Page MenuHomeFreeBSD

markj (Mark Johnston)
User

Projects (9)

User Details

User Since
Mar 12 2014, 1:00 AM (591 w, 4 d)

Recent Activity

Fri, Jul 11

markj abandoned D16659: Bump pageout_oom_seq..
Fri, Jul 11, 3:18 PM
markj added inline comments to D51223: libusb: implement libusb_hotplug_get_user_data.
Fri, Jul 11, 2:57 PM
markj added inline comments to D51223: libusb: implement libusb_hotplug_get_user_data.
Fri, Jul 11, 1:39 PM
markj added a comment to D51226: libusb: implement libusb_fill_bulk_stream_transfer.

Approved.

Fri, Jul 11, 1:37 PM
markj added inline comments to D51222: libusb: implement libusb_dev_mem_{alloc,free}.
Fri, Jul 11, 1:36 PM
markj added inline comments to D51256: vmm: support INOUT manual decode..
Fri, Jul 11, 1:31 PM
markj accepted D51235: amd64 pmap: avoid an unnecessary demotion.
Fri, Jul 11, 1:23 PM
markj added a comment to D51221: libusb: implement libusb_free_pollfds.

Looks ok to me with the man page nits fixed.

Fri, Jul 11, 1:14 PM
markj accepted D51220: arm64 pmap: do not panic on inability to insert ptp into trie.

Do you have some local modifications to test the ADDR_IS_KERNEL(va) code path in pmap_enter_l2()?

Fri, Jul 11, 1:06 PM
markj accepted D51248: vm_domainset: Print correct function in KASSERT()/panic().
Fri, Jul 11, 12:48 PM

Thu, Jul 10

markj committed rG103763fab4f7: vm_pageout_scan_inactive: take a lock break (authored by rlibby).
vm_pageout_scan_inactive: take a lock break
Thu, Jul 10, 9:08 PM
markj updated the summary of D48337: vm_pageout: Scan inactive dirty pages less aggressively.
Thu, Jul 10, 9:06 PM
markj updated the diff for D48337: vm_pageout: Scan inactive dirty pages less aggressively.

Don't count unswappable pages as dirty. I believe this will address kib's concern
by side-stepping it: if a dirty page can't be swapped out, we simply don't count it
against the scan limit, so we will process as many pages as necessary.

Thu, Jul 10, 8:58 PM
markj updated the diff for D48337: vm_pageout: Scan inactive dirty pages less aggressively.
  • Rebase.
  • Drop the change to permit clustering with dirty pages in the inactive queue. It's logically a separate change, and it was incomplete: on its own it may cause the swap pager to call vm_page_deactivate_noreuse() on pages that hadn't yet made a full trip through the inactive queue. Compare with vm_pageout_flush(), which only moves pages to the head of the inactive queue if they are in the laundry.
Thu, Jul 10, 8:33 PM
markj committed rG791ac36668d5: random: Remove ARGSUSED annotations from random_harvestq.c (authored by markj).
random: Remove ARGSUSED annotations from random_harvestq.c
Thu, Jul 10, 8:08 PM
markj committed rG475a36762f9d: random: Define a macro for getting the CPU cycle count (authored by markj).
random: Define a macro for getting the CPU cycle count
Thu, Jul 10, 8:08 PM
markj committed rGcf1b83414f1e: random: Replace a comment with a static assertion (authored by markj).
random: Replace a comment with a static assertion
Thu, Jul 10, 8:08 PM
markj committed rG11d1a5c57900: random: Move entropy harvest queue lock macros to random_harvestq.c (authored by markj).
random: Move entropy harvest queue lock macros to random_harvestq.c
Thu, Jul 10, 8:08 PM
markj added inline comments to D51207: truncate: fix a minor nit + add a hole-punching test.
Thu, Jul 10, 4:42 PM
markj accepted D51233: inotify: do not call into namei() with a locked vnode.

Thank you. I guess the problem arises when the directory is a mount point, and we need to upgrade the lock?

Thu, Jul 10, 12:48 PM

Wed, Jul 9

markj updated the diff for D49442: pmap: Avoid clearing the accessed bit for wired mappings.

Modify the arm64 pmap to avoid clearing ATTR_AF on wired PTEs as well.

Wed, Jul 9, 8:49 PM
markj accepted D51039: release: Consistently use variables for invoking certain tools.
Wed, Jul 9, 4:22 PM
markj accepted D51211: Make NFS server aware of mnt_renamelock.
Wed, Jul 9, 3:53 PM
markj added a comment to D50820: libusb: implement libusb_setlocale to support different locale of error message as upstream does.
Wed, Jul 9, 3:50 PM
markj accepted D51212: fcntl(FD_SETFD): make FD_CLOFORK non-sticky.
Wed, Jul 9, 2:54 AM
markj added inline comments to D51211: Make NFS server aware of mnt_renamelock.
Wed, Jul 9, 12:46 AM
markj committed rG5098a5b48357: md: Use a larger buffer for the ident string (authored by markj).
md: Use a larger buffer for the ident string
Wed, Jul 9, 12:45 AM
markj committed rG10d14a37206e: localedef: Fix a wchar_t comparison (authored by markj).
localedef: Fix a wchar_t comparison
Wed, Jul 9, 12:45 AM
markj committed rGd162d59c393d: localedef: Use consistent sorting order (authored by markj).
localedef: Use consistent sorting order
Wed, Jul 9, 12:45 AM

Tue, Jul 8

markj added inline comments to D51206: kern_descrip.c: provide helpers to translate between fd flags namespace.
Tue, Jul 8, 9:55 PM
markj added a comment to D51206: kern_descrip.c: provide helpers to translate between fd flags namespace.

BTW there are some test failures when this change is applied. e.g.,

Tue, Jul 8, 8:06 PM
markj added inline comments to D51206: kern_descrip.c: provide helpers to translate between fd flags namespace.
Tue, Jul 8, 7:51 PM
markj added inline comments to D51206: kern_descrip.c: provide helpers to translate between fd flags namespace.
Tue, Jul 8, 7:22 PM
markj accepted D51206: kern_descrip.c: provide helpers to translate between fd flags namespace.
Tue, Jul 8, 5:49 PM
markj added a comment to D51204: libc: only append exterr herald and text if exterr was recorded.

I verified that the test failures are gone now, thank you.

Tue, Jul 8, 3:46 PM
markj accepted D51204: libc: only append exterr herald and text if exterr was recorded.
Tue, Jul 8, 3:45 PM
markj committed rG9911028f53d5: libc: mkostemp(3) should support O_CLOFORK (authored by rbranco_suse.com).
libc: mkostemp(3) should support O_CLOFORK
Tue, Jul 8, 3:45 PM
markj committed rG1e570722dced: cap_dns/tests/dns_test: mark tests as needing network access (authored by guest-svmhdvn).
cap_dns/tests/dns_test: mark tests as needing network access
Tue, Jul 8, 12:57 PM
markj committed rG26e31700fc7d: sysctl_test: Print standard error if sysctl -ao fails (authored by markj).
sysctl_test: Print standard error if sysctl -ao fails
Tue, Jul 8, 12:57 PM
markj added a comment to D50680: libusb: replace DEBUG_LEVEL with LOG_LEVEL.

Approved, but see the inline comment.

Tue, Jul 8, 12:15 PM
markj added inline comments to D50820: libusb: implement libusb_setlocale to support different locale of error message as upstream does.
Tue, Jul 8, 12:11 PM
markj added a comment to D49851: Document the named attribute interface.

Perhaps modify extattr.2 and extattr.3 to reference this page?

Tue, Jul 8, 11:59 AM

Mon, Jul 7

markj committed rG393fefa3d9c7: mkimg: Make output reproducible using a specific timestamp (authored by markj).
mkimg: Make output reproducible using a specific timestamp
Mon, Jul 7, 8:29 PM
markj closed D51077: mkimg: Make output reproducible using a specific timestamp.
Mon, Jul 7, 8:29 PM
markj committed rG12bef37a824c: dtrace: fix symbol address resolution (authored by zldrobit_gmail.com).
dtrace: fix symbol address resolution
Mon, Jul 7, 8:17 PM
markj committed rGbc6b526e2b71: inotify: Use memcmp() to compare embedded file names (authored by markj).
inotify: Use memcmp() to compare embedded file names
Mon, Jul 7, 8:17 PM
markj committed rG4ea7a920de22: inotify tests: Use unsigned integers for event masks (authored by markj).
inotify tests: Use unsigned integers for event masks
Mon, Jul 7, 8:17 PM
markj committed rG3a20f630a9fc: git-arc: Fix handling of author names when we have to make a guess (authored by markj).
git-arc: Fix handling of author names when we have to make a guess
Mon, Jul 7, 8:17 PM
markj closed D51188: dtrace: fix symbol address resolving.
Mon, Jul 7, 8:17 PM · DTrace
markj added a comment to D51188: dtrace: fix symbol address resolving.

@markj I found that we should also delete st_value = 0 in dt_module_symsort32/64(), otherwise dmp->dm_aslen would always be lesser than dmp->dm_asrsv by at least one. As a result, dt_module_symsort32/64() ignore the last element in dmp->dm_asmap and dt_module_symaddr32/64() wouldn't find it. I had to update the patch accordingly. Sorry for interrupting the test suite procedure.

Mon, Jul 7, 6:52 PM · DTrace
markj added a comment to D51188: dtrace: fix symbol address resolving.

@markj I found that we should also delete st_value = 0 in dt_module_symsort32/64(), otherwise dmp->dm_aslen would always be lesser than dmp->dm_asrsv by at least one. As a result, dt_module_symsort32/64() ignore the last element in dmp->dm_asmap and dt_module_symaddr32/64() wouldn't find it. I had to update the patch accordingly. Sorry for interrupting the test suite procedure.

Mon, Jul 7, 6:41 PM · DTrace
markj accepted D51188: dtrace: fix symbol address resolving.

I'll commit this once I've gone through the test suite.

Mon, Jul 7, 4:56 PM · DTrace
markj updated the diff for D51154: random: Add NIST SP 800-90B entropy source health test implementations.

Apply feedback from Conrad.

Mon, Jul 7, 4:41 PM
markj updated the diff for D51155: random: Treat writes to /dev/random as separate from /entropy.

Bump __FreeBSD_version.

Mon, Jul 7, 4:34 PM
markj committed rGeda96744b434: lposix: Clean up the posix namespace definitions (authored by markj).
lposix: Clean up the posix namespace definitions
Mon, Jul 7, 3:44 PM
markj closed D51158: lposix: Clean up the posix namespace definitions.
Mon, Jul 7, 3:44 PM
markj committed rG9940c974029b: random: Change the entropy harvest event queuing scheme (authored by markj).
random: Change the entropy harvest event queuing scheme
Mon, Jul 7, 3:43 PM
markj closed D51112: random: Change the entropy harvest event queuing scheme.
Mon, Jul 7, 3:43 PM
markj added inline comments to D51112: random: Change the entropy harvest event queuing scheme.
Mon, Jul 7, 2:57 PM
markj added a comment to D51188: dtrace: fix symbol address resolving.

This change looks reasonable to me. I think the problem with st_value = 0 only applies to ET_REL KLDs.

Mon, Jul 7, 2:45 PM · DTrace
markj accepted D51126: opendir, fdopendir: Add tests, improve robustness..
Mon, Jul 7, 2:10 PM
markj committed rG97452b097a4a: examples/inotify: Handle a zero-length name (authored by markj).
examples/inotify: Handle a zero-length name
Mon, Jul 7, 2:01 PM

Sun, Jul 6

markj committed rGea4042246d61: libsysdecode: Add O_CLOFORK flags to sysdecode (authored by rbranco_suse.com).
libsysdecode: Add O_CLOFORK flags to sysdecode
Sun, Jul 6, 11:20 PM
markj committed rG20ee243707ff: tests: Add Illumos oclo tests for O_CLOEXEC & O_CLOFORK (authored by rbranco_suse.com).
tests: Add Illumos oclo tests for O_CLOEXEC & O_CLOFORK
Sun, Jul 6, 11:20 PM
markj committed rG4140012f8369: tests: Adapt oclo tests to FreeBSD (authored by rbranco_suse.com).
tests: Adapt oclo tests to FreeBSD
Sun, Jul 6, 11:20 PM
markj committed rG71a0af25a535: libopenbsd: Add recallocarray() (authored by rbranco_suse.com).
libopenbsd: Add recallocarray()
Sun, Jul 6, 11:20 PM
markj committed rG63503cbe5a48: tests: Adapt existing tests for O_CLOEXEC to O_CLOFORK (authored by rbranco_suse.com).
tests: Adapt existing tests for O_CLOEXEC to O_CLOFORK
Sun, Jul 6, 11:20 PM
markj committed rGefadb5c2a908: Add manpages for O_CLOFORK flag and others (authored by rbranco_suse.com).
Add manpages for O_CLOFORK flag and others
Sun, Jul 6, 11:19 PM
markj committed rG971f738679e2: kern: Make dup3() support O_CLOFORK (authored by rbranco_suse.com).
kern: Make dup3() support O_CLOFORK
Sun, Jul 6, 11:19 PM
markj committed rGe6b64eea5e77: kern: Make close_range() support CLOSE_RANGE_CLOEXEC (authored by rbranco_suse.com).
kern: Make close_range() support CLOSE_RANGE_CLOEXEC
Sun, Jul 6, 11:19 PM
markj committed rGbe6c405c4d34: socket: Add MSG_CMSG_CLOFORK flag (authored by rbranco_suse.com).
socket: Add MSG_CMSG_CLOFORK flag
Sun, Jul 6, 11:19 PM
markj committed rG04a481b743c5: socket: Add SOCK_CLOFORK flag (authored by rbranco_suse.com).
socket: Add SOCK_CLOFORK flag
Sun, Jul 6, 11:19 PM
markj committed rG02ffca404e65: kern: Add support for POSIX O_CLOFORK flag (authored by rbranco_suse.com).
kern: Add support for POSIX O_CLOFORK flag
Sun, Jul 6, 11:19 PM
markj accepted D51183: amd64: move efi_map_regs definition.

Please be sure to include a Fixes: tag in the commit message.

Sun, Jul 6, 9:21 PM
markj accepted D51180: amd64 pmap: convert some panics to KASSERTs.
Sun, Jul 6, 6:40 PM
markj accepted D51179: vm_fault: drop never-true busy_sleep test.
Sun, Jul 6, 5:51 PM
markj accepted D51173: amd64: fix physmap entry addition logic.
Sun, Jul 6, 5:49 PM

Sat, Jul 5

markj accepted D51124: cp: Add descriptions to all test cases..
Sat, Jul 5, 4:38 PM
markj accepted D46608: kern_domainset: Split domainset validation logic into a separate function.
Sat, Jul 5, 3:56 PM
markj added a comment to D44567: bhyve: Add support for specifying VM NUMA configuration.

Looks pretty good to me.

Sat, Jul 5, 3:52 PM
markj added inline comments to D44566: libvmmapi: Add support for setting up and configuring guest NUMA domains.
Sat, Jul 5, 3:49 PM
markj added a comment to D44565: vmm: Add support for specifying NUMA configuration.

This is looking pretty good, I just had one question about compatibility, the other comments are just nits.

Sat, Jul 5, 3:39 PM
markj added a comment to D50740: libusb: fix incorrect member name in bos_descriptor.

In libusb20_desc.h I see a macro LIBUSB20_BOS_DESCRIPTOR which references the old field name. I'm not really sure what that macro is doing--should that be updated as well?

The BOS_DESCRIPTOR in libusb20 is a dedicated structure for decode in libusb20 and the definition resides inside the libusb20_desc.c. Thus there is no reference from libusb10 structure.

For libusb10, we require a xfer and copy it into buffer directly. It will not be parsed by libusb20.

Sat, Jul 5, 3:15 PM
markj committed rG667ef8875bad: nuageinit: Add wrappers for chmod and chown (authored by markj).
nuageinit: Add wrappers for chmod and chown
Sat, Jul 5, 2:56 PM
markj closed D51159: nuageinit: Add wrappers for chmod and chown.
Sat, Jul 5, 2:55 PM
markj added a comment to D50740: libusb: fix incorrect member name in bos_descriptor.

In libusb20_desc.h I see a macro LIBUSB20_BOS_DESCRIPTOR which references the old field name. I'm not really sure what that macro is doing--should that be updated as well?

Sat, Jul 5, 2:52 PM

Fri, Jul 4

markj committed rG88d94ead7f7c: lposix: Use reentrant passwd and group lookup functions (authored by markj).
lposix: Use reentrant passwd and group lookup functions
Fri, Jul 4, 8:15 PM
markj closed D46555: lposix: Use reentrant passwd and group lookup functions.
Fri, Jul 4, 8:15 PM
markj requested review of D51159: nuageinit: Add wrappers for chmod and chown.
Fri, Jul 4, 8:12 PM
markj requested review of D51158: lposix: Clean up the posix namespace definitions.
Fri, Jul 4, 8:12 PM
markj committed rGf9cf745a5084: RELNOTES: Document the addition of inotify (authored by markj).
RELNOTES: Document the addition of inotify
Fri, Jul 4, 3:15 PM
markj committed rGf56e0d4755c5: procstat: Add handling for inotify descriptors (authored by markj).
procstat: Add handling for inotify descriptors
Fri, Jul 4, 2:57 PM
markj committed rG79102e108099: libprocstat: Add handling for inotify descriptors (authored by markj).
libprocstat: Add handling for inotify descriptors
Fri, Jul 4, 2:57 PM
markj committed rG1d8664d69041: examples: Add a demo program for inotify (authored by markj).
examples: Add a demo program for inotify
Fri, Jul 4, 2:57 PM
markj committed rG3965de642c29: linux: Add inotify support (authored by markj).
linux: Add inotify support
Fri, Jul 4, 2:57 PM
markj committed rGd7025d3a3f17: linux: Regenerate system call definitions (authored by markj).
linux: Regenerate system call definitions
Fri, Jul 4, 2:57 PM
markj committed rG739fc14938d5: tests: Add some regression tests for inotify (authored by markj).
tests: Add some regression tests for inotify
Fri, Jul 4, 2:57 PM
markj committed rG5202cfcf93b1: inotify: Add man pages for the new syscalls and for VOP_INOTIFY (authored by markj).
inotify: Add man pages for the new syscalls and for VOP_INOTIFY
Fri, Jul 4, 2:57 PM
markj committed rG4d08e7815d6c: truss: Add a decoder for inotify_add_watch_at(2) (authored by markj).
truss: Add a decoder for inotify_add_watch_at(2)
Fri, Jul 4, 2:57 PM
markj committed rGe02d4fab5453: kdump: Add support for decoding inotify events (authored by markj).
kdump: Add support for decoding inotify events
Fri, Jul 4, 2:57 PM