:06 PM
kib committed rS316767: Map DMAP as nx..
Map DMAP as nx.
Thu, Apr 13, 3:50 PM
kib accepted D10385: Remove many write-only variables from kernel.

I did not found anything obviously wrong from an eye inspection.

Thu, Apr 13, 3:31 PM
kib accepted D10382: Set the arm64 Execute-never bits in more places..
Thu, Apr 13, 12:45 PM
kib accepted D10371: Add support for Privileged Access Never (PAN).
In D10371#214953, @kib wrote:

Do we need to set SCTLR_EL1.SPAN to 1 ?

No, we need to clear it so PSTATE.PAN is set to 1 on exception entry.

Thu, Apr 13, 12:42 PM
kib committed rS316753: MFC r316566:.
MFC r316566:
Thu, Apr 13, 8:21 AM

Wed, Apr 12

kib added a comment to D10371: Add support for Privileged Access Never (PAN).

Do we need to set SCTLR_EL1.SPAN to 1 ?

Wed, Apr 12, 10:02 PM
kib committed rS316739: Report _SC_SEM_NSEMS_MAX and _SC_SEM_VALUE_MAX which show parameters.
Report _SC_SEM_NSEMS_MAX and _SC_SEM_VALUE_MAX which show parameters
Wed, Apr 12, 7:27 PM
kib accepted D9674: Handle ENXIO bufs..
Wed, Apr 12, 9:26 AM
kib committed rS316727: MFC r316528:.
MFC r316528:
Wed, Apr 12, 9:22 AM
kib committed rS316726: MFC r316526:.
MFC r316526:
Wed, Apr 12, 9:20 AM
kib committed rS316725: MFC r316525:.
MFC r316525:
Wed, Apr 12, 9:13 AM
kib committed rS316724: MFC r316524:.
MFC r316524:
Wed, Apr 12, 9:11 AM

Tue, Apr 11

kib accepted D10360: Use the execute-never bits on arm64.
Tue, Apr 11, 4:52 PM
kib committed rS316698: Remove debugging printf..
Remove debugging printf.
Tue, Apr 11, 8:29 AM

Mon, Apr 10

kib accepted D10349: Busy the map in vm_map_protect().

Thank you.

Mon, Apr 10, 7:49 PM
kib accepted D10349: Busy the map in vm_map_protect().
Mon, Apr 10, 6:52 PM
kib committed rS316679: Do not lose dirty bits for removing PROT_WRITE on arm64..
Do not lose dirty bits for removing PROT_WRITE on arm64.
Mon, Apr 10, 3:32 PM
kib added a comment to D10339: Do not loose dirty bits for removing PROT_WRITE on arm64..
In D10339#214208, @alc wrote:

pmap_remove() needs the same fix.

Mon, Apr 10, 2:55 PM
kib updated the summary of D10339: Do not loose dirty bits for removing PROT_WRITE on arm64..
Mon, Apr 10, 7:21 AM
kib created D10339: Do not loose dirty bits for removing PROT_WRITE on arm64..
Mon, Apr 10, 7:18 AM

Sat, Apr 8

kib accepted D10313: rendevous -> rendezvous.

Note that the change would affect KBI/drivers, at least because counters(9). If merging to stable branches, you would need to keep compat symbol for rendevous name.

Sat, Apr 8, 7:00 AM

Thu, Apr 6

kib committed rS316576: Improvements for the brand detection and prioritization..
Improvements for the brand detection and prioritization.
Thu, Apr 6, 3:19 PM
kib committed rS316566: Remove spl*() calls from the nfsclient code. Style adjustments in the.
Remove spl*() calls from the nfsclient code. Style adjustments in the
Thu, Apr 6, 12:45 PM

Wed, Apr 5

kib committed rS316532: Make nfs pageout coherent with the dirty state of the buffers..
Make nfs pageout coherent with the dirty state of the buffers.
Wed, Apr 5, 5:26 PM
kib closed D10241: Make nfs pageout coherent with the buffer cache. by committing rS316532: Make nfs pageout coherent with the dirty state of the buffers..
Wed, Apr 5, 5:26 PM
kib committed rS316531: Handle nfs IO_ASYNC write requests asynchronously..
Handle nfs IO_ASYNC write requests asynchronously.
Wed, Apr 5, 5:20 PM
kib committed rS316529: Handle possible vnode reclamation after ncl_vinvalbuf() call..
Handle possible vnode reclamation after ncl_vinvalbuf() call.
Wed, Apr 5, 5:12 PM
kib committed rS316528: Add V_VMIO flag for vinvalbuf(9) to indicate that the flush request.
Add V_VMIO flag for vinvalbuf(9) to indicate that the flush request
Wed, Apr 5, 4:58 PM
kib committed rS316526: Extract calculation of ioflags from the vm_pager_putpages flags into a.
Extract calculation of ioflags from the vm_pager_putpages flags into a
Wed, Apr 5, 4:56 PM
kib committed rS316525: Some style fixes for vnode_pager_generic_putpages(), in the local.
Some style fixes for vnode_pager_generic_putpages(), in the local
Wed, Apr 5, 4:45 PM
kib committed rS316524: Use int instead of boolean_t for flags argument type in.
Use int instead of boolean_t for flags argument type in
Wed, Apr 5, 4:30 PM

Tue, Apr 4

kib updated the diff for D10241: Make nfs pageout coherent with the buffer cache..

Fix flags member for VI_DOOMED, remove np != NULL chunk.

Tue, Apr 4, 5:17 PM
kib added inline comments to D10241: Make nfs pageout coherent with the buffer cache..
Tue, Apr 4, 5:16 PM
kib updated the diff for D10241: Make nfs pageout coherent with the buffer cache..

Handle markj feedback: more accurate handling of reclaim in several places, rename sync to flags.

Tue, Apr 4, 6:58 AM

Mon, Apr 3

kib committed rS316449: MFC r316011:.
MFC r316011:
Mon, Apr 3, 9:42 AM
kib committed rS316448: MFC r316009:.
MFC r316009:
Mon, Apr 3, 9:36 AM
kib created D10241: Make nfs pageout coherent with the buffer cache..
Mon, Apr 3, 9:23 AM

Sun, Apr 2

kib committed rS316392: MFC r315968:.
MFC r315968:
Sun, Apr 2, 7:11 AM

Sat, Apr 1

kib committed rS316346: MFC r315934:.
MFC r315934:
Sat, Apr 1, 8:40 AM
kib committed rS316345: MFC r315933:.
MFC r315933:
Sat, Apr 1, 8:37 AM
kib committed rS316344: MFC r315932:.
MFC r315932:
Sat, Apr 1, 8:35 AM

Fri, Mar 31

kib committed rS316335: MFC r309551:.
MFC r309551:
Fri, Mar 31, 3:48 PM
kib added a comment to D10156: use counter(9) in vmmeter.

Kostik is right. I lost counter here. The problem is that sparc64 doesn't have a C part of interrupt handler. It even hardcodes into assembly offset of ih_handler in the structure struct intr_handler. I could resurrect counter in assembly, but it seems to me correct way is to create C function, called from assembly, just like other arches do. Can anyone provide help with sparc64 assembly?

Fri, Mar 31, 9:50 AM

Thu, Mar 30

kib committed rS316258: Only activate __EXT1_VISIBLE block when using sys/errno.h in userspace..
Only activate __EXT1_VISIBLE block when using sys/errno.h in userspace.
Thu, Mar 30, 6:25 AM
kib committed rS316213: Implement the memset_s(3) function as specified by the C11 ISO/IEC.
Implement the memset_s(3) function as specified by the C11 ISO/IEC
Thu, Mar 30, 4:58 AM
kib closed D9903: Add C11 Appendix K function memset_s by committing rS316213: Implement the memset_s(3) function as specified by the C11 ISO/IEC.
Thu, Mar 30, 4:58 AM
kib committed rS316211: A followup to r315749, two more places where brand->interp_path was.
A followup to r315749, two more places where brand->interp_path was
Thu, Mar 30, 4:21 AM
kib added a comment to D10156: use counter(9) in vmmeter.
In D10156#210714, @lidl wrote:

'vmstat -i 1' give reasonable looking stats while the machine is idling, and running hard.

Thu, Mar 30, 4:15 AM

Wed, Mar 29

kib committed rS316135: MFC r315331:.
MFC r315331:
Wed, Mar 29, 11:04 AM
kib committed rS316134: MFC r309068 (by jhb):.
MFC r309068 (by jhb):
Wed, Mar 29, 11:00 AM
kib added inline comments to D10156: use counter(9) in vmmeter.
Wed, Mar 29, 10:37 AM
kib updated the diff for D10161: Continuation of D9903..

Use lock and free old handler pointer.

Wed, Mar 29, 10:24 AM

Tue, Mar 28

kib added inline comments to D10161: Continuation of D9903..
Tue, Mar 28, 2:02 PM
kib added a comment to D9903: Add C11 Appendix K function memset_s.

I noted that set_constrain_handler_s() is not thread-safe. Please see D10161 for my take on it and several other style editings.

Tue, Mar 28, 9:17 AM
kib created D10161: Continuation of D9903..
Tue, Mar 28, 9:15 AM
kib committed rS316073: MFC r315281:.
MFC r315281:
Tue, Mar 28, 6:08 AM

Mon, Mar 27

kib committed rS316011: Timeout DMAR commands..
Timeout DMAR commands.
Mon, Mar 27, 7:06 AM
kib committed rS316009: Fix TUNABLE_UINT64() on 32bit architectures..
Fix TUNABLE_UINT64() on 32bit architectures.
Mon, Mar 27, 6:37 AM

Sun, Mar 26

kib accepted D10143: Add kern_mincore() and use it in compat..

I suggest to split this into two commits: introduction of kern_mincore(), and reference to it in linuxolator. This would make the merge to stable easier.

Sun, Mar 26, 6:43 PM

Mar 26 2017

kib committed rS315972: MFC r315588:.
MFC r315588:
Mar 26 2017, 1:11 AM
kib committed rS315971: MFC r315552:.
MFC r315552:
Mar 26 2017, 12:59 AM
kib committed rS315970: MFC r315588:.
MFC r315588:
Mar 26 2017, 12:56 AM
kib committed rS315969: MFC r315586:.
MFC r315586:
Mar 26 2017, 12:53 AM
kib committed rS315968: Provide less laborius way to enable busdma DMAR to only short list of devices..
Provide less laborius way to enable busdma DMAR to only short list of devices.
Mar 26 2017, 12:41 AM

Mar 25 2017

kib committed rS315934: Avoid leaking allocated but unused context after creation race..
Avoid leaking allocated but unused context after creation race.
Mar 25 2017, 10:47 AM
kib committed rS315933: Do not create RMRR entries for identity-mapped domains..
Do not create RMRR entries for identity-mapped domains.
Mar 25 2017, 10:45 AM
kib committed rS315932: Slight cleanup of the comment..
Slight cleanup of the comment.
Mar 25 2017, 10:42 AM
kib accepted D9903: Add C11 Appendix K function memset_s.

Only style issues are left.

Mar 25 2017, 10:18 AM

Mar 24 2017

kib added a comment to D9903: Add C11 Appendix K function memset_s.

The remaining issues are minor, what do you plan about the committing stuff ? Do you have commit bit yourself or somebody with src commit bit who is willing to commit this ?

Mar 24 2017, 10:02 PM
kib committed rS315894: MFC r315453:.
MFC r315453:
Mar 24 2017, 7:22 AM
kib committed rS315893: MFC r315453:.
MFC r315453:
Mar 24 2017, 7:15 AM

Mar 22 2017

kib committed rS315754: Update r315753 with the proper flag name..
Update r315753 with the proper flag name.
Mar 22 2017, 10:28 PM
kib committed rS315753: Add a flag BI_BRAND_ONLY_STATIC to specify that the brand only.
Add a flag BI_BRAND_ONLY_STATIC to specify that the brand only
Mar 22 2017, 10:23 PM
kib committed rS315749: Adjust r314851 to not require every brand to specify interpreter path..
Adjust r314851 to not require every brand to specify interpreter path.
Mar 22 2017, 10:07 PM
kib accepted D10100: Assert that the align parameter to uma_zcreate() is valid..
Mar 22 2017, 9:07 PM

Mar 21 2017

kib added a comment to D10067: X86 pmap_qenter needs to always invalidate.
In D10067#208488, @bz wrote:

See reference [1]

Mar 21 2017, 8:26 PM
kib accepted D10068: Support 1 and 2 byte cmpset on x86.
Mar 21 2017, 6:28 PM