kib (Konstantin Belousov)
User

Projects

User Details

User Since
May 16 2014, 7:35 PM (169 w, 6 d)

Recent Activity

Yesterday

kib added inline comments to D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..
Thu, Aug 17, 4:20 PM
kib committed rS322617: MFC r322493:.
MFC r322493:
Thu, Aug 17, 11:36 AM
kib committed rS322616: MFC r322494:.
MFC r322494:
Thu, Aug 17, 10:56 AM
kib committed rS322615: MFC r322493:.
MFC r322493:
Thu, Aug 17, 10:49 AM
kib accepted D11801: Discover CPU topology on multi-die AMD Zen systems.
Thu, Aug 17, 10:45 AM

Wed, Aug 16

kib added a comment to D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..
In D11435#249799, @alc wrote:

By the way, once upon a time, you asked me about whether we should continue using the blist allocator for swap space or switch to vmem. Perhaps the most compelling reason, which I failed to give at the time, is that every allocated range from an arena consumes a boundary tag. In other words, while vmem coalesces free ranges, it maintains a boundary tag for each allocated range.

Ok.

Wed, Aug 16, 3:54 PM
kib updated the diff for D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..

Edit comments, use consistent loop iteration construct.

Wed, Aug 16, 3:18 PM
kib added inline comments to D11943: Modify vm_page_wire() to not dequeue the specified page.
Wed, Aug 16, 9:00 AM

Tue, Aug 15

kib added a comment to D11866: Allow vt_timer() callout to run during mountroot input loop in vt(4).

However, we can't call DROP/PICKUP because they have to be in the same block.

I don't follow what you mean here.

Tue, Aug 15, 9:02 PM
kib committed rS322550: Typo, the '-6' option selects inet6..
Typo, the '-6' option selects inet6.
Tue, Aug 15, 7:29 PM
kib committed rS322543: MFC r322256:.
MFC r322256:
Tue, Aug 15, 2:22 PM
kib accepted D12035: loader.efi: repace XXX with real comments in trap.c.
Tue, Aug 15, 1:56 PM
kib accepted D11688: Add OBJ_PG_DTOR flag to VM object.
Tue, Aug 15, 12:10 PM
kib accepted D11113: Intel SGX driver.
Tue, Aug 15, 10:46 AM
kib added inline comments to D11113: Intel SGX driver.
Tue, Aug 15, 10:06 AM
kib added a comment to D12010: Support armv7 builds for userland.

You may want to take a look at lib/libc/arm/sys/__vdso_gettc.c and at least use proper asm if possible, instead of the hardcoded instruction byte encoding, if possible after the patch.

Tue, Aug 15, 9:59 AM
kib accepted D11984: Add vm_page_alloc_after().
Tue, Aug 15, 9:51 AM
kib committed rS322534: MFC r322456:.
MFC r322456:
Tue, Aug 15, 9:44 AM
kib committed rS322533: MFC r322426:.
MFC r322426:
Tue, Aug 15, 9:41 AM
kib committed rS322532: MFC r322256:.
MFC r322256:
Tue, Aug 15, 9:36 AM

Mon, Aug 14

kib added inline comments to D11113: Intel SGX driver.
Mon, Aug 14, 2:55 PM
kib updated the summary of D12023: Make WRFSBASE and WRGSBASE functional..
Mon, Aug 14, 12:43 PM
kib updated the summary of D12023: Make WRFSBASE and WRGSBASE functional..
Mon, Aug 14, 12:22 PM
kib created D12023: Make WRFSBASE and WRGSBASE functional..
Mon, Aug 14, 11:42 AM
kib committed rS322496: Print whole machine state on double fault..
Print whole machine state on double fault.
Mon, Aug 14, 11:23 AM
kib committed rS322495: Add {rd,wr}{fs,gs}base C wrappers for instructions..
Add {rd,wr}{fs,gs}base C wrappers for instructions.
Mon, Aug 14, 11:21 AM
kib committed rS322494: Style..
Style.
Mon, Aug 14, 11:20 AM
kib committed rS322493: Remove confusion in the line explaining syntax of the msr read..
Remove confusion in the line explaining syntax of the msr read.
Mon, Aug 14, 11:19 AM
kib committed rS322492: MFC r322175:.
MFC r322175:
Mon, Aug 14, 11:06 AM
kib committed rS322491: MFC r322171:.
MFC r322171:
Mon, Aug 14, 11:04 AM

Sun, Aug 13

kib committed rS322456: Fix a regression in r321608..
Fix a regression in r321608.
Sun, Aug 13, 2:42 PM
kib added inline comments to D11113: Intel SGX driver.
Sun, Aug 13, 12:19 PM

Sat, Aug 12

kib added inline comments to D11113: Intel SGX driver.
Sat, Aug 12, 4:55 PM
kib committed rS322427: Improve standard compliance for memset_s() and abort_handler_s()..
Improve standard compliance for memset_s() and abort_handler_s().
Sat, Aug 12, 3:18 PM
kib closed D11991: memset_s() and abort_handler_s() missed some standard notes by committing rS322427: Improve standard compliance for memset_s() and abort_handler_s()..
Sat, Aug 12, 3:18 PM
kib committed rS322426: Fix indent..
Fix indent.
Sat, Aug 12, 2:58 PM
kib accepted D11991: memset_s() and abort_handler_s() missed some standard notes.
Sat, Aug 12, 2:35 PM
kib added inline comments to D11991: memset_s() and abort_handler_s() missed some standard notes.
Sat, Aug 12, 2:05 PM
kib added inline comments to D11991: memset_s() and abort_handler_s() missed some standard notes.
Sat, Aug 12, 9:17 AM
kib committed rS322422: MFC r322077:.
MFC r322077:
Sat, Aug 12, 8:52 AM
kib added inline comments to D11984: Add vm_page_alloc_after().
Sat, Aug 12, 8:50 AM

Fri, Aug 11

kib accepted D11984: Add vm_page_alloc_after().

I think that the explanation from the revision' description should be added as a comment.

Fri, Aug 11, 6:12 PM
kib committed rS322399: MFC r322059:.
MFC r322059:
Fri, Aug 11, 11:38 AM
kib accepted D11942: Have vm_page_grab_pages() support VM_ALLOC_NOWAIT.
Fri, Aug 11, 11:37 AM
kib added inline comments to D11943: Modify vm_page_wire() to not dequeue the specified page.
Fri, Aug 11, 11:35 AM
kib committed rS322396: MFC r322059:.
MFC r322059:
Fri, Aug 11, 10:08 AM
kib committed rS322395: MFC r322050:.
MFC r322050:
Fri, Aug 11, 10:06 AM

Thu, Aug 10

kib added inline comments to D11943: Modify vm_page_wire() to not dequeue the specified page.
Thu, Aug 10, 12:01 PM
kib committed rS322346: MFC r321607:.
MFC r321607:
Thu, Aug 10, 9:02 AM
kib committed rS322345: MFC r321608:.
MFC r321608:
Thu, Aug 10, 9:00 AM
kib accepted D11945: Micro-optimize kmem_unback()..
Thu, Aug 10, 8:47 AM
kib accepted D11946: Make vm_page_sunbusy() assert that the page is not locked.
Thu, Aug 10, 8:45 AM

Wed, Aug 9

kib committed rS322303: MFC r321919:.
MFC r321919:
Wed, Aug 9, 9:16 AM
kib committed rS322301: MFC r321919:.
MFC r321919:
Wed, Aug 9, 8:57 AM

Tue, Aug 8

kib accepted D11926: Add vm_page_grab_pages().
Tue, Aug 8, 4:40 PM
kib committed rS322256: Fix logic error in the the assert, causing the condition to be always true..
Fix logic error in the the assert, causing the condition to be always true.
Tue, Aug 8, 3:46 PM
kib accepted D11919: Remove useless adev adjustment.

I see this was already committed.

Tue, Aug 8, 8:10 AM

Mon, Aug 7

kib committed rS322175: Avoid DI recursion when reclaim_pv_chunk() is called from.
Avoid DI recursion when reclaim_pv_chunk() is called from
Mon, Aug 7, 5:30 PM
kib committed rS322171: Explain why delayed invalidation is not required in pmap_protect() and.
Explain why delayed invalidation is not required in pmap_protect() and
Mon, Aug 7, 5:23 PM
kib accepted D11913: x86: bump MAX_APIC_ID to 512.
Mon, Aug 7, 11:09 AM
kib accepted D11912: x86: make the arrays that depend on MAX_APIC_ID dynamic.
Mon, Aug 7, 11:09 AM
kib accepted D11911: apic_enumerator: only set mp_ncpus and mp_maxid at probe cpus phase.
Mon, Aug 7, 11:09 AM

Sun, Aug 6

kib committed rS322118: Remove trivial comments. Remove and-ing with UINT_MAX for minor(),.
Remove trivial comments. Remove and-ing with UINT_MAX for minor(),
Sun, Aug 6, 12:27 PM
kib committed rS322117: MFC r321730:.
MFC r321730:
Sun, Aug 6, 10:07 AM

Sat, Aug 5

kib committed rS322077: Provide more detailed specification for major(), minor() and makedev()..
Provide more detailed specification for major(), minor() and makedev().
Sat, Aug 5, 7:52 AM
kib added inline comments to D11826: Make it possible to use any executable as init..
Sat, Aug 5, 7:38 AM

Fri, Aug 4

kib committed rS322059: Fix off by one in calculation of the number of buckets for the pc.
Fix off by one in calculation of the number of buckets for the pc
Fri, Aug 4, 6:03 PM
kib closed D11864: Fix off by one in calculation of the number of buckets for the pc addresses. by committing rS322059: Fix off by one in calculation of the number of buckets for the pc.
Fri, Aug 4, 6:03 PM
kib created D11864: Fix off by one in calculation of the number of buckets for the pc addresses..
Fri, Aug 4, 1:50 PM
kib added a comment to D11826: Make it possible to use any executable as init..

If your init is broken, your theoretical wrapper would probably be broken in the exact same way. It's just another moving part that can cause problems. Moving it into the kernel, on the other hand, is intuitive to the user.

Fri, Aug 4, 10:07 AM
kib committed rS322050: Relax visibility for some termios symbols..
Relax visibility for some termios symbols.
Fri, Aug 4, 9:45 AM
kib added inline comments to D11819: Allow blist allocations to span leaf boundaries, but not meta boundaries.
Fri, Aug 4, 8:47 AM
kib committed rS322048: MFC r321349:.
MFC r321349:
Fri, Aug 4, 8:27 AM
kib committed rS322047: MFC r321348:.
MFC r321348:
Fri, Aug 4, 8:26 AM
kib committed rS322046: MFC r321347:.
MFC r321347:
Fri, Aug 4, 8:25 AM
kib committed rS322045: MFC r321349:.
MFC r321349:
Fri, Aug 4, 8:20 AM
kib committed rS322044: MFC r321348:.
MFC r321348:
Fri, Aug 4, 8:18 AM
kib committed rS322043: MFC r321347:.
MFC r321347:
Fri, Aug 4, 8:16 AM
kib committed rS322042: MFC r321652:.
MFC r321652:
Fri, Aug 4, 8:12 AM

Thu, Aug 3

kib accepted D11601: Make p1003_1b.aio_listio_max a tunable.

I am fine with the move of all aio-related sysctls.

Thu, Aug 3, 4:12 PM
kib added a comment to D11826: Make it possible to use any executable as init..

I agree that a separate application would be more elegant. However, it would make it more complicated from the user point of view - the approach I took doesn't require the user to learn anything new; the init_path option existed since forever, and there are no special userspace tools to remember about. Also, we kind of already have that: it should be possible to get init(8) to do exactly that, although I don't exactly know how.

Get init to do what ?

Thu, Aug 3, 12:13 PM
kib committed rS321978: MFC r321627:.
MFC r321627:
Thu, Aug 3, 7:30 AM
kib committed rS321977: MFC r321625:.
MFC r321625:
Thu, Aug 3, 7:29 AM
kib updated the diff for D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..

Update comments.

Thu, Aug 3, 7:23 AM

Wed, Aug 2

kib added a comment to D11826: Make it possible to use any executable as init..

I would suggest a different approach for this, since ability to run a shell without configuring tty port in init in single-user mode is sometimes useful. Create a program which opens /dev/console for fds 0,1,2 and execs /bin/sh or whatever specified by some kenv var.

Wed, Aug 2, 8:29 PM
kib added inline comments to D11826: Make it possible to use any executable as init..
Wed, Aug 2, 8:19 PM
kib committed rS321930: For makedev(), cast the minor argument to unsigned type explicitely,.
For makedev(), cast the minor argument to unsigned type explicitely,
Wed, Aug 2, 2:55 PM
kib committed rS321920: Change major()/minor() to work with 64bit dev_t..
Change major()/minor() to work with 64bit dev_t.
Wed, Aug 2, 10:14 AM
kib committed rS321919: Do not call trapsignal() after handling usermode fault or interrupt,.
Do not call trapsignal() after handling usermode fault or interrupt,
Wed, Aug 2, 10:12 AM
kib accepted D11789: Implement the ksyms mmap interface using d_mmap_single..
Wed, Aug 2, 8:12 AM
kib committed rS321910: MFC r321581:.
MFC r321581:
Wed, Aug 2, 8:07 AM
kib committed rS321909: MFC r321580:.
MFC r321580:
Wed, Aug 2, 8:03 AM
kib committed rS321908: MFC r321512:.
MFC r321512:
Wed, Aug 2, 7:59 AM
kib updated the diff for D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..

Rearrange loops. Update comment.

Wed, Aug 2, 7:46 AM

Tue, Aug 1

kib accepted D11808: Keep top page on CloudABI to work around AMD Ryzen stability issues..

IMO this is an overreaction. For non-buggy CPUs, I do not see why do you want this.

Tue, Aug 1, 7:58 PM
kib accepted D11780: Lower shared page for amd64 on Ryzen to work around bug with code near top of user space.
Tue, Aug 1, 4:16 PM
kib added a comment to D11435: Replace global swhash in swap pager with per-object trie to track swap blocks assigned to the object pages..
In D11435#243818, @alc wrote:

Yes, let's try this patched version on a swap-intensive workload.

Tue, Aug 1, 2:15 PM
kib added inline comments to D11113: Intel SGX driver.
Tue, Aug 1, 1:31 PM
kib added a comment to D11801: Discover CPU topology on multi-die AMD Zen systems.

Please do not consider my note below as an objection against the patch, rather it is a generic question. As a side effect, it might affect the patch.

Tue, Aug 1, 1:27 PM

Mon, Jul 31

kib added inline comments to D11113: Intel SGX driver.
Mon, Jul 31, 8:49 PM