Page MenuHomeFreeBSD

glebius (Gleb Smirnoff)
User

Projects (6)

User Details

User Since
May 28 2014, 2:27 PM (594 w, 1 d)

Recent Activity

Wed, Oct 15

glebius committed rG4548b9f3a816: unix/stream: plug a corner case when control externalization failed (authored by glebius).
unix/stream: plug a corner case when control externalization failed
Wed, Oct 15, 8:48 PM
glebius committed rGa091d6919c31: socket(2): refactor the manual page (authored by glebius).
socket(2): refactor the manual page
Wed, Oct 15, 6:59 PM
glebius committed rGbbfaff26bf36: unix/stream: repair SO_SNDTIMEO (authored by glebius).
unix/stream: repair SO_SNDTIMEO
Wed, Oct 15, 6:59 PM
glebius committed rG794299ac6876: netlink: in snl_init_writer() don't overwrite error in case of failure (authored by Bruno Silvestre <bruno.silvestre@gmail.com>).
netlink: in snl_init_writer() don't overwrite error in case of failure
Wed, Oct 15, 6:59 PM
glebius committed rG8596810d02a1: tests/carp: kill routed(8) before destroying interfaces (authored by glebius).
tests/carp: kill routed(8) before destroying interfaces
Wed, Oct 15, 6:56 PM
glebius committed rG47b273265025: linux: add translation for MCAST_JOIN_GROUP family of socket options (authored by glebius).
linux: add translation for MCAST_JOIN_GROUP family of socket options
Wed, Oct 15, 6:56 PM
glebius closed D52937: linux: add translation for MCAST_JOIN_GROUP family of socket options.
Wed, Oct 15, 6:56 PM
glebius committed rG479e29157259: linux: make linux_to_bsd_sockaddr() use memory supplied by caller (authored by glebius).
linux: make linux_to_bsd_sockaddr() use memory supplied by caller
Wed, Oct 15, 6:56 PM
glebius committed rGbe7bdb1cf0b0: netinet: do route lookup when asked to join multicast group on ifindex 0 (authored by glebius).
netinet: do route lookup when asked to join multicast group on ifindex 0
Wed, Oct 15, 6:56 PM
glebius closed D52918: netinet: do route lookup when asked to join multicast group on ifindex 0.
Wed, Oct 15, 6:56 PM
glebius closed D52936: linux: make linux_to_bsd_sockaddr() use memory supplied by caller.
Wed, Oct 15, 6:55 PM
glebius committed rGf8b76ec4cac7: tests/hpts: provide stub KTEST module compiled without TCP_HPTS_KTEST (authored by nickbanks_netflix.com).
tests/hpts: provide stub KTEST module compiled without TCP_HPTS_KTEST
Wed, Oct 15, 5:40 PM
glebius accepted D53114: socket: Annotate sooptcopyin() and sooptcopyout() with __result_use_check.
Wed, Oct 15, 4:14 PM
glebius updated the diff for D52918: netinet: do route lookup when asked to join multicast group on ifindex 0.
  • Rebase the test after the flakyness fix.
Wed, Oct 15, 12:50 AM
glebius closed D53081: unix/stream: repair SO_SNDTIMEO.
Wed, Oct 15, 12:30 AM
glebius committed rGead721935251: unix/stream: repair SO_SNDTIMEO (authored by glebius).
unix/stream: repair SO_SNDTIMEO
Wed, Oct 15, 12:30 AM

Tue, Oct 14

glebius committed rGc8359b327104: tests/multicast: fix test flakyness (authored by glebius).
tests/multicast: fix test flakyness
Tue, Oct 14, 11:45 PM
glebius accepted D53108: conf: enable TCP_HPTS_KTEST on debug builds.
Tue, Oct 14, 11:35 PM
glebius committed rG86d17239233e: socket(2): refactor the manual page (authored by glebius).
socket(2): refactor the manual page
Tue, Oct 14, 6:42 PM
glebius committed rG663a571514fb: sockets: add PF_HYPERV alias (authored by glebius).
sockets: add PF_HYPERV alias
Tue, Oct 14, 6:41 PM
glebius closed D52771: socket(2): refactor the manual page.
Tue, Oct 14, 6:41 PM
glebius added a comment to D53081: unix/stream: repair SO_SNDTIMEO.

Do we not block at all when sending on a dgram socket and the receiver has insufficient space?

Tue, Oct 14, 4:29 PM
glebius updated the summary of D53081: unix/stream: repair SO_SNDTIMEO.
Tue, Oct 14, 4:27 PM
glebius updated the diff for D53081: unix/stream: repair SO_SNDTIMEO.
  • Improve test per Alan's review.
Tue, Oct 14, 4:26 PM
glebius added a comment to D53070: pf: remove unused variable allrulelist.
In D53070#1212661, @mjg wrote:

This *is* used by pf_rule_counter_u64_periodic. The goal is to sum up the 32 bit values frequently enough(tm) that they don't overflow.

Tue, Oct 14, 4:09 PM
glebius requested review of D53081: unix/stream: repair SO_SNDTIMEO.
Tue, Oct 14, 4:48 AM
glebius resigned from D53079: unionfs: fix NULL deref on closing an fd passed through SCM_RIGHTS.

I trust Alan's expertise here. IMHO, given that this patch is analog to fusefs one, should be fine. But not expert in the area.

Tue, Oct 14, 4:44 AM
glebius added a comment to D53070: pf: remove unused variable allrulelist.

In pf_krule we got remaining:

Tue, Oct 14, 4:41 AM

Mon, Oct 13

glebius added a comment to D53045: openssl: add a simple smoke test for the legacy provider.

The test uses OPENSSL_CONF to specify the test config file.

Mon, Oct 13, 2:56 AM

Sun, Oct 12

glebius accepted D52991: login.conf: Remove never used ignoretime/ignoretime@.

If it ever existed it existed in lib/libutil/login_ok.c and lib/libutil/login_times.c. Even the original import to FreeBSD does mention ignoretime.

Sun, Oct 12, 10:16 PM
glebius added a comment to D52991: login.conf: Remove never used ignoretime/ignoretime@.

This options plays together with times.allow and times.deny, that are still documented in the login.conf. I think deeper digging is required to see if this is indeed a cruft or actually a functional thing. If it is cruft, then documentation in login.conf on times.allow/times.deny is also cruft.

Sun, Oct 12, 10:10 PM
glebius added a comment to D53045: openssl: add a simple smoke test for the legacy provider.

Does OpenSSL just picks up openssl.conf from current directory?

Sun, Oct 12, 10:03 PM

Fri, Oct 10

glebius added inline comments to D53021: inline most of the considered lost condensation..
Fri, Oct 10, 4:17 PM

Wed, Oct 8

glebius added inline comments to D52975: etc/mtree: Add package tags for Kerberos directories.
Wed, Oct 8, 4:52 PM

Tue, Oct 7

glebius committed rGe7e7bf757ace: krb5: create the KDC database directory (authored by glebius).
krb5: create the KDC database directory
Tue, Oct 7, 9:50 PM
glebius closed D52970: krb5: create the KDC database directory.
Tue, Oct 7, 9:50 PM
glebius requested review of D52970: krb5: create the KDC database directory.
Tue, Oct 7, 9:08 PM
glebius committed rGd9c94e5eefc0: tests/netinet: fix so_reuseport_lb_test flakyness (authored by glebius).
tests/netinet: fix so_reuseport_lb_test flakyness
Tue, Oct 7, 9:07 PM
glebius committed rG488718ff4234: netlink: in snl_init_writer() don't overwrite error in case of failure (authored by Bruno Silvestre <bruno.silvestre@gmail.com>).
netlink: in snl_init_writer() don't overwrite error in case of failure
Tue, Oct 7, 5:49 AM
glebius accepted D52938: x86: Reduce amount of time the MCA lock is held while emitting records.
Tue, Oct 7, 5:43 AM
glebius accepted D52942: x86: Add a way to inject artificial MCA events for testing.
Tue, Oct 7, 5:43 AM
glebius accepted D52943: x86: print MCAs to the console before MCA initialization.
Tue, Oct 7, 5:42 AM
glebius accepted D52946: x86: Allow MCA messages to be rate-limited.
Tue, Oct 7, 5:42 AM

Mon, Oct 6

glebius added a comment to D52942: x86: Add a way to inject artificial MCA events for testing.

Maybe put that under DIAGNOSTIC or similar hiding ifdef? Only one or two people in the world would ever need this. Those who do, definitely would be compiling kernels.

Mon, Oct 6, 6:26 PM
glebius accepted D52934: tcp: improve SEG.ACK validation in SYN-RECEIVED.
Mon, Oct 6, 6:24 PM
glebius requested review of D52937: linux: add translation for MCAST_JOIN_GROUP family of socket options.
Mon, Oct 6, 3:12 PM
glebius requested review of D52936: linux: make linux_to_bsd_sockaddr() use memory supplied by caller.
Mon, Oct 6, 3:12 PM

Sun, Oct 5

glebius requested review of D52918: netinet: do route lookup when asked to join multicast group on ifindex 0.
Sun, Oct 5, 9:52 PM

Sat, Oct 4

glebius updated the diff for D52771: socket(2): refactor the manual page.
  • Add a paragraph on non-blocking mode
Sat, Oct 4, 5:56 AM
glebius added a comment to D52873: sockets: bump socket buffer limit.
In D52873#1207907, @jtl wrote:

My main concern about these three revisions is that they are somewhat susceptible of remote manipulation, and that could make it easier to DoS a server. However, I view that as a tradeoff that the user needs to make, and think an appropriate release note should suffice to warn about these issues.

Sat, Oct 4, 5:19 AM
glebius updated the diff for D52771: socket(2): refactor the manual page.
  • Added sentence on how to set record boundaries with SOCK_SEQPACKET.
Sat, Oct 4, 5:06 AM

Thu, Oct 2

glebius accepted D52872: tcp: bump max snd buffer size for autoscaling.
Thu, Oct 2, 7:50 PM
glebius accepted D52731: sys/netinet6: Use atomic(9) for dad_failures counter.
Thu, Oct 2, 4:56 PM
glebius committed rGa733ea831a00: tests/netinet: add a few missing atf_checks in multicast.sh (authored by glebius).
tests/netinet: add a few missing atf_checks in multicast.sh
Thu, Oct 2, 1:23 PM
glebius added a comment to D52852: vxlan: support moving to a different vnet.

if_vmove bites again? I'm fine with adding more kludges around this problem as long as we all agree that eventually this thing needs to be removed and interfaces shall be fully destroyed and fully instantiated in a different jail.

Thu, Oct 2, 12:41 PM
glebius committed rGb8975e052e80: ucode: use wrmsr_early_safe() for early CPU microcode update (authored by glebius).
ucode: use wrmsr_early_safe() for early CPU microcode update
Thu, Oct 2, 12:01 AM

Wed, Oct 1

glebius accepted D52840: tcp: keep SYN-cache entry when sending of challenge ACK fails.
Wed, Oct 1, 11:14 PM
glebius committed rGd0c8a0b3e6fc: tests/netinet: extend IPv4 multicast testing framework (authored by glebius).
tests/netinet: extend IPv4 multicast testing framework
Wed, Oct 1, 9:25 PM
glebius committed rG36f2eda25171: tests/netinet: rename and extend sendto-IP_MULTICAST_IF.c (authored by glebius).
tests/netinet: rename and extend sendto-IP_MULTICAST_IF.c
Wed, Oct 1, 9:25 PM
glebius committed rGe91fc88bca02: ip(4): fix description of struct ip_mreqn (authored by glebius).
ip(4): fix description of struct ip_mreqn
Wed, Oct 1, 9:25 PM
glebius added a comment to D52816: tcp: cleanup syncache_expand().

I agree with Nick that previous comments were better. You may add assertions that match the checks in the syncookie_expand() case.

Wed, Oct 1, 1:15 PM

Tue, Sep 30

glebius added inline comments to D52731: sys/netinet6: Use atomic(9) for dad_failures counter.
Tue, Sep 30, 4:49 PM

Mon, Sep 29

glebius accepted D52731: sys/netinet6: Use atomic(9) for dad_failures counter.
Mon, Sep 29, 7:35 AM

Sun, Sep 28

glebius committed rGbbec2c9a6d9a: newsyslog: fix one hour rotation with frequent execution (authored by glebius).
newsyslog: fix one hour rotation with frequent execution
Sun, Sep 28, 2:56 PM
glebius committed rG81df9522c8b7: newsyslog: provide -I option to specify default signal (authored by glebius).
newsyslog: provide -I option to specify default signal
Sun, Sep 28, 2:55 PM
glebius committed rGaab45f2aeec8: tests/netlink: fix flaky netlink_sockets:sizes (authored by glebius).
tests/netlink: fix flaky netlink_sockets:sizes
Sun, Sep 28, 2:55 PM
glebius closed D52700: newsyslog: fix one hour rotation with frequent execution.
Sun, Sep 28, 2:55 PM
glebius closed D52701: newsyslog: provide -I option to specify default signal.
Sun, Sep 28, 2:55 PM
glebius accepted D52651: rpcbind: Fix a coredump that occurs when rpcinfo is done.
Sun, Sep 28, 2:44 PM
glebius added a comment to D52651: rpcbind: Fix a coredump that occurs when rpcinfo is done.
if (asprintf(&ret, "%s", (char *)nbuf->buf) < 0)
Sun, Sep 28, 2:24 PM
glebius added inline comments to D52651: rpcbind: Fix a coredump that occurs when rpcinfo is done.
Sun, Sep 28, 2:17 PM
glebius added a comment to D50825: libexec/kgdb: Add a new VNET function and add more scaffolding.

On amd64 only, or does this work on all platforms? IIRC the amd64 flavour is a bit different, see the zpcpu_* macros in amd64/include/pcpu.h vs sys/pcpu.h.

Sun, Sep 28, 12:48 PM
glebius added a comment to D50825: libexec/kgdb: Add a new VNET function and add more scaffolding.

JFYI, for counter(9) this is all what is needed:

define counter_fetch
  set $sum = 0
  set $c = (uintptr_t )$arg0 + (uintptr_t )&__pcpu[0]
  set $n = 0
  while ( $n < mp_ncpus )
    set $sum = $sum + *(uint64_t *)$c
    set $c = (uint64_t *)((char *)$c + sizeof(struct pcpu))
    set $n = $n + 1  
  end
  p/u $sum
end
document counter_fetch
display a counter_u64 value
end
Sun, Sep 28, 10:08 AM
glebius added a reviewer for D52771: socket(2): refactor the manual page: roy_marples.name.
Sun, Sep 28, 9:57 AM
glebius requested review of D52771: socket(2): refactor the manual page.
Sun, Sep 28, 9:57 AM
glebius requested changes to D52651: rpcbind: Fix a coredump that occurs when rpcinfo is done.

I see, in

Sun, Sep 28, 7:31 AM

Sat, Sep 27

glebius added a reviewer for D52701: newsyslog: provide -I option to specify default signal: delphij.
Sat, Sep 27, 3:35 PM
glebius added a reviewer for D52700: newsyslog: fix one hour rotation with frequent execution: delphij.
Sat, Sep 27, 3:35 PM
glebius committed rGd9f03a43f2fe: ucode: use wrmsr_early_safe() for early CPU microcode update (authored by glebius).
ucode: use wrmsr_early_safe() for early CPU microcode update
Sat, Sep 27, 2:31 PM
glebius closed D52643: ucode: use wrmsr_early_safe() for early CPU microcode update.
Sat, Sep 27, 2:31 PM
glebius retitled D52643: ucode: use wrmsr_early_safe() for early CPU microcode update from ucode: try Kostik's wrmsr_early_safe() to ucode: use wrmsr_early_safe() for early CPU microcode update.
Sat, Sep 27, 2:29 PM
glebius added a comment to D52731: sys/netinet6: Use atomic(9) for dad_failures counter.

@jtl is absolutely right with this suggestion. However, what's the point on saving bits and going with uint8_t, especially with portability considerations. We aren't saving any memory here. Why not just use u_int?

Sat, Sep 27, 10:07 AM
glebius updated the diff for D52643: ucode: use wrmsr_early_safe() for early CPU microcode update.
  • Update Intel
  • Remove efi_boot check
Sat, Sep 27, 9:55 AM

Thu, Sep 25

glebius updated the diff for D52643: ucode: use wrmsr_early_safe() for early CPU microcode update.

Make it compilable on i386.

Thu, Sep 25, 9:07 AM
glebius updated the diff for D52701: newsyslog: provide -I option to specify default signal.

Address Benedict's review.

Thu, Sep 25, 8:27 AM
glebius accepted D52713: ng_tty: copy to the correct mbuf.
Thu, Sep 25, 8:22 AM
glebius committed rGb0091036f724: psignal(3): document return values of sig2str() and str2sig (authored by glebius).
psignal(3): document return values of sig2str() and str2sig
Thu, Sep 25, 8:20 AM
glebius closed D52698: psignal(3): document return values of sig2str() and str2sig.
Thu, Sep 25, 8:20 AM
glebius committed rGb00e65ff70a4: newsyslog: use str2sig() instead of own implementation (authored by glebius).
newsyslog: use str2sig() instead of own implementation
Thu, Sep 25, 8:19 AM
glebius closed D52697: newsyslog: use str2sig() instead of own implementation.
Thu, Sep 25, 8:19 AM

Wed, Sep 24

glebius accepted D52702: ng_tty: don't increment m_data.

Thanks a lot for the fix and sorry for the breakage.

Wed, Sep 24, 11:33 AM
glebius accepted D52703: netgraph: ng_device: set curvnet when writing from the cdev.
Wed, Sep 24, 10:49 AM
glebius added a comment to D52651: rpcbind: Fix a coredump that occurs when rpcinfo is done.

What code does call into this function with netlink argument? IMHO, this transport layer abstraction of RPC is a big code bloat where most of the code is never executed. Of course a crash if it happens needs to be fixed.

Wed, Sep 24, 6:36 AM

Tue, Sep 23

glebius requested review of D52701: newsyslog: provide -I option to specify default signal.
Tue, Sep 23, 8:11 PM
glebius requested review of D52700: newsyslog: fix one hour rotation with frequent execution.
Tue, Sep 23, 8:10 PM
glebius added a reviewer for D52698: psignal(3): document return values of sig2str() and str2sig: Doc Committers.
Tue, Sep 23, 7:55 PM
glebius requested review of D52698: psignal(3): document return values of sig2str() and str2sig.
Tue, Sep 23, 7:52 PM
glebius requested review of D52697: newsyslog: use str2sig() instead of own implementation.
Tue, Sep 23, 7:52 PM
glebius committed rGd5a5975f8430: newsyslog: remove leftovers of 20-year old debugging code (authored by glebius).
newsyslog: remove leftovers of 20-year old debugging code
Tue, Sep 23, 7:34 PM
glebius committed rGf221f410920d: tests/newsyslog: don't hardcode -F for newsyslog execution (authored by glebius).
tests/newsyslog: don't hardcode -F for newsyslog execution
Tue, Sep 23, 7:34 PM