Activity
From 2013-04-18 to 2013-05-17
2013-05-17
-
11:30 PM Bug #3772: consider raising default descriptor soft limit
- SmartOS patches the limit (via the default /etc/system) such that soft == hard by default, at 65536
-
- We by default have the rlimit file descriptor soft limit set to 256 (or at least, that's what it is on my systems). ...
-
06:21 PM Feature #3113: ztest should use watchpoints to protect frozen arc bufs
- Ok, that's good to know. After adapting the debug patch for Linux to use mprotect(2) I'm still able to fairly easily...
-
01:10 PM Feature #3113: ztest should use watchpoints to protect frozen arc bufs
- The buffer modified while frozen issue still exists. I have been trying to isolate this but have not had much time re...
-
03:48 PM Bug #3664: pkg/Makefile openssl logic should be removed
- Done,
Webrev coming up
2013-05-16
-
- Was the root cause of the 'buffer modified while frozen!' error ever identified? I see that the arc watchpoint patch...
-
09:56 PM Bug #3771: kmdb hangs when reading an invalid address
- Just quickly glancing through the code it looks like kmdb_kvm.c does eventually include setjmp.h which does define se...
-
- Wonderful analysis, thanks for that.
The compiler should be issueing a warning in this case (perhaps another that ... -
- Jeremy Jones did some heroic investigation of this. His write-up of the problem is:...
2013-05-15
-
01:36 PM Bug #3368: NFSv4 and slow OS X Mountain Lion (10.8.2) Client
- Marcel Telka wrote:
> Okay. I'm closing this. Once you have some info showing what can we do from the OI/illumos sid...
2013-05-14
-
- Resolved in a63a4ea
-
- Resolved in 3197aa6
2013-05-13
-
11:40 PM Bug #3769: Implement SOCK_NONBLOCK flag to socket()
- A more complete test suite that finds existing problems including the related socketpair issues may be found at https...
-
06:08 PM Bug #3769 (Resolved): Implement SOCK_NONBLOCK flag to socket()
- With the integration of 3666, the socket() implementation treats SOCK_NONBLOCK as a noop. This causes applications to...
-
09:57 PM Bug #3770 (Resolved): ipmi drivers hangs due to attach/detach/attach cycle
- 36657:May 11 11:06:02 myhost ipmi: [ID 550463 kern.warning] WARNING: Timed out waiting for GET_DEVICE_ID
Noticed t... -
04:29 PM Bug #3736: cp -r shouldn't allow infinite loops
- Possible fix is attached
2013-05-11
-
07:56 PM Bug #3768 (Closed): fnmatch(5) is worded poorly
- Specifically, "The following patterns matching a single character match a single character". This seems like a tautol...
-
03:59 PM Feature #3731 (Feedback): Update nawk to version 20121220
- Yuri Pankov wrote:
> At least, webrev(1) breaks with newer nawk versions... Please make sure you don't break anythin...
2013-05-10
-
- Resolved in dea290b
-
- http://cr.illumos.org/~webrev/0xffea/nawk-continue-01/
-
07:23 AM Bug #3767 (New): nwamcfg dns-nameservice-service not dual stack
- When setting up a DNS server for dual stack IPv6/IPv4 network connectivity I noticed that nwamcfg's dns-nameservice-s...
2013-05-09
-
- The auditd(1M) and similar documentation that ships with illumos refers one to files in /etc/security (specifically /...
-
11:24 AM Bug #3326: VMWare: mutex panic during shutdown
- No, I'm not using open-vm-tools, but the solaris tools provided by ESXi. This only seems to be an issue shutting dow...
-
- Ah, it's part of the vmware tools.
Nothing stands out as overly broken looking in the open-vm-tools code I can fin... -
- In those dumps we have:...
-
- vmcore.0.gz and unix.0.gz
-
- Sent review request.
2013-05-08
-
- Will do. Hopefully tonight...
-
- If this is easily repeatable, could you please boot with -kd, and when you stop at the kmdb prompt on the console, ut...
-
07:45 PM Bug #3326 (In Progress): VMWare: mutex panic during shutdown
- I added the panic stack from vmcore.5 to the Description field and reopened the bug.
-
01:05 PM Bug #3751: ALUA mode: "/usr/demo/comstar/bin/aluaadm enable 1" sets "ALUA Node" = 0
- The bug and the fix come from Paul Griffiths-Todd. I also included a small manpage fix from Paul.
Webrev: http://c... -
12:12 AM Bug #3765 (New): fast reboot fails when kernel parameters get too long
- a fast reboot initiated via
@/usr/sbin/reboot -f -- dataset -B kernel,parameters@
will fail if the part after "...
2013-05-07
-
06:57 PM Feature #3525: Persistent L2ARC
- On a modern CPU fletcher-4 achieves in excess of 4GB/s/core in hashing performance, whereas SHA-256 (in hand-written ...
-
06:47 PM Feature #3525: Persistent L2ARC
- Perhaps use SHA256 rather than Fletcher, or offer it as an option when an L2ARC device is created? SHA256 has instruc...
2013-05-06
-
- Old nawk:...
-
- s/recursive/reentrant/ of course. I blame Monday morning.
-
- This reentrant variant of qsort(3C) is used in modern software such as libgit2. Both BSD and GNU have implementations.
-
- Newer nawk versions don't permit code like this:...
-
- The Datagram Congestion Control Protocol (DCCP) is a transport protocol that provides bidirectional unicast connectio...
2013-05-05
-
03:31 PM Bug #3758: RFE: Would like "hostname -s"
- I'd recommend combining the two changes into one. You can certainly have one webrev that references two bug reports.
-
- Got the code done, needs review. Waiting on 3430 (this bug modifies the same lines in the files, so a conflict is ine...
2013-05-04
-
- Awaiting review feedback for webrev at: http://cr.illumos.org/~webrev/skiselkov/2987/
2013-05-03
-
11:58 PM Bug #3736: cp -r shouldn't allow infinite loops
- I've started taking a look at this. Don't have permissions to assign bug to myself, but will update with progress.
... -
11:36 PM Bug #3719 (Resolved): so_socketpair syscall should preserve FD_CLOEXEC flag
- garrett@openindiana{115}> git log
commit 5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e
Author: Theo Schlossnagle <jesus@... -
- garrett@openindiana{115}> git log
commit 5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e
Author: Theo Schlossnagle <jesus@... -
- garrett@openindiana{115}> git log
commit 5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e
Author: Theo Schlossnagle <jesus@... -
- garrett@openindiana{115}> git log
commit 5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e
Author: Theo Schlossnagle <jesus@... -
- garrett@openindiana{115}> git log
commit 5dbfd19ad5fcc2b779f40f80fa05c1bd28fd0b4e
Author: Theo Schlossnagle <jesus@... -
09:36 PM Feature #3760 (New): svcadm restart should accept -s (synchronous) option
- svcadm enable and disable accept -s to work synchronously and return non-zero status in case of failure. Without the ...
-
- Resolved in 6136c58
-
06:52 PM Bug #3759 (Resolved): egrep(1) and fgrep(1) -s flag does not hide -c output
- The legacy option -s in egrep(1) and fgrep(1) (which documented to behave the same as POSIX option -q instead of the ...
-
-
03:20 PM Bug #2702: ZFS sharenfs strangeness with anon=
- It looks like this issue has not been touched in for ever. It is most definitely still an issue. I ran into it today,...
2013-05-02
-
-
- I've got a webrev for this up at http://labs.omniti.com/people/cnehren/webrevs/issue3430/. It implements -f rather th...
-
- The bug also exists in oi_151a3 and is 100% reproducible too.
-
08:51 PM Feature #3737 (In Progress): grep does not support -H option
- Currently waiting for reviews.
-
- The -s flag (only print the hostname, not the fqdn) is available in other Unixes and would be nice to have here.
2013-05-01
-
- This crash dump is from oi151a7 after a "zfs diff"
In case the vmcore.0 or vmdump.0 are needed - let me know where I... -
09:05 AM Feature #3756 (Closed): want lz4 support for metadata compression
- Metadata on ZFS is compressed using the *lzjb* algorithm by default.
The advantages of lz4 compression may be used ... -
- What's going on is when we ask the kernel for probes in dt_probe_lookup, the provider isn't part of the pattern we ha...
-
- We're crashing when parsing the script actually. And actually, provider name is immaterial, dtrace_xstr2desc isn't f...
-
- Most of the work actually done; just need to run a lint build and submit for review.
2013-04-30
-
10:02 PM Bug #3755: dladm show-aggr documentation
- In man dladm, "OBYTES Number of bytes received on this link." should be "OBYTES Number of bytes sent on this link."
-
- "OBYTES Number of bytes received on this link." should be "OBYTES Number of bytes sent on this link."
-
- which(1) does lie to you, but that's because it's intended for csh, not ksh. For ksh, you should use whence(1) inste...
-
- When ksh93 integrated, it was _supposed_ to only interpose over real commands based on $PATH if the real command was ...
2013-04-29
-
09:11 PM Bug #3594: Accessing a password protected CIFS share from an android device always fail
- Add info from original author that he posted on jcifs@lits.samba.org
X CIMO
2013-04-07 18:42:57 GMT
I have see... -
- The assert trips because the provider name we get back is "ilc" unadorned, but the provider name associated with the ...
2013-04-28
-
08:14 PM Feature #3753: Add zfs receive property support (zfs recv -o)
- This was meant as a feature request and not a bug. I missed specifying the feature type on creation. My apologies.
-
- This is a request to add zfs receive property support (zfs recv -o) to Illumos. Solaris 11 added this feature after ...
-
03:10 AM Feature #3731: Update nawk to version 20121220
- At least, webrev(1) breaks with newer nawk versions... Please make sure you don't break anything.
2013-04-25
-
11:56 PM Bug #3691: setgroups() needs a sorted GID list for more than 16 groups
- also see my comment here in the initial bug report about the 16group issue: https://www.illumos.org/issues/3577#note-2
-
10:13 PM Bug #3752 (New): want more verifiable dbuf user eviction
- The current dbuf user eviction process has a few issues that make it difficult to verify that they work properly:
- ... -
- This is a bug with /usr/demo/comstar/bin/aluaadm; essentially the code portion that sets the nodeid is incorrect as i...
-
01:16 PM Bug #3740: Poor ZFS send / receive performance due to snapshot hold / release processing
- Webrev for this is now here: http://cr.illumos.org/~webrev/steveh/illumos-3740/
-
10:03 AM Bug #3374: usage of 'dladm' does not match to its man page
- Also the "dladm show-link -S" option isn't documented either on the command line or in the man page.
-
- Works for me:...
-
- I have been informed by Steven Hartland <smh@FreeBSD.org> that enoent_ok needs to be respected in this case, too.
At... -
- Debugging a change to libdtrace, I lofs mounted my development copies over the copies in /usr (I had my reasons...)
...
2013-04-24
-
- From the original change log:
This restriction appears to have been put in place to avoid errors with
upd... -
- New patch version attached, this eliminates the separate "release temporary holds" code path creating a single path f...
-
- ...
-
03:52 AM Bug #3721: sys/ctype.h should pragma inline isxdigit
- Testing with the current Studio C compiler, the pragma is ignored. The function is inlined from -XO4. If the functi...
2013-04-23
-
- Changes made to ZFS should to retain C++ compatibility primarily involve renaming variable names that use C++ reserve...
-
- From the original change log:
Change 477353 by justing@justing-ns1 on 2011/02/04 10:11:30
Remove C cons... -
- This enables ZFS to be used directly in C++ programs. The most notable current example is FreeBSD's zfsd, which is a...
-
- There are bugs that preclude the txg commit callbacks from working at all:
- The callback list move arguments are re... -
- From the original change log:
It was possible for a reference to be added even with the lock held, and
for refere... -
- From the original change log:
Make zpool create -m <mountpoint> and -O mountpoint=<mountpoint> equivalent.
... -
- zfs_unmount_snap() is currently void. We should, at minimum, return an error when one occurs internally, and if feas...
-
- Specific areas that currently need to be fixed:
- zap_cursor_fini() is not always called in error cases.
- zap_tabl... -
- A number of areas are vulnerable to refcount leaks, particularly in error paths. Existing refcount leaks need to be ...
-
- Various areas of ZFS have style issues, e.g.:
- Lack of (or too much) whitespace in areas
- Bad spelling
- Dense c... -
- Various areas of ZFS need better comments. This bug is probably more of an ongoing effort than a point-in-time insta...
-
- ...
-
- ...
-
- Okay, I think I figured out why the dump zvol was too small. I only made an 8GB boot disk :( So I created a dump zvo...
-
- Well, damn, omnios created the dump zvol as only being 512MB so I think the dump is crap. I'm recreating it - stay t...
-
- Okay, unix.4.gz and vmcore.4.gz are at:
http://www.druber.com/illumos/3326/ -
04:09 PM Bug #3326: VMWare: mutex panic during shutdown
- I have ported open-vm-tools to DilOS and have no crashes with it on VMware ESXi 4,5,5.1.
but i have no illumos-gate ... -
- I can repro this at will using omnios as well. I will provide a dump. I am surprised no-one has tried to reproduce ...
-
- For our own uses we intentionally try to prevent the compiler making inlining decisions we haven't directed. More th...
-
- Is the pragma necessary at all? Studio or gcc would inline these in an optimised build either way.
-
- I've got a fairly simple setup here which has a hierarchy of zfs volumes which we're looking to sync between two node...
-
- The GNU manual states "This option is probably never needed on modern computer systems.". Is adding this worth the e...
-
- This issue only appears if the variable is initialised at declaration. Fortunately most "static volatile" variables ...
2013-04-22
-
- I am attaching a combined simplified patch, including removal of the three lines from dsl_prop.c
-
08:23 PM Bug #3739: cannot set zfs quota or reservation on pool version < 22
- The new patch looks like it will work. I think you can simplify the code a little bit by moving the history_log into...
-
- I am attaching an updated patch.
-
- Your patch would have use set the "quota" and "reservation" properties in dd_props_obj even on old version pools. Is...
-
09:43 PM Bug #1552: BMC on Dell r815 stops functioning when OI boots
- Last I tried it, the blob from Broadcom directly panics/fails to load (I forget which) on Illumos because of some sym...
-
- I'll try version 7.4.1 from Broadcom tomorrow on one of my machines. If that fixes the issue, I suggest we close this.
-
- AFAICS there's not been any changes other than a rename to the blob since it was committed.
https://github.com/joy... -
-
08:01 PM Bug #1552: BMC on Dell r815 stops functioning when OI boots
- I've found that the same issue affects the iDRAC on a Dell T610 server. The iDRAC integrated management controller is...
-
-
-
- Resolved in 461686c
-
- Resolved in a713f26
-
- Resolved in b44bad0
2013-04-21
-
- Starting with revision 13973:4972ab336f54 (3464 zfs synctask code needs restructuring) the process of setting quota a...
2013-04-20
-
- I've now dropped the LOGNAME_MAX_ILLUMOS symbol and instead increased LOGNAME_MAX from 8 to 32. This change means th...
-
11:57 AM Feature #3737 (Resolved): grep does not support -H option
- Our grep could be more compatible with GNU grep if we added support for the -H option, which causes grep to always pr...
-
- http://www.opensource.apple.com/source/mDNSResponder/
-
08:52 AM Bug #3595: Mac OS X clients have problems copying files with Finder
- Which version of OS X?
With or without an extended attribute on an original file, *expect an extended attribute t... -
- Resolved in 5918f984
-
- Resolved in ef16f6b5
-
- Resolved in ef16f6b5
-
- Resolved in 9e647765
-
- Make the xpg4 grep the default one, and kill the {,e,f}grep monstrosity already.
2013-04-19
-
- If you do...
-
- It'd be great, for those of us who sometimes want to pass in extra compiler flags to facilitate debugging (be that pa...
-
- Resolved in db8a384
-
- Resolved in 22ce014
-
- Resolved in 22ce014
-
- Resolved in 22ce014
-
- Resolved in 22ce014
-
- Resolved in 8c430e59
-
- Resolved in b89a2c3
-
- Resolved in 43fab1a
-
- It's saying "executable file name" to try to suggest it matches on the basename, not the full path. That's admittedl...
-
- support for the -q option was added to grep, but not to egrep.
-
- tar could support the -s option (detailed here http://www.gnu.org/software/tar/manual/html_node/Same-Order.html) for ...
-
- http://www.cs.princeton.edu/~bwk/btl.mirror/
-
01:17 AM Feature #3730 (New): expose additional functions from getifaddrs.c
- The file "getifaddrs.c":http://src.illumos.org/source/xref/illumos-gate/usr/src/lib/libsocket/inet/getifaddrs.c conta...
-
- _NB: After #3728 integrates, there will be a getifaddrs(3SOCKET) which mentions this bug. Once this is fixed, the *B...
-
- Presently *getifaddrs* appears to enumerate network interface addresses only from *AF_INET* and *AF_INET6*. This is ...
-
- I have written this manual page and will be integrating it into *illumos-joyent* soon. After that, I will complete a...
-
- We should provide a *getifaddrs(3SOCKET)* manual page.
_NB: This bug mirrors Joyent bug OS-2149._ -
- The theory that we're being too strict to R_AMD64_32 addends is accurate. In the amd64 do_reloc_* implementations, w...
2013-04-18
-
- If you try to use spell -b, it segvs behind the scenes and acts as if everything is correct. I'm guessing constant s...
-
07:45 PM Bug #3726 (New): libdtrace's use of libelf is non-portable
- When processing an object file that contains USDT probes in a static function, libdtrace does some libelf magic that ...
-
- Since the MIN macro evaluates both arguments twice, the value of count might be set to NFS_MAXDATA unexpectedly here ...
-
- Minor update, I forgot to remove the comments around the STACKLIMIT lines. Attached is a fix for it.
-
- So it appears that I was indeed wrong about my original assumptions on performance with heap allocation. Turns out, I...
-
04:47 AM Bug #3722: link-editor is over restrictive of R_AMD64_32 addends
- You suggestion of using "-fPIC" seems to work as you expected.
I'll therefore add a hack.mk in the meanwhile setti... -
- Rich Lowe wrote:
> You have 3 GCC's:
>
> - 1. 4.7.2 defaulting 64bit (evidenced in the first paste)
> - 2. 4.4.4... -
- The C kstat reads the wait stats from the wrong array, and thus reports the CPU times as wait times (all of which sho...
Also available in: Atom