Page MenuHomeFreeBSD

kib (Konstantin Belousov)
User

Projects

User Details

User Since
May 16 2014, 7:35 PM (517 w, 5 d)

Recent Activity

Yesterday

kib added a reverting change for rGf7fd7f8b88a1: acpidump: add printfield() helper: rG7cf0b203f57d: Revert "acpidump: dump AMD IVRS table describing IOMMU layout".
Wed, Apr 17, 9:42 AM
kib added a reverting change for rGf95a254aa722: acpidump: dump AMD IVRS table describing IOMMU layout: rG7cf0b203f57d: Revert "acpidump: dump AMD IVRS table describing IOMMU layout".
Wed, Apr 17, 9:42 AM
kib added a reverting change for rG256ee428f5a3: acpidump: add Foundation copyright: rG7cf0b203f57d: Revert "acpidump: dump AMD IVRS table describing IOMMU layout".
Wed, Apr 17, 9:42 AM
kib committed rG7cf0b203f57d: Revert "acpidump: dump AMD IVRS table describing IOMMU layout" (authored by kib).
Revert "acpidump: dump AMD IVRS table describing IOMMU layout"
Wed, Apr 17, 9:42 AM
kib committed rG02dc74cc21af: pciconf(8): dump AMD IOMMU Base Capability (authored by kib).
pciconf(8): dump AMD IOMMU Base Capability
Wed, Apr 17, 8:14 AM
kib committed rG6b771774c5c6: pcireg.h: Add AMD IOMMU Base Cap definitions (authored by kib).
pcireg.h: Add AMD IOMMU Base Cap definitions
Wed, Apr 17, 8:14 AM
kib committed rG6f8e8d0723cc: pcireg.h: add include guard (authored by kib).
pcireg.h: add include guard
Wed, Apr 17, 8:14 AM
kib committed rG256ee428f5a3: acpidump: add Foundation copyright (authored by kib).
acpidump: add Foundation copyright
Wed, Apr 17, 8:14 AM
kib committed rGf95a254aa722: acpidump: dump AMD IVRS table describing IOMMU layout (authored by kib).
acpidump: dump AMD IVRS table describing IOMMU layout
Wed, Apr 17, 8:14 AM
kib committed rGf7fd7f8b88a1: acpidump: add printfield() helper (authored by kib).
acpidump: add printfield() helper
Wed, Apr 17, 8:14 AM
kib committed rG84944bf5e282: acpidump: bump WARNS to 6 (authored by kib).
acpidump: bump WARNS to 6
Wed, Apr 17, 8:14 AM
kib committed rG930c76a733d0: acpidump: the acpi_get_fadt_revision() argument is unused (authored by kib).
acpidump: the acpi_get_fadt_revision() argument is unused
Wed, Apr 17, 8:14 AM
kib committed rGf77abc3526d6: acpidump: silent the warning about unaligned uuid (authored by kib).
acpidump: silent the warning about unaligned uuid
Wed, Apr 17, 8:14 AM
kib committed rGa6caa7bc5423: stop_all_proc(): skip traced or signal-stoped processes (authored by kib).
stop_all_proc(): skip traced or signal-stoped processes
Wed, Apr 17, 8:14 AM

Tue, Apr 16

kib added a comment to D43306: arm64: Initial SVE support.

Rework to not guarentee the SVE registers are saved over a syscall other than sigreturn.
The ABI doesn't require SVE registers to be saved on a function call so extend this to a system call.
As dropping the SVE state on all system calls could get expensive only do it on context switch when in one.

Tue, Apr 16, 4:20 PM
kib added a comment to D44807: arm64: Clear td_frame when returning to userspace.

Do you mean that td_frame is not constant on arm64?

Tue, Apr 16, 4:15 PM

Sun, Apr 14

kib accepted D44788: unionfs_rename: fix numerous locking issues.

Could you try to (greatly) simplify unionfs rename by using ERELOOKUP? For instance, it can be split into two essentially independent cases: 1. need to copy fdvp from lower to upper (and return ERELOOKUP) 2. Just directly call VOP_RENAME() on upper if copy is not needed.

Sun, Apr 14, 2:28 PM

Sat, Apr 13

kib accepted D44773: nullfs: Show correct exported flag..
Sat, Apr 13, 8:04 AM

Fri, Apr 12

kib accepted D44465: Reduce reliance on sys/sysproto.h pollution.
Fri, Apr 12, 11:36 PM
kib accepted D44766: sys/malloc.h: move sys/proc.h MALLOC_DECLAREs here.
Fri, Apr 12, 11:35 PM
kib accepted D44474: acl_to_text: include sys/param.h for MAXLOGNAME.
Fri, Apr 12, 11:35 PM
kib accepted D44752: defaults/rc.conf: Remove /usr/lib32 from ldconfig32_paths.
Fri, Apr 12, 1:24 AM
kib accepted D44751: rc.d/ldconfig: Compute ldconfig paths in a function.
Fri, Apr 12, 1:23 AM
kib committed rGfb8eb2042dbe: pciconf(8): dump AMD IOMMU Base Capability (authored by kib).
pciconf(8): dump AMD IOMMU Base Capability
Fri, Apr 12, 1:19 AM
kib committed rGfd02cfeb9ade: pcireg.h: Add AMD IOMMU Base Cap definitions (authored by kib).
pcireg.h: Add AMD IOMMU Base Cap definitions
Fri, Apr 12, 1:19 AM
kib committed rG83fe75d6c16e: pcireg.h: add include guard (authored by kib).
pcireg.h: add include guard
Fri, Apr 12, 1:19 AM
kib committed rG80bf64732ce5: acpidump: add Foundation copyright (authored by kib).
acpidump: add Foundation copyright
Fri, Apr 12, 1:19 AM
kib committed rGec9952bd472e: acpidump: dump AMD IVRS table describing IOMMU layout (authored by kib).
acpidump: dump AMD IVRS table describing IOMMU layout
Fri, Apr 12, 1:19 AM
kib committed rG9ab216e2646a: acpidump: add printfield() helper (authored by kib).
acpidump: add printfield() helper
Fri, Apr 12, 1:19 AM
kib committed rG0a7de174b779: acpidump: bump WARNS to 6 (authored by kib).
acpidump: bump WARNS to 6
Fri, Apr 12, 1:19 AM
kib committed rG0b5c1913b0fe: acpidump: the acpi_get_fadt_revision() argument is unused (authored by kib).
acpidump: the acpi_get_fadt_revision() argument is unused
Fri, Apr 12, 1:19 AM
kib committed rG2bb6e35f3b9f: acpidump: silent the warning about unaligned uuid (authored by kib).
acpidump: silent the warning about unaligned uuid
Fri, Apr 12, 1:19 AM
kib committed rGa648f68c9254: stop_all_proc(): skip traced or signal-stoped processes (authored by kib).
stop_all_proc(): skip traced or signal-stoped processes
Fri, Apr 12, 1:19 AM

Thu, Apr 11

kib accepted D44734: Create little-endian hints files on all architectures.
Thu, Apr 11, 11:43 AM
kib added inline comments to D44734: Create little-endian hints files on all architectures.
Thu, Apr 11, 9:42 AM

Wed, Apr 10

kib committed rG1e6db7be6921: pciconf(8): dump AMD IOMMU Base Capability (authored by kib).
pciconf(8): dump AMD IOMMU Base Capability
Wed, Apr 10, 5:29 AM
kib committed rG33adb388c78e: pcireg.h: Add AMD IOMMU Base Cap definitions (authored by kib).
pcireg.h: Add AMD IOMMU Base Cap definitions
Wed, Apr 10, 5:29 AM
kib committed rG1cd9868f04c3: pcireg.h: add include guard (authored by kib).
pcireg.h: add include guard
Wed, Apr 10, 5:29 AM
kib closed D44732: pciconf(8): dump AMD IOMMU Base Capability.
Wed, Apr 10, 5:29 AM
kib updated the diff for D44732: pciconf(8): dump AMD IOMMU Base Capability.

Reference the IOMMU spec for pcireg.h

Wed, Apr 10, 12:27 AM
kib added inline comments to D44732: pciconf(8): dump AMD IOMMU Base Capability.
Wed, Apr 10, 12:26 AM

Tue, Apr 9

kib requested review of D44732: pciconf(8): dump AMD IOMMU Base Capability.
Tue, Apr 9, 11:15 PM
kib added a comment to D44623: vm: add macro to mark arguments used when NUMA is defined.

Go ahead with the current patch.

Tue, Apr 9, 6:14 AM

Mon, Apr 8

kib accepted D38852: vm: improve kstack_object pindex calculation scheme to avoid pindex holes.
In D38852#1018757, @alc wrote:

Just so we're all on the same page, I want to point out the following: While this patch achieves contiguity, it doesn't guarantee 2 MB alignment. Let 'F' represent a fully populated 2 MB reservation, 'E', represent a partially populated reservation, where the population begins in the middle and goes to the end, and 'B' is the complement of 'E', where the population begins at the start and ends in the middle. Typically, the physical memory allocation for one chunk of stacks on amd64 looks like 'EFFFB'. While it would be nice to achieve 'FFFF', this patch is already a great improvement over the current state of affairs.

Mon, Apr 8, 9:58 PM
kib added a comment to D44581: Introduce a new kernel-only errno EUNSDEP for unsatisfied dependency.
In D44581#1017675, @kib wrote:

I do not like it. Errno has the defined scope of providing the communication between kernel and userspace. Some of the 'out of band' values like ERESTART and EJUSTRETURN and valid extensions because they modify the kernel->user edge behavior, all other in the list are hacks.

If "Errno" above means the standard macro errno, of course, but then ERESTART and EJUSTRETURN are also just hacks: They never reach userland (i.e., are never placed in errno).

They are hacks, but not in the same sense as the proposed value. ERESTART and EJUSTRETURN directly affect the content of the user context restored by the syscall return to userspace, same as any other userspace errno values, unlike the EUNSDEP (and unlike ENOIOCTL/EDIRIOCTL/ERELOOKUP).

Mon, Apr 8, 9:54 PM

Sun, Apr 7

kib added inline comments to D38852: vm: improve kstack_object pindex calculation scheme to avoid pindex holes.
Sun, Apr 7, 8:12 PM

Sat, Apr 6

kib added inline comments to D38852: vm: improve kstack_object pindex calculation scheme to avoid pindex holes.
Sat, Apr 6, 11:19 PM

Fri, Apr 5

kib accepted D44646: swap_pager: Unbusy readahead pages after an I/O error.
Fri, Apr 5, 9:27 PM
kib added a comment to D44581: Introduce a new kernel-only errno EUNSDEP for unsatisfied dependency.

Adding anything to errno is the wrong direction. If you want to get arbitrary error back from kldload(2), then perhaps kldload2(2) is due, with a signature allowing to pass back a text describing the error.
This would be similar to the interface of rtld itself, where dl*() functions do not try to extend errno but provide dlerror() with the description of the problem.

Fri, Apr 5, 9:27 PM
kib added inline comments to D44646: swap_pager: Unbusy readahead pages after an I/O error.
Fri, Apr 5, 8:56 PM
kib added inline comments to D44646: swap_pager: Unbusy readahead pages after an I/O error.
Fri, Apr 5, 8:00 PM
kib committed rGf1519a8229d7: acpidump: add Foundation copyright (authored by kib).
acpidump: add Foundation copyright
Fri, Apr 5, 3:19 PM
kib committed rG6d789b6126af: acpidump: dump AMD IVRS table describing IOMMU layout (authored by kib).
acpidump: dump AMD IVRS table describing IOMMU layout
Fri, Apr 5, 3:19 PM
kib committed rG969a4b8be5a7: acpidump: add printfield() helper (authored by kib).
acpidump: add printfield() helper
Fri, Apr 5, 3:18 PM
kib committed rG29a6b72775ee: acpidump: bump WARNS to 6 (authored by kib).
acpidump: bump WARNS to 6
Fri, Apr 5, 3:18 PM
kib committed rG4262ad56a074: acpidump: the acpi_get_fadt_revision() argument is unused (authored by kib).
acpidump: the acpi_get_fadt_revision() argument is unused
Fri, Apr 5, 3:18 PM
kib committed rGde937ecb75ed: acpidump: silent the warning about unaligned uuid (authored by kib).
acpidump: silent the warning about unaligned uuid
Fri, Apr 5, 3:18 PM
kib closed D44634: acpidump: dump AMD IVRS table describing IOMMU layout.
Fri, Apr 5, 3:18 PM
kib added a comment to D44623: vm: add macro to mark arguments used when NUMA is defined.
In D44623#1017481, @kib wrote:

I have some bad feeling about all that <feature>_unused macros. Would it better to change it to unconditionally apply unused globall?

I guess __unused has no effect if the compiler can observe that the annotated variable is in fact used? Maybe it would trigger warnings?

I don't really like all the __unused aliases either, but I also don't want to just use __unused everywhere, it would be confusing.

Fri, Apr 5, 3:06 PM
kib committed rG235436d6311e: stop_all_proc(): skip traced or signal-stoped processes (authored by kib).
stop_all_proc(): skip traced or signal-stoped processes
Fri, Apr 5, 2:53 PM
kib closed D44638: stop_all_proc(): skip traced or signal-stoped processes.
Fri, Apr 5, 2:53 PM

Thu, Apr 4

kib added a comment to D44581: Introduce a new kernel-only errno EUNSDEP for unsatisfied dependency.

There is a wart in ELF, called copy relocation. When a binary references data object defined in dso, static linker allocates the symbol in the binary and issues the copy relocation, to copy the content of the dso at this symbol to binary. The dso is resolved to reference the binary' object, not its own instance. This was done to emulate the semantic of the .a archives, as everything in ELF.

Thu, Apr 4, 8:13 PM
kib requested review of D44638: stop_all_proc(): skip traced or signal-stoped processes.
Thu, Apr 4, 7:39 PM
kib added a comment to D44523: thread_single(9): decline external requests for traced or debugger-stopped procs.

The fundamental problem with P_TOTAL_STOP is that it is external to the stopped process but uses the same mechanism (suspending threads) as any kind of internal stop (single threading, SIGSTOP, ptracestop()). So having a thread suspended by P_TOTAL_STOP would make internal stop logic to consider the thread handled, and in reverse.

Thu, Apr 4, 7:15 PM
kib added a comment to D44581: Introduce a new kernel-only errno EUNSDEP for unsatisfied dependency.
In D44581#1017675, @kib wrote:

I do not like it. Errno has the defined scope of providing the communication between kernel and userspace. Some of the 'out of band' values like ERESTART and EJUSTRETURN and valid extensions because they modify the kernel->user edge behavior, all other in the list are hacks.
More, the hack you propose is for very special and local situation, where error can be propagated by other means, if ever needed.

Would you like if it get a positive value and goes under #ifndef _POSIX_SOURCE like other our errnos that are extensions?

Thu, Apr 4, 6:47 PM
kib requested changes to D44581: Introduce a new kernel-only errno EUNSDEP for unsatisfied dependency.

I do not like it. Errno has the defined scope of providing the communication between kernel and userspace. Some of the 'out of band' values like ERESTART and EJUSTRETURN and valid extensions because they modify the kernel->user edge behavior, all other in the list are hacks.

Thu, Apr 4, 5:46 PM
kib updated the diff for D44634: acpidump: dump AMD IVRS table describing IOMMU layout.

Markj' changes

Thu, Apr 4, 3:48 PM
kib added a comment to D44634: acpidump: dump AMD IVRS table describing IOMMU layout.
/*
  IVRS: Length=382, Revision=2, Checksum=127,
	OEMID=SUPERM, OEM Table ID=AMD IVRS, OEM Revision=0x1,
	Creator ID=AMD, Creator Revision=0x0
	VAsize=64 PAsize=48 GVAsize=2
	ATS_resp_res=0 DMA_preboot_remap=0 EFRsup=1
Thu, Apr 4, 3:47 PM
kib added a comment to D44626: tarfs: Implement VOP_BMAP.

BTW we never call bmap with runp NULL and runb non-NULL? Perhaps a debugpre hook for VOP_BMAP stating this would be due.

Thu, Apr 4, 3:33 PM
kib accepted D44626: tarfs: Implement VOP_BMAP.
Thu, Apr 4, 3:31 PM
kib requested review of D44634: acpidump: dump AMD IVRS table describing IOMMU layout.
Thu, Apr 4, 4:55 AM
kib accepted D44628: tarfs.5: Document the importance of zstd frames.
Thu, Apr 4, 2:55 AM
kib added inline comments to D44626: tarfs: Implement VOP_BMAP.
Thu, Apr 4, 2:54 AM
kib accepted D44627: tarfs: Inherit mnt_iosize_max from the lower vnode.
Thu, Apr 4, 2:49 AM
kib accepted D44623: vm: add macro to mark arguments used when NUMA is defined.

I have some bad feeling about all that <feature>_unused macros. Would it better to change it to unconditionally apply unused globall?

Thu, Apr 4, 2:47 AM
kib committed rGbfafc6a86ddb: thread_single(9): decline external requests for traced or debugger-stopped procs (authored by kib).
thread_single(9): decline external requests for traced or debugger-stopped procs
Thu, Apr 4, 2:46 AM
kib committed rGdf9c60beea7c: thread_single(9): decline external requests for traced or debugger-stopped procs (authored by kib).
thread_single(9): decline external requests for traced or debugger-stopped procs
Thu, Apr 4, 2:45 AM

Wed, Apr 3

kib accepted D44601: unionfs_lookup(): fix wild accesses to vnode private data.
Wed, Apr 3, 1:11 AM

Tue, Apr 2

kib committed rGb79fb02b32d9: x86: handle MXCSR from XSAVEOPT when x87 state was optimized (authored by kib).
x86: handle MXCSR from XSAVEOPT when x87 state was optimized
Tue, Apr 2, 3:19 PM
kib committed rGc787fedf1262: amd64: initialize td_frame stack area for init(8) main thread (authored by kib).
amd64: initialize td_frame stack area for init(8) main thread
Tue, Apr 2, 3:18 PM
kib committed rGd4676f565d9a: x86: test the right CPUID bit when checking for XSAVEOPT support (authored by kib).
x86: test the right CPUID bit when checking for XSAVEOPT support
Tue, Apr 2, 3:18 PM
kib committed rG38fdb37047ea: x86: handle MXCSR from XSAVEOPT when x87 state was optimized (authored by kib).
x86: handle MXCSR from XSAVEOPT when x87 state was optimized
Tue, Apr 2, 9:00 AM
kib committed rG38c74c46b220: amd64: initialize td_frame stack area for init(8) main thread (authored by kib).
amd64: initialize td_frame stack area for init(8) main thread
Tue, Apr 2, 9:00 AM
kib committed rGb40ed1134d67: x86: test the right CPUID bit when checking for XSAVEOPT support (authored by kib).
x86: test the right CPUID bit when checking for XSAVEOPT support
Tue, Apr 2, 9:00 AM

Mon, Apr 1

kib added inline comments to D38852: vm: improve kstack_object pindex calculation scheme to avoid pindex holes.
Mon, Apr 1, 10:37 PM
kib added inline comments to D44222: ipsec_accel: handle TSO if supported.
Mon, Apr 1, 2:27 PM

Sun, Mar 31

kib added a comment to D44453: coredirector - Intel TD/HFI driver - Part1: Add ITD/HFI related MSR/CPUID defines to specialregs.h.

Please mail me (kib@f.o) the updated patch in the format of git-format-patch, with the correct metadata. Most important, put your author's email and name as you want them to appear in the git commit log.

Sun, Mar 31, 9:00 PM
kib added inline comments to D44222: ipsec_accel: handle TSO if supported.
Sun, Mar 31, 8:55 PM
kib accepted D44564: ptrace: Add a test using PROC_REAP_KILL to kill a stopped debuggee.
Sun, Mar 31, 6:03 PM

Sat, Mar 30

kib added inline comments to D44523: thread_single(9): decline external requests for traced or debugger-stopped procs.
Sat, Mar 30, 7:43 PM
kib added inline comments to D44523: thread_single(9): decline external requests for traced or debugger-stopped procs.
Sat, Mar 30, 7:20 PM
kib committed rG9241ebc796c1: thread_single(9): decline external requests for traced or debugger-stopped procs (authored by kib).
thread_single(9): decline external requests for traced or debugger-stopped procs
Sat, Mar 30, 2:51 PM
kib closed D44523: thread_single(9): decline external requests for traced or debugger-stopped procs.
Sat, Mar 30, 2:51 PM
kib accepted D44564: ptrace: Add a test using PROC_REAP_KILL to kill a stopped debuggee.

Why do you need the killer child to kill the debuggee?

Sat, Mar 30, 2:43 PM
kib accepted D44559: linker: Don't invoke dtors without having invoked ctors.
Sat, Mar 30, 1:24 PM
kib committed rGf13afa7f5478: cdevpriv(9): document devfs_foreach_cdevpriv() (authored by kib).
cdevpriv(9): document devfs_foreach_cdevpriv()
Sat, Mar 30, 8:33 AM
kib committed rG8d9ed1dd5088: cdevpriv(9): add iterator (authored by kib).
cdevpriv(9): add iterator
Sat, Mar 30, 8:33 AM
kib committed rG36d16f5c5b50: cdevpriv(9): document devfs_foreach_cdevpriv() (authored by kib).
cdevpriv(9): document devfs_foreach_cdevpriv()
Sat, Mar 30, 8:33 AM
kib committed rGcde0347ca70b: cdevpriv(9): add iterator (authored by kib).
cdevpriv(9): add iterator
Sat, Mar 30, 8:32 AM

Fri, Mar 29

kib added inline comments to D44559: linker: Don't invoke dtors without having invoked ctors.
Fri, Mar 29, 8:52 PM