Activity
From 2012-05-30 to 2012-06-28
2012-06-28
-
10:11 PM Bug #2946 (New): USB sometimes stops working completely
- Sometimes, when I unplug all USB devices, the USB subsystem completely stops working. Nothing happens at all when I ...
-
10:02 PM Bug #2942: CTF tools need to handle files which legitimately lack data
- The fix I have for this employs a simple heuristic "any file which contains executable text must also contain actual ...
-
- There are, within our source trees, files which compile to nothing (who can imagine why).
The CTF tools need to ha... -
09:33 PM Bug #2945 (Resolved): mdb needs to export _mdb_ks_pageoffset
- Modules are supposed to use mdb_param.h instead of sys/param.h. However, anyone who does so and then references a ma...
-
08:06 PM Bug #2944: cscope-fast/invlib.[ch] use reserved word as identifier
- UPD. C99 is not enabled.
But @bool@ is bad name for a variable. -
- @usr/src/tools/cscope-fast/invlib.h@ uses @bool@ as a variable name.
@bool@ is a typename in C99 and cscope-fast i... -
- ...
-
-
12:42 PM Bug #2911: blkdev-related panic on DEBUG kernels
- Got it. kmem_cache_alloc only expects the constructor to always succeed when KM_NOSLEEP is not set. Let me fix blkdev.
-
12:02 PM Bug #2911: blkdev-related panic on DEBUG kernels
- No. KM_NOSLEEP allocations won't be in that code path. Or rather, the rc will be -1, indicating a KM_NOSLEEP failed...
-
- The first call is still made with KM_NOSLEEP, so the ctor would still fail on the first call, and cause the assert in...
-
- No, you misunderstand. KM_NOSLEEP should still be fine. And it is. The problem is that we used KM_PUSHPAGE instead...
-
- The ctor would fail from the first allocation that is using KM_NOSLEEP:...
-
- I think the problem is that bd_xfer_ctor() needs to consider KM_PUSHPAGE as well as KM_SLEEP. (KM_PUSHPAGE has simil...
-
06:46 AM Bug #2885: Rapid incremental zfs sends can leave hidden clones
- Could you attach your reproducer script, please?
Maybe you should ask on the developer mailing list for some advice. -
- Certain foolhardy compilers have begun to assert that their C++ compiler complies with C99. It can't, that's a contr...
2012-06-27
-
- http://188.120.236.176/webrev/2940/
-
- ...
-
- ...
-
- A simulation was included in the RTI:...
-
05:32 PM Bug #1961: investigate stopping uniquifying CTF information
- Rich, could you please add some lines here about possible increazing of miniroot/bootarchive size (from your tests).
-
04:21 PM Bug #2939 (New): dladm show-linkprop -p cpus fails on ncpus > 32
- I can't tell if this is merely cosmetic, but:...
-
12:46 AM Support #2935: ksocket_sendmsg EINTR error
- This looks like a question and not bug report and should probably be asked on networking@ and/or developer@ mailing l...
2012-06-26
-
- Igor Kozhukhov wrote:
> I have a comment:
> I thought about it and had idea: configure env with additional userla... -
12:26 PM Bug #2936 (New): FIONBIO cause ksocket_recvmsg speed down?
- i have one kernel module run in two units. one unit act client and one act server. client send the message to server ...
-
- ksocket_sendmsg return EINTR, if it always return EINTR, the send speed will be very low.
-
- i have one program to act client and server, when the client send msg to server, it's normally at start, but after ma...
-
-
03:25 AM Bug #2724 (Resolved): snoop contains few problems with accessing uninitialized data
- ...
2012-06-25
-
- At present, for several reasons, we have gagged some high-value (and some low-value) compiler warnings globally. Thi...
-
04:21 PM Feature #2932 (Resolved): support crash dumps to raidz, etc. pools
- Most systems are configured with a dedicated system pool: a pool which contains the installed OS, swap device, dump d...
-
03:13 PM Bug #2930: beadm should sort by creation date
- I think sort by name is correct.
but we can implement additional flag for additional sort types.
-
03:06 PM Bug #2930 (Closed): beadm should sort by creation date
- This is something I've found annoying for some time....
2012-06-24
-
- The attributes(5) man page refers to Sun instead of Ilumos. Example:...
-
- I have a comment:
I thought about it and had idea: configure env with additional userland soft for illumos-gate and ... -
- Currently, illumos is not self-contained; its build encoded runtime dependencies based on a number of objects found o...
-
- lex's undocumented -Y option would be useful if it did anything. Yacc honors a very similar option to allow finding ...
-
- We've already got "LSO":http://en.wikipedia.org/wiki/Large_segment_offload support, but no "LRO":http://en.wikipedia....
-
- The description probably should be added to the cmn_err man page, but I'd leave it to a native speaker.
-
- Recently added dev_err() should have a manpage describing it.
-
- Many manpages are asking for tbl preprocessing (via @'\" t@), but don't actually contain any tables.
2012-06-23
-
- None of the manpages in the gate need eqn (except for the eqn(1), of course), but almost all of them ask for eqn prep...
-
06:22 PM Feature #2919 (In Progress): add hal-set-property --direct
- HAL 0.5.13 added --direct to hal-set-property which is usefull for vmware guest
-
- So, how should we handle this? Is it possible that a KM_SLEEP allocation would cause some pages to be swapped out, th...
2012-06-22
-
06:38 PM Bug #2901 (Resolved): ZFS receive fails for exabyte sparse files
-
- Hmm, I don't think you are right.
bd_xfer_alloc() is called from bd_strategy():... -
- Reading the code, it appears that it is ill equipped to deal with constructor failure here. That's a problem. We ca...
-
- When trying to do an illumos build on a DEBUG kernel, I'm getting this a few hours into the build: ...
-
- Resolved in r13739 commit:f6a4c9dc9b80
-
- I am seeing this nearly every system shutdown.
-
- Resolved in r13738 commit:e10d315d1821
-
- Resolved in r13738 commit:e10d315d1821
-
- Resolved in r13737 commit:1189735fe286
-
02:23 PM Bug #2917 (Resolved): DTrace in a zone should have limited provider access
- Users in zones cannot have the *dtrace_kernel* privilege, and thus cannot presently utilise various stable probe prov...
-
- webrev: http://cr.illumos.org/~webrev/jclulow/2916/
-
- Users in zones cannot have the *dtrace_kernel* privilege, and thus cannot presently gain access to the contents of DT...
-
- webrev: http://cr.illumos.org/~webrev/jclulow/2915/
-
- Users in zones cannot have the *dtrace_kernel* privilege, and thus cannot presently gain access to potentially intere...
2012-06-21
-
11:25 PM Bug #2910 (Resolved): svc.startd core dump
- Tracked down that this was due to svc.startd dumping core and restarting.
> $C
f5af9da8 libc_hwcap1.so.1`_lwp_kil... -
07:25 PM Bug #2909: zfs: assert at ../../common/fs/zfs/bptree.c: line 195
- for OpenIndiana 151a3 (before async destroy changes) assert is:
debugging crash dump vmcore.5 (64-bit) from oi... -
- for OpenIndiana 151a4 assert is differ:
::status
debugging crash dump vmcore.2 (64-bit) from oi151
operating s... -
- headcommit: 13716:3db0c1979163
image uuid: 407e5701-35ce-661a-f1ee-d3820a8b5da5
panic message: assertion failed: ... -
- Hi.
I'm getting this kind if result when trying to run ::findleaks.
[1]> ::findleaks
*** Debugger Fault (CPU 1... -
- dsl_pool->dp_config_rwlock is krwlock_t. Writer has priority over readers and is barrier for readers.
So nested R... -
- I've checked other OSes, and only FreeBSD doesn't have bel defined for the 'dumb' terminal type, Solaris 11 and Debia...
-
- While looking at this issue, I noticed the following (not directly related to the issue itself) - __ct_rune_t is type...
-
- In a couple of places, hxge does:
ASSERT(0 <= rhp->index < HXGE_MAX_TDCS);
That's always true, and not useful... -
- after pkg install gcc-3 or pkg install //openindiana.org/system/kernel, the dtrace -s vicm.d show the follow error me...
2012-06-20
-
- Why isn't this fix in illumos-gate? (Or at the very least OI's child of it?)
Need review cycles? More tests? Mor... -
- Resolved in r13735 commit:1c48a05ba37d
-
07:43 PM Bug #2901: ZFS receive fails for exabyte sparse files
- The size of the file is < 2^63, so it is not actually too large for a uint64_t.
from zstreamdump:
WRITE object ... -
- Our *zettabyte* file system has a problem with *exabyte* sparse files......
-
-
-
- After integration of #2831 only the last leak is resolved here.
-
06:49 PM Bug #2869: duplicate packets with vnics over aggrs
- For reference, here is a proposed fix: http://cr.illumos.org/~webrev/postwait/2869/
This fix is not optimal as the... -
05:20 PM Bug #2903 (Resolved): zfs destroy -d does not work
- It fails if the snapshot has clones (it is supposed to mark the snapshot for deferred destroy). A test for this case ...
-
- pushed:
commit 462453d2d0c563559a4caf186db76954e563bd1a
Author: Matthew Ahrens <matt@delphix.com>
Date: Tue Oc... -
01:33 PM Bug #2702: ZFS sharenfs strangeness with anon=
- Yuri Pankov wrote:
> Could you please do the same for the failing anon=-1?... -
- Could you please do the same for the failing anon=-1?
-
- Yuri Pankov wrote:
> Try `snoop rpc nfs` to check what happens.
in nautilus I did a refresh to /net/smicro/export... -
- Try `snoop rpc nfs` to check what happens.
-
- Yuri Pankov wrote:
> Using which uid are you trying to access the mount in all cases? Which security mode is used?
... -
- Using which uid are you trying to access the mount in all cases? Which security mode is used?
-
- During creation of the target with EUI iSCSI name, itadm folds the name uppercase, while MS iSCSI initiator sends the...
-
- Rich，
thanks for your reply.
on dtrace fail system:
root@vs1:~# pkg info system/kernel network/iscsi/target
...
2012-06-19
-
- The "zfs snapshot" command can create a single snapshot, or with the -r flag, a snapshot for every filesystem under a...
-
- Resolved in r13734 commit:41fe4dd8e23d
-
- Generally, that's because of CTF mismatches between closed-bins and the rest of the system. I can't imagine how suc...
-
- when i install openindiana, dtrace -s vicm.d can work normally. but after i install comstar. it appear 2 boot item. ...
-
- Here is another deadlock after introduced getzfsvfs_ds()
This deadlock is due to race between ds->ds_rwlock and vf... -
11:20 AM Bug #2897 (Closed): "zpool split" documentation missing from manpage
- Found this option in some Oracle documentation and wanted to check out the zpool manpage on it in OI. Unfortunately i...
2012-06-18
-
- By default, ls(1) reads the entire contents of a directory into memory, sorts the entires, and calculates how to opti...
-
- usr/src/tools/scripts/build_osnet.sh is the script which used to be delivered to solaris source customers to allow th...
-
- Attached is a version of the driver after addressing the recommendations from Garrett D'Amore. This one might be the...
-
02:23 PM Bug #2892 (New): flows use datalink IDs in persistent configuration
- The /etc/dladm/flowadm.conf configuration for flows stores the datalink over which flows are configured using a datal...
-
-
11:01 AM Bug #1487: some binaries fail to work in a chroot environment.
- Looks like you were right. /devices/psuedo was not mounted so /dev/zero was just a dangling symlink.
-
- svc.startd was stripped by https://www.illumos.org/issues/2831 from umem debugging for non-debug (release) build but ...
2012-06-17
-
- Yuri Pankov wrote:
> Ok. Now I see why you want this change, guess it's related to #2849 so that we could test <code... -
- Some 64-bit operating systems and virtualization software running on Intel CPU hardware are vulnerable to a local pri...
-
- pargs core_zpool
core 'core_zpool' of 2624: zpool create -f data raidz1 c2t5000C50016AFEC4Fd0 c2t5000C50016B276D3d0... -
- boot environments should be created under 'beadm create' not 'beadm activate'.
I don't understand how the failure ...
2012-06-16
-
- Ok. Now I see why you want this change, guess it's related to #2849 so that we could test ...
-
- Yuri Pankov wrote:
> Better question is why this change is needed at all.
???
Cause ru_RU *does* use 24-hours ... -
- Better question is why this change is needed at all.
-
06:31 PM Bug #2127: 'zpool create' core dumps
- https://www.dropbox.com/s/kg89zkapjzeu74o/core_zpool.bz2
-
- we have use case:
1. beadm create <newBE>
2. beadm destroy <newBE>
new zbe-x for zones have been created in (1) ...
2012-06-15
-
- As the locale data for this was taken directly from CLDR, at the Unicode organization, we should work to fix that ups...
-
- Note that I pruned a lot of this out of SVR4 packaging at one time. Not sure if I *pushed* those changes though.
-
- While looking further at #1961, I've noticed that we're still carrying all the Solaris patch release machinery in the...
-
- Resolved in r13726 commit:056b82d21d71
-
04:05 PM Bug #2885 (New): Rapid incremental zfs sends can leave hidden clones
- A couple of my guys discovered that rapid succession "zfs send/recv" can result in some hidden clones being left arou...
-
07:46 AM Bug #2883 (Resolved): changing "canmount" property to "on" should not always remount dataset
- If a dataset has the "canmount" property set to "noauto" it may be mounted.
Changing this property to "on" on a moun...
2012-06-14
-
08:46 PM Bug #2778: zpool replace fails - 2TB advanced format drives
- Just a final follow up on this issue. Apparently there are at least 3 firmware revisions out in the world. CC3C and C...
-
08:23 PM Feature #2881: usbftdi needs support for the SheevaPlug JTAG adapter
- Webrev can be found here: http://grumpf.hope-2000.org/illumos-2881-webrev/
I took the liberty to add a few ASSERTs... -
- Usbftdi currently supports single-port USB-to-serial adapters based on the FTDI232R chip. Some of these chips support...
-
- We should implement libzfs_core, under the principles proposed here:
http://blog.delphix.com/matt/2012/01/17/the-fut... -
- Fixed in:
garrett@neon{291}> hg head
changeset: 13725:9a3ca91fb74e
tag: tip
user: Garrett D'... -
- Resolved in r13724 commit:7740792727e0
-
03:14 AM Bug #2873 (Resolved): sysretq doesn't properly handle non-canonical addresses
- The syscall and sysret instruction expect to be given a canonical x86 address when called. On Intel processors it is ...
2012-06-13
-
- Closed . dup.
-
- This bug should be closed as a duplicate of 2605, but I can't see how to do that through this interface.
-
- Dup of 2067.
-
- Rejecting this as more investigation needed.
-
- Resolved in r13723 commit:3297c26a553a
-
- The above diff is incorrect: it checks if the filesystem's readonly property is 'on'. You need to check if the pool ...
-
- When the system is booting zfs_mountroot() is called to mount the root file system, but if the kernel being booted do...
-
- The ZFS test suite uses an undocumented __ZFS_POOL_RESTRICT environment variable to allow it to test things like 'zpo...
-
- The issue appears to fall along the lines of hardware support.
If the underlying ethernet devices support hardware... -
- ...
2012-06-12
-
- On a current system with two e1000g NICs in an LACP aggregate, on which vnics are placed, we see duplicate packets. (...
2012-06-11
-
08:35 PM Feature #2761: ZFS checksum/compression algorithm number indirection
- Would it make sense to add a better-compressing opensourced algorithm, like bzip2, p7zip or xz? For example, WORM dat...
-
- Resolved in r13721 commit:5b51a16a186f
-
- Certain well-formed commands can cause the iscsit module to kernel panic.
We have a well-tested solution. Because... -
07:22 AM Bug #2740: zfs_rmnode() fails to clean up znodes
- [spam]
-
- [spam]
-
- It should be mentioned that while we can do this part, it doesn't actually help for the specific case in which it is ...
-
- When elfsign is built for the host rather than target machine, in usr/src/tools, it uses -I.../uts/common/sys but doe...
2012-06-10
-
- Fixed in:
garrett@velocity{25}> hg head
changeset: 13719:da056fdd2d14
tag: tip
user: Bryan Can...
2012-06-08
-
- From Joyent OS-987:
the w command is hanging because it's trying to get the pr_p_lock
on the prnode_t associated ... -
- Tried to add a pair of cache devices to a zpool, got back:
# zpool add nosuch3_a0_1 cache c2t1d0 c2t2d0
Assertion f... -
08:06 PM Bug #2852 (New): mdb -o nostop -p <pid> hangs forked children forever
- shell 1:
$ echo $$
196398
shell 2:
$ mdb -o nostop -p 196398
>
shell 1:
$ sleep 1
<hangs forever>
shell 2:
... -
- By reading some SMF scripts I discovered that "/sbin/zonename" takes some parameters, at least "-t" to return the ip-...
2012-06-07
-
- I believe that the deadlock mentioned in update #2 is caused by the zfs_ioc_recv() thread recursively holding the dat...
-
- ...
-
- Currently @uptime@ printf 12-hour time no matter what locale is set:...
-
- Because of some implementation quirks, we're going to be doing this in the compilers for 4.4.4, at the least.
2012-06-06
-
- Resolved in r13716 commit:3db0c1979163
-
- Resolved in:
changeset: 13715:351036203e4b
tag: tip
user: Garrett D'Amore <garrett@damore.org>
... -
- Reproducer test for case 1:
mkfile 100m diskfile
zpool create testpool $PWD/diskfile
zfs create testpool/... -
- remove print/lp* from gate and use CUPS from userland
we have dependencies on print/lp ipp to apache-1.3
apache-1... -
- Change backed out due to missing packaging:
changeset: 13714:e2f6dabb84ef
tag: tip
user: Garret... -
- stmfadm, itadm and iscsiadm (probably there are other places which need updating) should support T11 Network Address ...
2012-06-05
-
- Resolved in r13713 commit:a5b409d8d6a8
-
- Fixed in:
changeset: 13712:236e80401f3d
tag: tip
user: Garrett D'Amore <garrett@damore.org>
da... -
- Resolved in:
changeset: 13711:0765d0ec4e2a
tag: tip
user: Vitaliy Gusev <gusev.vitaliy@nexen... -
- Resolved in:
changeset: 13710:3cee4ba952f0
tag: tip
user: Milan Jurik <milan.jurik@xylab.cz>
... -
- Resolved in:
changeset: 13710:3cee4ba952f0
tag: tip
user: Milan Jurik <milan.jurik@xylab.cz>
... -
-
09:43 AM Feature #212: Atheros AR8132 / L1c Gigabit Ethernet Adapter
- I can confirm both version of the driver (atge-18.5 and atge-01.6) work on 151a and 151a4 for me (Acer Aspire One 722...
2012-06-04
-
- Apparently libsocket is already there, so this applies to libnsl only.
-
- Like libw, libpthread, and others before it, libsocket and libnsl should move into libc and enter retirement as filte...
-
- OS-1186 svc.startd allocates excessive memory for instance names
OS-1187 svc.startd should reduce footprint by se...
2012-06-03
-
- In some situations (see bug #2807) ds_rwlock on READ can not be obtained immediately and caller goes to uninterruptib...
-
08:48 PM Feature #212 (Feedback): Atheros AR8132 / L1c Gigabit Ethernet Adapter
- Work fine for me with pciex1969,2062 Atheros AR8152 v2.0 Fast Ethernet on oi151a4 even after coldboot.
Can't wait un... -
- I've attached a more polished version of the driver, in case people want to do some more testing. This might be the ...
-
- Here's some information from a couple of other Unix live CDs. Freebsd 9.0 only detects two serial ports on the Tyan ...
-
- The last leak is in libscf_extract_runlevel(), val returned from safe_scf_value_create() - libscf_object_create() is ...
-
- webrev: http://grumpf.hope-2000.org/illumos-2812-webrev/
-
- The generic topology support for AMD CPUs isn't enabled for family 0x10 model 10. The kernel support is there and wor...
-
- ...
-
-
01:33 AM Bug #2811: missing implementation: zfs send -r
- Verified in xref: /illumos-gate/usr/src/cmd/zfs/zfs_main.c
It appears that the usage flags include:
252 case...
2012-06-02
-
- Resolved in r13704 commit:1f68af63b8d9
-
-
- The original libcmd which implemented defread() and defopen() was removed in build 51 of opensolaris.
But pmconfig... -
03:13 PM Bug #2811: missing implementation: zfs send -r
- Relevant discussion from #openindiana IRC channel:
2012-06-02 08:50 < lattera> http://pastebin.com/9MPt5uk1 <- wha... -
- The documentation shows that zfs send should have the -r flag implemented. The usage output for zfs send shows the sa...
2012-06-01
-
- For prev message (correction):
zfs list {pool}/{dataset}
zfs destroy {pool}/{dataset} -
- So "1" case can occur if just do:
zfs list {pool}/{snapshot}
zfs destroy {pool}/{dataset}
stack of thi...
2012-05-31
-
- Both init.d scripts are obsolete from Solaris 10 and with limited use. SMF must be used directly.
-
-
- Basics implemented. More precise detection of return values and handling of temporary tarfiles during update need som...
-
- 2. Second problem is:...
-
- Also another stack that shows real deadlock:...
-
- It can be:
1.
Because ds_rwlock is taken on READ in recursive (twice or more).
If another thread N2 tri... -
- System can hang due to obtaining dsl_dataset_t->ds_rwlock.
2012-05-30
-
-
- changeset: 13705:70355eedf57a
user: Milan Jurik <milan.jurik@xylab.cz>
date: Sun May 27 16:49:00 20... -
- From investigation I think it is the same issue as SmartOS 1188. I am not sure about the last leak yet.
-
- The upcoming push of SmartOS will be syncing out changes to startd and configd that clean up memory leaks and excessi...
-
- Resolved in 13708:6d0517b1a52b.
-
- For some distros, requiring the use (or installation!) of perl is onerous, but we still want to use certain system ut...
-
- Here's the proposed trivial fix for it:...
-
- ZFS tools treat the GUID for an object as a numeric and try pretty printing, which leads to garbage.
# zfs get gui... -
- Note that its not just filesystem/local, but also the devices, which need to be explicitly represented in the depende...
-
- After discussion with Dan McDonald, I have developed the following solution:...
-
- The ipsecalgs service tries to start very asynchronously, which is ambitious. It doesn't operate properly if /usr/ ha...
-
-
- Resolved in 13707:3f41b24ce657.
Also available in: Atom