Page MenuHomeFreeBSD

glebius (Gleb Smirnoff)
User

Projects (6)

User Details

User Since
May 28 2014, 2:27 PM (581 w, 1 h)

Recent Activity

Today

glebius added a reviewer for D51072: libbsnmp: make snmp_parse_server() more robust when lacking info: ae.
Wed, Jul 16, 4:40 AM

Yesterday

glebius accepted D51329: pf: delay taking the rules lock in pf_test().
Tue, Jul 15, 4:25 PM
glebius added a comment to D50947: ethernet: Move the assertion of ether header sizes back into ethernet.h.

P.S. Why do you want to merge that to a stable branch?

The structs has __packed attribute so it is unlikely to have wrong size. Well I'm a little paranoiac that may happen in some way :)

Tue, Jul 15, 5:10 AM
glebius retitled D51031: libbsnmp: fix local client socket creation from bsnmp: fix local client socket creation to libbsnmp: fix local client socket creation.
Tue, Jul 15, 12:02 AM

Mon, Jul 14

glebius accepted D50947: ethernet: Move the assertion of ether header sizes back into ethernet.h.

Please include the link to the exp-run bugzilla into the commit message.

Mon, Jul 14, 6:01 PM

Fri, Jul 11

glebius added inline comments to D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
Fri, Jul 11, 4:05 AM
glebius updated the diff for D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.

Add missing 'break;'. We aren't using _SAFE iterator here, and we don't
expect any more matches.

Fri, Jul 11, 3:49 AM
glebius updated the summary of D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
Fri, Jul 11, 3:45 AM
glebius updated the diff for D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.

As Alexander noted, a single label doesn't store information about the
entire pool, but only about one top-level vdev. Thus, we should blow
away only this part of the vdev tree, not entire tree.

Fri, Jul 11, 3:40 AM
glebius committed rG8e8fd66ca06f: libsa/zfs: fail earlier in vdev_probe() if can't find ZPOOL_CONFIG_GUID (authored by glebius).
libsa/zfs: fail earlier in vdev_probe() if can't find ZPOOL_CONFIG_GUID
Fri, Jul 11, 3:28 AM
glebius committed rG03f3dc3189e5: libsa/zfs: don't hardcode root vdev name, use the real pool name (authored by glebius).
libsa/zfs: don't hardcode root vdev name, use the real pool name
Fri, Jul 11, 3:21 AM
glebius committed rG8949cdd9af53: libsa/zfs: don't read and store ZPOOL_CONFIG_VDEV_CHILDREN in vdev_probe() (authored by glebius).
libsa/zfs: don't read and store ZPOOL_CONFIG_VDEV_CHILDREN in vdev_probe()
Fri, Jul 11, 3:21 AM
glebius committed rGf1eceb963caf: libsa/zfs: add error printf() in zfs_spa_init() (authored by glebius).
libsa/zfs: add error printf() in zfs_spa_init()
Fri, Jul 11, 3:07 AM
glebius accepted D51234: nlmsg_report_err_msg: add __printflike annotation.
Fri, Jul 11, 12:48 AM

Thu, Jul 10

glebius added a comment to D51143: sysctl net.inet.tcp.ktlslist: allow snd_tag_status_str() to sleep.

All this hardware TLS assist is a crazy layer violation. But this is a wave a can't stand against.

Thu, Jul 10, 5:30 AM

Wed, Jul 9

glebius accepted D51143: sysctl net.inet.tcp.ktlslist: allow snd_tag_status_str() to sleep.

I can't say that I am happy with this change, sorry. But looks like you need it, so let it be so.

Wed, Jul 9, 7:02 PM
glebius added a comment to D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
In D51219#1169582, @mav wrote:

TXGs are committed once there is some minimal amount of dirty data reached, but no less than every 5 seconds if there were any changes. Under constant data stream TXGs are committed as fast as underlying storage can handle metadata updates and cache flushes, which may vary between some fractions of a second to multiple seconds. Many zpool and zfs commands do it forcefully. zpool sync does just that, for example, if you need up to date stats, etc.

Wed, Jul 9, 6:11 PM
glebius accepted D51086: rpcsec_gss.c: Avoid panics in clnt_nl_call.

Works as a panic prevention :)

Wed, Jul 9, 5:49 PM
glebius accepted D51084: clnt_rc.c: Fix the case where a TLS handshake fails.
Wed, Jul 9, 5:48 PM
glebius added a comment to D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
In D51219#1169554, @mav wrote:

The labels are updated on each TXG commit, that may be many times per second. So the races are quite probable there, and unlike resurrecting old disk are completely unpredictable and also harder to recover by just removing that disk.

Wed, Jul 9, 5:24 PM
glebius accepted D51083: rpctls_impl.c: Fix handling of hung rpc tls daemon.
Wed, Jul 9, 5:12 PM
glebius added a comment to D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
In D51219#1169548, @mav wrote:

I wonder what happen if system crash after updating one label on only one disk out of several. In such case vdev_label_read_config() will return config of that vdev for that TXG. But no other vdev have config for that TXG yet. So if we already probed some other vdevs, the added vdev_free(spa->spa_root_vdev); will wipe root vdev and all the children configs for the pool and restart it from scratch, populating only vdev where the currently probed disk belongs. But already probed vdevs I guess will not be re-probed, leaving pool-wide configuration incomplete. It seems not implemented here, looking on UINT64_MAX always passed to vdev_label_read_config(), but IIRC normal pool import code tries to find maximum common TXG where it has sufficient quorum. It might not be the maximum TXG, but the previous one, or theoretically allowed even one before that, if something went very wrong.

Wed, Jul 9, 5:02 PM
glebius requested review of D51219: libsa/zfs: prefer a vdev with newer txg to a stale one.
Wed, Jul 9, 5:55 AM

Fri, Jun 27

glebius committed rG24ebe9fa0ec1: psm: fix build (authored by glebius).
psm: fix build
Fri, Jun 27, 10:47 PM
glebius updated the summary of D51070: libbsnmp: make binding of client UNIX socket optional and configurable.
Fri, Jun 27, 10:31 PM
glebius updated the diff for D51070: libbsnmp: make binding of client UNIX socket optional and configurable.

Make mktemp+bind optional only for the SOCK_STREAM case.

Fri, Jun 27, 10:31 PM
glebius accepted D51076: gone_in(9): update man page.

Thanks! Didn't know this page exists.

Fri, Jun 27, 6:29 PM
glebius added a comment to D51070: libbsnmp: make binding of client UNIX socket optional and configurable.

Note: the binding is actually needed when we run in SOCK_DGRAM mode. So this isn't a final revision.

Fri, Jun 27, 5:17 PM

Thu, Jun 26

glebius accepted D51069: tcp: disable nolocaltimewait and deprecate it.
Thu, Jun 26, 9:56 PM
glebius added a comment to D51031: libbsnmp: fix local client socket creation.

Note: 4 more reviews stacked on top of this one, if anyone interested.

Thu, Jun 26, 7:59 PM
glebius added a reviewer for D51031: libbsnmp: fix local client socket creation: syrinx.
Thu, Jun 26, 7:59 PM
glebius added a reviewer for D51070: libbsnmp: make binding of client UNIX socket optional and configurable: syrinx.
Thu, Jun 26, 7:58 PM
glebius added a reviewer for D51071: libbsnmp: define SNMP_DEFAULT_LOCAL that matches default server config: syrinx.
Thu, Jun 26, 7:58 PM
glebius added a reviewer for D51072: libbsnmp: make snmp_parse_server() more robust when lacking info: syrinx.
Thu, Jun 26, 7:58 PM
glebius added a reviewer for D51073: bsnmpget: print SNMP library reported error on failures: syrinx.
Thu, Jun 26, 7:57 PM
glebius requested review of D51073: bsnmpget: print SNMP library reported error on failures.
Thu, Jun 26, 7:57 PM
glebius requested review of D51072: libbsnmp: make snmp_parse_server() more robust when lacking info.
Thu, Jun 26, 7:57 PM
glebius requested review of D51071: libbsnmp: define SNMP_DEFAULT_LOCAL that matches default server config.
Thu, Jun 26, 7:57 PM
glebius requested review of D51070: libbsnmp: make binding of client UNIX socket optional and configurable.
Thu, Jun 26, 7:56 PM
glebius added a comment to D51069: tcp: disable nolocaltimewait and deprecate it.

We could also use SYSCTL_PROC() instead of SYSCTL_BOOL() and then add gone_in() to emphasis that the sysctl-variable is deprecated. Just let me know if this is wanted.

Thu, Jun 26, 7:54 PM
glebius accepted D50912: tcp: restrict the contents of tcp_timer.h to the kernel context.

We will sort out how to correctly put #error later.

Thu, Jun 26, 3:19 PM

Wed, Jun 25

glebius added a comment to D51031: libbsnmp: fix local client socket creation.
In D51031#1164827, @des wrote:

I assume the original concern here was that mktemp() is subject to TOCTTOU races. Thus, it would be better to use mkdtemp() to create a unique temporary directory, and create our socket inside that directory.

Wed, Jun 25, 4:44 PM
glebius added a comment to D51031: libbsnmp: fix local client socket creation.

I am really puzzled how D34550 was created and how did it pass your review :)

Wed, Jun 25, 2:18 AM
glebius added a reverting change for rG81e0e7b9e36d: bsnmp: Use mkstemp when creating clients local socket: D51031: libbsnmp: fix local client socket creation.
Wed, Jun 25, 2:18 AM
glebius requested review of D51031: libbsnmp: fix local client socket creation.
Wed, Jun 25, 2:18 AM
glebius added a reverting change for D34550: bsnmp: Use mkstemp when creating clients local socket: D51031: libbsnmp: fix local client socket creation.
Wed, Jun 25, 2:17 AM
glebius committed rG96a241a35905: bsddialog: in textbox mode differentiate between Exit and Extra buttons (authored by emax).
bsddialog: in textbox mode differentiate between Exit and Extra buttons
Wed, Jun 25, 1:18 AM
glebius committed rG06ec77a42d9f: gone_in: initialize va before passing it to gone_panic() (authored by glebius).
gone_in: initialize va before passing it to gone_panic()
Wed, Jun 25, 1:18 AM
glebius closed D48668: bsddialog: in textbox mode differentiate between Exit and Extra buttons.
Wed, Jun 25, 1:17 AM

Tue, Jun 24

glebius committed rG994bec47ee7f: tests/kqueue: repair flaky test (authored by glebius).
tests/kqueue: repair flaky test
Tue, Jun 24, 10:31 PM
glebius closed D50996: tests/kqueue: repair flaky test.
Tue, Jun 24, 10:30 PM
glebius committed rGd17cbe46983c: sendfile: retire SF_SYNC (authored by glebius).
sendfile: retire SF_SYNC
Tue, Jun 24, 10:25 PM
glebius committed rG01addd846c64: gone_in: make it __printflike() (authored by glebius).
gone_in: make it __printflike()
Tue, Jun 24, 10:25 PM
glebius committed rGeabb9f600001: gone_in: enforce printing only once (authored by glebius).
gone_in: enforce printing only once
Tue, Jun 24, 10:25 PM
glebius closed D50783: gone_in: make it __printflike().
Tue, Jun 24, 10:25 PM
glebius closed D50784: gone_in: enforce printing only once.
Tue, Jun 24, 10:25 PM
glebius added a comment to D50996: tests/kqueue: repair flaky test.

Agreed, let's try 495 and see what CI reports after a week.

Tue, Jun 24, 2:10 AM

Mon, Jun 23

glebius requested review of D50996: tests/kqueue: repair flaky test.
Mon, Jun 23, 10:00 PM
glebius accepted D50796: counter(9): rate limit periods may be more than 1 second.

No objection from me. Thanks for the change and sorry for the delay. But please wait for the stabweek closure.

Mon, Jun 23, 6:59 PM
glebius accepted D50975: ifnet: Eliminate unnecessary synchronization of the interface index in the link layer address.

Thanks!

Mon, Jun 23, 6:58 PM
glebius added a comment to D50912: tcp: restrict the contents of tcp_timer.h to the kernel context.

Thanks, @imp! What if instead of allowing the header but masking its internals, we want to eliminate any use of a header by userland applications? Some headers have #error no user serviceable parts inside and AFAIU all of them are listed in badfiles.inc. Is this the way to go?

Mon, Jun 23, 6:56 PM

Sat, Jun 21

glebius added a comment to D50962: rpctls_impl.c: Add KRPC_CURVNET_SET() and KRPC_CURVNET_RESTORE().

What is the problem we are trying to fix here? What is the panic trace?

Sat, Jun 21, 7:56 PM
glebius added a comment to D50961: rpctls_impl.c: Fix handling of daemon failure.

I clicked on "Accept", trusting your expertise. The change looks correct, with small comment nit. But I didn't test it. And to be fair the gist of the code already washed out of my L3 brain cache :)

Sat, Jun 21, 5:38 PM
glebius accepted D50961: rpctls_impl.c: Fix handling of daemon failure.
Sat, Jun 21, 5:37 PM

Fri, Jun 20

glebius committed rG220cdd1b3941: libc/rpc/svc_nl: plg a memory leak (authored by glebius).
libc/rpc/svc_nl: plg a memory leak
Fri, Jun 20, 10:38 PM

Thu, Jun 19

glebius committed rG8d98d889e1fc: libc/getaddrinfo(2): return EAI_AGAIN on nameserver timeout (authored by glebius).
libc/getaddrinfo(2): return EAI_AGAIN on nameserver timeout
Thu, Jun 19, 8:01 PM
glebius committed rG421d49371d33: mount_nfs: make temporary DNS failure non-fatal with background mode (authored by glebius).
mount_nfs: make temporary DNS failure non-fatal with background mode
Thu, Jun 19, 8:01 PM
glebius closed D50891: tcp: provide sysctl for the maximum retransmission timeout.
Thu, Jun 19, 3:29 AM

Wed, Jun 18

glebius committed rG46023d54c7c2: tcp: fixup wording in comment (authored by glebius).
tcp: fixup wording in comment
Wed, Jun 18, 3:48 AM
glebius added a reviewer for D50912: tcp: restrict the contents of tcp_timer.h to the kernel context: imp.
Wed, Jun 18, 3:45 AM
glebius added a comment to D50912: tcp: restrict the contents of tcp_timer.h to the kernel context.

There is no #warning or #error in case the header file is not included in the kernel context, since then this header file would need to be added to tools/build/test-includes/badfiles.inc, but that list should not grow.

Wed, Jun 18, 3:44 AM

Tue, Jun 17

glebius committed rGa7596f6c63e9: snmp_mibII: gc fetching of kern.clockrate (authored by glebius).
snmp_mibII: gc fetching of kern.clockrate
Tue, Jun 17, 5:28 PM
glebius committed rGb3d0562dd845: LINT: add KDTRACE_MIB_SDT to catch inconsistencies with struct tcpstat (authored by glebius).
LINT: add KDTRACE_MIB_SDT to catch inconsistencies with struct tcpstat
Tue, Jun 17, 4:54 PM
glebius committed rGb59753f1d55d: tcp: provide a useful comment for struct tcpstat (authored by glebius).
tcp: provide a useful comment for struct tcpstat
Tue, Jun 17, 4:54 PM
glebius committed rG2d300d376d5b: tcp: add missing SDT probes (authored by glebius).
tcp: add missing SDT probes
Tue, Jun 17, 4:54 PM
glebius added a comment to D50796: counter(9): rate limit periods may be more than 1 second.

Manual page counter.9 also needs to be updated.

Tue, Jun 17, 4:17 PM
glebius added a comment to D50796: counter(9): rate limit periods may be more than 1 second.

I totally agree with the period feature. Thanks!

Tue, Jun 17, 4:15 PM
glebius added a reviewer for D50796: counter(9): rate limit periods may be more than 1 second: jtl.
Tue, Jun 17, 4:15 PM
glebius added a comment to D50840: tcp: cleanup timer initialisations.

bsnmpd done. We can put the entire header under _KERNEL. We also can add #error in there to catch all incorrect inclusions of this header.

Tue, Jun 17, 3:56 PM
glebius committed rGdf4b9fa460ab: snmp_mibII: use sysctl(3) to read min/max TCP retransmission timeouts (authored by glebius).
snmp_mibII: use sysctl(3) to read min/max TCP retransmission timeouts
Tue, Jun 17, 3:55 PM
glebius committed rG552d17804629: tcp: provide sysctl for the maximum retransmission timeout (authored by glebius).
tcp: provide sysctl for the maximum retransmission timeout
Tue, Jun 17, 3:55 PM
glebius committed rGfaa6aa7723aa: tcp: remove CTLFLAG_NEEDGIANT from sysctl(9) handlers related to timers (authored by glebius).
tcp: remove CTLFLAG_NEEDGIANT from sysctl(9) handlers related to timers
Tue, Jun 17, 3:55 PM
glebius closed D50892: snmp_mibII: use sysctl(3) to read min/max TCP retransmission timeouts.
Tue, Jun 17, 3:54 PM
glebius closed D50890: tcp: remove CTLFLAG_NEEDGIANT from sysctl(9) handlers related to timers.
Tue, Jun 17, 3:54 PM
glebius committed rG1396e87a37b6: tcp: count two previously ignored kinds of syncookie failures (authored by glebius).
tcp: count two previously ignored kinds of syncookie failures
Tue, Jun 17, 3:51 PM
glebius committed rG0809d26de24d: tcp: unlock before syncookie_expand() (authored by glebius).
tcp: unlock before syncookie_expand()
Tue, Jun 17, 3:51 PM
glebius closed D50898: tcp: unlock before syncookie_expand().
Tue, Jun 17, 3:51 PM
glebius closed D50897: tcp: increase tcps_sc_recvcookie only in the syncache_expand().
Tue, Jun 17, 3:51 PM
glebius committed rG3ed8d5645dd4: tcp: increase tcps_sc_recvcookie only in the syncache_expand() (authored by glebius).
tcp: increase tcps_sc_recvcookie only in the syncache_expand()
Tue, Jun 17, 3:51 PM
glebius committed rGe9e6a025b452: tcp: refactor debug function syncookie_cmp() (authored by glebius).
tcp: refactor debug function syncookie_cmp()
Tue, Jun 17, 3:51 PM
glebius closed D50899: tcp: count two previously ignored kinds of syncookie failures.
Tue, Jun 17, 3:51 PM
glebius committed rG6538742c1aac: tcp: rename syncookie_lookup() into syncookie_expand() and make it bool (authored by glebius).
tcp: rename syncookie_lookup() into syncookie_expand() and make it bool
Tue, Jun 17, 3:51 PM
glebius closed D50896: tcp: refactor debug function syncookie_cmp().
Tue, Jun 17, 3:51 PM
glebius closed D50895: tcp: rename syncookie_lookup() into syncookie_expand() and make it bool.
Tue, Jun 17, 3:51 PM
glebius updated the diff for D50891: tcp: provide sysctl for the maximum retransmission timeout.
  • Don't forget rack/bbr.
  • Document the new sysctl and improve paragraph.
Tue, Jun 17, 4:35 AM
glebius requested review of D50898: tcp: unlock before syncookie_expand().
Tue, Jun 17, 4:16 AM
glebius requested review of D50899: tcp: count two previously ignored kinds of syncookie failures.
Tue, Jun 17, 4:16 AM
glebius requested review of D50897: tcp: increase tcps_sc_recvcookie only in the syncache_expand().
Tue, Jun 17, 4:16 AM
glebius requested review of D50896: tcp: refactor debug function syncookie_cmp().
Tue, Jun 17, 4:16 AM
glebius requested review of D50895: tcp: rename syncookie_lookup() into syncookie_expand() and make it bool.
Tue, Jun 17, 4:15 AM