Page MenuHomeFreeBSD

kib (Konstantin Belousov)
User

Projects

User Details

User Since
May 16 2014, 7:35 PM (337 w, 16 h)

Recent Activity

Yesterday

kib added a comment to D27025: [RFC] Fix abort in malloc extent coalescing..

I think this change is fine as is.

Fri, Oct 30, 7:28 PM

Thu, Oct 29

kib added a comment to D26988: Move syscall_thread_{enter,exit}() to the slow path.

If doing such reorg, it might make sense to de-commission syscall_thread_enter/exit, renaming underscored functions back to syscall_thread_.... Then you would fetch SY_THR_STATIC once and cache it in local var.

Thu, Oct 29, 3:59 PM
kib accepted D26423: Integrate 4.4BSD-Lite2 changes to IOC_* definitions.
Thu, Oct 29, 11:48 AM
kib committed rS367129: MFC r366932:.
MFC r366932:
Thu, Oct 29, 11:20 AM
kib committed rS367128: MFC r366932:.
MFC r366932:
Thu, Oct 29, 11:10 AM

Wed, Oct 28

kib added a comment to D26988: Move syscall_thread_{enter,exit}() to the slow path.

enter is needed for syscalls from _loadable_ modules, not unloadable.

Wed, Oct 28, 10:26 PM
kib committed rS367120: Check for process group change in tty_wait_background()..
Check for process group change in tty_wait_background().
Wed, Oct 28, 10:13 PM
kib committed rS367116: MFC r366918:.
MFC r366918:
Wed, Oct 28, 9:02 PM
kib committed rS367115: MFC r366904:.
MFC r366904:
Wed, Oct 28, 9:01 PM
kib updated the summary of D26136: Handle LoR in flush_pagedep_deps()..
Wed, Oct 28, 7:42 PM
kib updated the summary of D26136: Handle LoR in flush_pagedep_deps()..
Wed, Oct 28, 7:31 PM
kib updated the diff for D26136: Handle LoR in flush_pagedep_deps()..

Provide extensive comment about get_parent_vp(), which in essence overviews the whole patch with its ERELOOKUP mechanism.

Wed, Oct 28, 7:09 PM
kib updated the diff for D26136: Handle LoR in flush_pagedep_deps()..

Last batch of fixes.
Passed full stress2 run.

Wed, Oct 28, 6:57 PM

Tue, Oct 27

kib added inline comments to D26981: Note that sys/systm.h is special too.
Tue, Oct 27, 10:12 PM
kib accepted D26981: Note that sys/systm.h is special too.
Tue, Oct 27, 9:45 PM
kib added inline comments to D26981: Note that sys/systm.h is special too.
Tue, Oct 27, 9:44 PM
kib accepted D26977: Add local variable to store sysent pointer.
Tue, Oct 27, 4:45 PM
kib added a comment to D26923: vm_map: Add fences around pmap rundown.

refcount_release_if_last sounds almost like refcount_release, and I suspect that refcount_release() might be enough. As I understand, the problem is that pmap_remove_pages() has to operate on current pmap, which causes all that troubles. What if we just use refcount_release() and keep p_vmspace with refcount zero until pmap_remove_pages() finish ?

Tue, Oct 27, 11:11 AM
kib added a comment to D26964: ufs softdep: end-of-life truncate should depend on dirent write.

I am not sure about this approach. Note that vref()/usecount reference does not prevent the vnode reclaim. So for instance force umount results in vgone() which does inactivation and reclaim regardless of the active state (or rather does inactivation if the vnode is active). In this case, it seems to not fix the issue.

Tue, Oct 27, 8:24 AM

Mon, Oct 26

kib added a comment to D26955: Try to canonicalize the first argument to mount(8).

All pseudo fs do not do anything with from/source.
For real filesystems, resolving source is actually useful for users, I believe, but it would be a user-visible change. In fact, I sometimes want the behavior that your patch introduces.

Mon, Oct 26, 12:41 PM
kib accepted D26908: mmap(2): Clarify that guard is taken from the stack region..
Mon, Oct 26, 8:27 AM

Sun, Oct 25

kib added a comment to D26908: mmap(2): Clarify that guard is taken from the stack region..
In D26908#600942, @alc wrote:

After mulling this over for the past day or so, I'd like to propose the following alternative description:

"This option creates a memory region that grows downward on demand and an adjoining guard that both reserves address space for the memory region to grow into and limits the memory region's growth. Together, the memory region and the guard occupy len bytes of the address space. The guard starts at the returned address, and the memory region ends at the returned address plus len bytes. Upon access to the guard, the memory region automatically grows in size, and the guard shrinks by an equal amount. Essentially, the boundary between the guard and the memory region moves downward so that the access falls within the enlarged memory region. However, the guard will never shrink to less than the number of pages specified by the sysctl security.bsd.stack_guard_page, thereby ensuring that a gap for detecting stack overflow always exists between the downward growing memory region and the closest memory region beneath it."

Sun, Oct 25, 11:20 AM

Sat, Oct 24

kib updated the diff for D26136: Handle LoR in flush_pagedep_deps()..

Rebase.
Fix build for !DIAGNOSTICS.
Latest batch of the fixes for bugs reported by Peter.

Sat, Oct 24, 7:55 PM
kib added a comment to D26923: vm_map: Add fences around pmap rundown.

vmspace != vm_map, and vm_map lock is only supposed to protect vm_map.

Sat, Oct 24, 7:52 PM
kib accepted D26934: Make it possible to mount nullfs(5) using plain mount(8).
Sat, Oct 24, 6:21 PM
kib accepted D26906: iommu_unmap_msi() added.
Sat, Oct 24, 4:59 PM
kib accepted D26910: Make pmap_invalidate_ept() wait synchronously for guest exits.
Sat, Oct 24, 9:32 AM
kib accepted D26922: audit: correct reporting of *execve(2) errors.
Sat, Oct 24, 12:41 AM
kib added a comment to D26922: audit: correct reporting of *execve(2) errors.

What about compat32 ?

Sat, Oct 24, 12:28 AM

Fri, Oct 23

kib accepted D26923: vm_map: Add fences around pmap rundown.
Fri, Oct 23, 10:48 PM
kib added inline comments to D26906: iommu_unmap_msi() added.
Fri, Oct 23, 10:43 PM
kib accepted D26906: iommu_unmap_msi() added.
Fri, Oct 23, 10:43 PM
kib added inline comments to D26923: vm_map: Add fences around pmap rundown.
Fri, Oct 23, 10:12 PM
kib accepted D26921: Remove the check against "loops" from mount_nullfs.
Fri, Oct 23, 10:06 PM
kib added a comment to D26922: audit: correct reporting of *execve(2) errors.

Instead of trying to outguess all syscalls, might be add a TDP flag to indicate that AUDIT on exit does not need to be done. Then, in the special-case syscalls like *execve() and perhaps all users of kern_posix_exit(), call AUDIT_SYSCALL_EXIT() manually when you know the correct errno.

Fri, Oct 23, 10:05 PM
kib added inline comments to D26922: audit: correct reporting of *execve(2) errors.
Fri, Oct 23, 10:03 PM
kib closed D26913: xhci: Handle the case when MSI-X BAR is same as IO BAR..
Fri, Oct 23, 6:18 PM
kib committed rS366978: xhci: Handle the case when MSI-X BAR is the same as IO BAR..
xhci: Handle the case when MSI-X BAR is the same as IO BAR.
Fri, Oct 23, 6:18 PM
kib updated the diff for D26908: mmap(2): Clarify that guard is taken from the stack region..

Reword text to distinguish between guard as is, and minimal guard size.

Fri, Oct 23, 6:11 PM
kib added inline comments to D26906: iommu_unmap_msi() added.
Fri, Oct 23, 5:06 PM
kib accepted D26887: Add bus_dma_iommu_set_buswide() stubs.

As far as this pass tinderbox ...

Fri, Oct 23, 4:58 PM
kib committed rS366964: MFC r366768:.
MFC r366768:
Fri, Oct 23, 10:25 AM
kib added a comment to D26887: Add bus_dma_iommu_set_buswide() stubs.
In D26887#600298, @br wrote:

Would it make sense to move these to subr_bus_dma.c in the non-IOMMU case?

I'm not sure. I guess we can't customize per-arch implementation of these functions in non-IOMMU case

Fri, Oct 23, 9:55 AM

Thu, Oct 22

kib requested review of D26913: xhci: Handle the case when MSI-X BAR is same as IO BAR..
Thu, Oct 22, 11:18 PM
kib updated the diff for D26908: mmap(2): Clarify that guard is taken from the stack region..

Ed' changes.

Thu, Oct 22, 7:01 PM
kib requested review of D26908: mmap(2): Clarify that guard is taken from the stack region..
Thu, Oct 22, 6:43 PM
kib accepted D26907: Conditionally compile struct vm_phys_seg's md_first field.
Thu, Oct 22, 6:37 PM
kib committed rS366937: MFC r339748, r340343, r340347, r341256, r366015-r366023, r366549:.
MFC r339748, r340343, r340347, r341256, r366015-r366023, r366549:
Thu, Oct 22, 3:24 PM
kib added inline comments to D26851: vm_map: Handle kernel map entry allocator recursion.
Thu, Oct 22, 12:41 PM
kib closed D26826: sysv_sem: semusz depends on semume..
Thu, Oct 22, 9:28 AM
kib committed rS366932: sysv_sem: semusz depends on semume..
sysv_sem: semusz depends on semume.
Thu, Oct 22, 9:28 AM

Wed, Oct 21

kib closed D26894: mmap(2): Document guard size and related EINVAL..
Wed, Oct 21, 9:40 PM
kib committed rS366918: mmap(2): Document guard size for MAP_STACK and related EINVAL..
mmap(2): Document guard size for MAP_STACK and related EINVAL.
Wed, Oct 21, 9:40 PM
kib updated the diff for D26894: mmap(2): Document guard size and related EINVAL..

Handle Ed' notes.
Use .Dv for sysctl mib name.
Bump date.

Wed, Oct 21, 6:37 PM
kib requested review of D26894: mmap(2): Document guard size and related EINVAL..
Wed, Oct 21, 5:48 PM
kib committed rS366910: MFC r366713, r366843 (by jhb):.
MFC r366713, r366843 (by jhb):
Wed, Oct 21, 3:07 PM
kib committed rS366909: MFC r366712:.
MFC r366712:
Wed, Oct 21, 3:04 PM
kib accepted D26887: Add bus_dma_iommu_set_buswide() stubs.
Wed, Oct 21, 2:38 PM
kib updated the diff for D26826: sysv_sem: semusz depends on semume..

Add parenthesis.

Wed, Oct 21, 1:35 PM
kib closed D26856: Improve FPU Tag Word reconstruction on i386 to indicate register states.
Wed, Oct 21, 12:15 AM
kib committed rS366904: Improve FPU Tag Word reconstruction on i386 to indicate register states..
Improve FPU Tag Word reconstruction on i386 to indicate register states.
Wed, Oct 21, 12:15 AM

Tue, Oct 20

kib added reviewers for D26826: sysv_sem: semusz depends on semume.: markj, jhb.
Tue, Oct 20, 11:32 PM
kib added a comment to D26856: Improve FPU Tag Word reconstruction on i386 to indicate register states.

I will handle the patch.

Tue, Oct 20, 11:30 PM
kib accepted D26856: Improve FPU Tag Word reconstruction on i386 to indicate register states.
Tue, Oct 20, 11:28 PM
kib added inline comments to D26816: Fix linprocfs(4) /proc/self/mem semantics.
Tue, Oct 20, 11:27 PM
kib accepted D26816: Fix linprocfs(4) /proc/self/mem semantics.
Tue, Oct 20, 2:20 PM
kib accepted D25618: Fix potential race condition in linux stat(2).
Tue, Oct 20, 1:45 PM
kib added a comment to D26856: Improve FPU Tag Word reconstruction on i386 to indicate register states.

Merging i386/npx.c and amd64/fpu.c is a lot of work for many reasons, I evaluated that several years ago. For now keeping them separate is better IMO.

Tue, Oct 20, 1:39 PM
kib added inline comments to D26816: Fix linprocfs(4) /proc/self/mem semantics.
Tue, Oct 20, 1:25 PM

Mon, Oct 19

kib added inline comments to D26816: Fix linprocfs(4) /proc/self/mem semantics.
Mon, Oct 19, 10:17 PM
kib committed rS366849: nullfs: ensure correct lock is taken after bypass..
nullfs: ensure correct lock is taken after bypass.
Mon, Oct 19, 7:23 PM
kib committed rS366848: vgonel(): avoid recursing into VOP_INACTIVE()..
vgonel(): avoid recursing into VOP_INACTIVE().
Mon, Oct 19, 7:20 PM
kib accepted D26857: IOMMU quirks added.
Mon, Oct 19, 4:54 PM
kib added inline comments to D26857: IOMMU quirks added.
Mon, Oct 19, 3:44 PM
kib added a reviewer for D26857: IOMMU quirks added: mav.
Mon, Oct 19, 3:44 PM
kib accepted D26859: Assign MSI entry to x86.
Mon, Oct 19, 3:29 PM
kib added a comment to D26857: IOMMU quirks added.

No, this should be in SATA driver.

Mon, Oct 19, 1:57 PM
kib updated the diff for D26136: Handle LoR in flush_pagedep_deps()..

Document softdep_pre*().

Mon, Oct 19, 11:31 AM
kib added inline comments to D26136: Handle LoR in flush_pagedep_deps()..
Mon, Oct 19, 11:30 AM
kib accepted D26705: Manage MSI iommu pages.
Mon, Oct 19, 11:06 AM
kib accepted D26845: Further refinements of ptsname_r(3) interface:.
Mon, Oct 19, 7:09 AM

Sun, Oct 18

kib added inline comments to D26705: Manage MSI iommu pages.
Sun, Oct 18, 9:12 PM
kib added inline comments to D26845: Further refinements of ptsname_r(3) interface:.
Sun, Oct 18, 9:08 PM
kib updated the diff for D26136: Handle LoR in flush_pagedep_deps()..

Newer snap.

Sun, Oct 18, 8:18 AM
kib added a comment to D26705: Manage MSI iommu pages.

I am curious. There is no arch restriction on the value of MSI base address ? System can map MSI page anywhere, and just programming the correct address into MSI address register would make it work ?

Sun, Oct 18, 6:02 AM
kib added a comment to D26818: cdefs: Assume __GNUC__ defined.
In D26818#598251, @imp wrote:
In D26818#598049, @kib wrote:
In D26818#598018, @imp wrote:
In D26818#597997, @kib wrote:

In fact having gnuc-isms clearly marked is a useful feature.

Everything in this file is a gnu ism... no?

No, I do not view it this way. First, there is enough stuff not related to GNU C extensions. Starting from the compat shims for c89 or even K&R (but some stuff like static assert or array sizing in args, is really for c99), POSIX/XSI/BSD visibility macros, ending in thread-local shims etc that do not require gnu-isms.

Do you have a compiler I can test whatever I come up with then? I don't mean to be difficult, but if there's no way to test it, there's no way to know if the ancient stuff even works.

What do you refer above as an ancient stuff ? Except for __P() and perhaps __inline/__const/__volatile everything else is not about ancient. [We should get rid of __inline/__const etc IMO].

Sun, Oct 18, 2:12 AM
kib added inline comments to D26827: Fix sleepq_add panic happening with too wide net epoch..
Sun, Oct 18, 1:23 AM
kib accepted D26830: vmx: Implement pmap (de)activation in C.
Sun, Oct 18, 1:19 AM

Sat, Oct 17

kib requested review of D26826: sysv_sem: semusz depends on semume..
Sat, Oct 17, 9:07 AM
kib added inline comments to D25445: Complete the fib<4|6>_lookup_nh_<basic|ext> -> fib<4|6>_lookup() transition.
Sat, Oct 17, 8:18 AM · network
kib added a comment to D26818: cdefs: Assume __GNUC__ defined.
In D26818#598018, @imp wrote:
In D26818#597997, @kib wrote:

In fact having gnuc-isms clearly marked is a useful feature.

Everything in this file is a gnu ism... no?

Sat, Oct 17, 8:08 AM
kib added a comment to D26818: cdefs: Assume __GNUC__ defined.

The bigger part of !GNUC support is in machine/atomic.h.

Sat, Oct 17, 12:06 AM

Fri, Oct 16

kib committed rS366768: amd64 pmap.h: explicitly provide constants values instead of relying.
amd64 pmap.h: explicitly provide constants values instead of relying
Fri, Oct 16, 4:22 PM
kib accepted D26406: Stop calling set_syscall_retval() from linux_set_syscall_retval().
Fri, Oct 16, 2:24 PM

Thu, Oct 15

kib accepted D26802: link_elf_obj: Colour VM objects.

I think that memory saving is still more important than really minor TLB usage optimization.

Thu, Oct 15, 11:53 PM
kib accepted D26781: Use a template assembly file to generate the embedded MFS..
Thu, Oct 15, 11:37 PM
kib added inline comments to D26781: Use a template assembly file to generate the embedded MFS..
Thu, Oct 15, 4:19 AM
kib added a comment to D26668: Add native system call for eventfd.

Ok, the capsicum argument is convincing enough to not go with /dev/eventfd route. But on the other hand having a single syscall that can also serve timerfd in future would be nice indeed. I do not think that shmfd should be plugged into this case.

Thu, Oct 15, 1:03 AM · Linux Emulation
kib accepted D26770: vmem: Allocate btags before looping in vmem_xalloc().
Thu, Oct 15, 1:00 AM
kib committed rS366714: MFC r366551:.
MFC r366551:
Thu, Oct 15, 12:50 AM