Activity
From 2011-06-28 to 2011-07-27
2011-07-27
-
09:59 PM Bug #693: Opensource replacement of sunwlibm
- Would much prefer lib/libm/common as where the actual sources are (They are, after all, all versions of libm)
-
08:30 PM Bug #693: Opensource replacement of sunwlibm
- discussed with Gordon about structure:
lib/libm1 - libm.so.1
lib/libm2 - libm.so.2
lib/libmvec - libmvec.so.1
all... -
06:57 PM Bug #1282 (Resolved): nsmb_open has a race
- While comparing the nsmb driver with another driver, I noticed that it
prematurely drops a mutex protecting some glo... -
06:51 PM Feature #1281 (Resolved): Expose all TSS definitions
- uts/intel/sys/tss.h defines the structures for the Task-State Structure from the Intel Architecture. Currently the ts...
-
04:34 PM Bug #614: catman should execute nawk
- Taking bug away from Roland... he's not working on it.
-
- There are numerous third party modules that may use this (probably do!) I suggest just flagging a warning for now.
-
- I suspect you may have hit a known ZFS corruption bug that was fixed a while ago (883)? Please let us know if you se...
-
04:32 PM Bug #1174: fails to boot on Thinkpad T520
- As far as I can tell it dies before it can dump. From the looks of things, it dies on the write to the msr itself.
... -
- Interesting. I wonder if we have current enough ACPI tables. I have a Thinkpad W510, Core-i7 Q 820, and I don't see...
-
- Resolved in:
changeset: 13414:b42c1f0432b6
tag: tip
user: Garrett D'Amore <garrett@nexenta.co... -
- Resolved in:
changeset: 13414:b42c1f0432b6
tag: tip
user: Garrett D'Amore <garrett@nexenta.co... -
- Resolved in r13410 commit:16832af22909
-
- Resolved in r13411 commit:814f11fdc297
-
- resolved in r13412 commit:706f98f17323
-
- Resolved in r13413 commit:22409bb43a34
-
- Resolved in r13413 commit:22409bb43a34
-
- The err() family of functions exit with a given message and failure code, and never return. They should be marked as...
2011-07-25
2011-07-24
-
- ше was example for using -YP,<dir> flag
as you can see we can't use it because we have problems with 'cw' wrapper.
... -
- Using -YP /lib/amd64 looks wrong to me. That's asking explicitly to
link against libraries on the build system inst...
2011-07-23
-
- I can see LDLIBS32 and LDLIBS64 at Makefile.master, but I can't understand issue:
(file /usr/lib/64/values-xpg6.o ty... -
- cw(1) only knows how to parse flags we use in the build. If you add more out of the way flags, you normally have to ...
-
- /home/igor/illumos-libm4/illumos-gate/usr/src/tools/proto/root_i386-nd/opt/onbld/bin/i386/cw -_cc -xO3 -m64 -Ui386 -U...
2011-07-22
-
11:17 PM Bug #1242 (Resolved): libwrap has undefined symbols
- tcpd.h prototypes 'allow_severity' and 'deny_severity' but libwrap leaves them undefined.
-
- In answer to both points: fractional-sleep is available on most other platforms (FreeBSD seems to do it, OSX does do ...
-
- Why shouldn't it be locale sensitive? I think it should respect the locale. If you want your script to work properl...
-
- the ksh93-ized sleep tries to support fractions of a second, unfortunately it is locale sensitive (and probably incor...
-
- Emulex has provided a code drop, which updates to support their very latest 10GbE hardware. (Another one for FC is c...
-
- I had contemplated using this, but had decided originally not to do it, because I wanted to minimize any edits or mas...
-
- Resolved in:
changeset: 13408:90b8961d252c
user: Bayard Bell <buffer.g.overflow@gmail.com>
date: ... -
- Fixed in:
changeset: 13409:a4e1558c0599
tag: tip
user: Jason King <jason.brian.king+illumoshg... -
- Conditions in va_list.h and va_impl.h use __GNUC_MINOR. The macro in question is __GNUC_MINOR__, the version without...
2011-07-21
-
12:53 PM Bug #1102: Resource exhaustion in sftp client
- I've submitted an RTI for this bug.
So far, nobody has offered to integrate it.
-
- Both Linux and FreeBSD seem to have a driver. It'd be userful to have it too.
-
02:59 AM Bug #1232: setlocale switch to C is broken
- Corrected patch attached.
Test case also attached.
Ran full nightly, resulting be w/ patch applied does not hav... -
- We should be using __builtin_va_start for any GCC >= 3.3, based on when the change was made to their ginclude copy, s...
-
- intrd needs the Sun::Solaris::Kstat module, which is in runtime/perl-510/modules/sun-solaris
2011-07-20
-
- Ugly historical mess...
SYSV defines _numeric (which appears to be exposed from libc) which has some locale data. ... -
07:25 PM Bug #1232 (Resolved): setlocale switch to C is broken
- Result is https://www.illumos.org/issues/1203 since perl uses setlocale(LC_NUMERIC, "C"); to check the version number...
-
- It would be nice to eliminate the libc_hwcap* files on x86 and
instead use "symbol capabilities" as was done for spa... -
07:07 PM Feature #1230: need support for recent gcc stdarg builtins
- Rich Lowe wrote:
> If you're building bits of illumos with GCC4 the entire branch may interest you: http://github.co... -
- If you're building bits of illumos with GCC4 the entire branch may interest you: http://github.com/richlowe/illumos-g...
-
- Also it can happens here while building illumos grub with gcc 4*. I used same fix as Rich.
-
- I have a fix for this in my gcc update branch. I confess I'm not sure how you're hitting this directly, as I had tho...
-
-
02:48 PM Feature #1230 (Resolved): need support for recent gcc stdarg builtins
- When i tried to compile something, what used macros va_start(), i get error:...
-
01:47 PM Feature #1163: zpool destroy should prompt user
- we should go ahead and close this issue then..
Thanks
2011-07-19
-
08:51 PM Bug #1118: /dev-il illumos version will break illumos onu
- I (or Rich) will be addressing this by changing the illumos branch version.
-
- Andrzej Szeszo wrote:
> Rich, regarding incorporations, branch version numbers and keeping packages in sync. How wou... -
05:39 PM Bug #1116 (Resolved): beadm *_snap commands shouldn't exist
- changeset: 13407:cec0afe32134
tag: tip
user: Alexander Eremin <a.eremin@nexenta.com>
date: ... -
- With the correct attribution:
changeset: 13407:cec0afe32134
tag: tip
user: Alexander Eremin <... -
- changeset: 13405:fee421268ea6
tag: tip
user: eschrock
date: Tue Jul 19 09:54:12 2011 -0700... -
12:54 PM Bug #1219: find shouldn't use FTW_CHDIR
- Rich Lowe wrote:
> The comments about exec and symlinks were separate.
They aren't so separate as you meant them ... -
- Yuri Pankov wrote:
> I've added -iname as well as -print0, -useracl, -groupacl (returning true if the file/directory... -
- CUPS has most (probably all) of the SysV and BSD interfaces down. Only hitches I am aware of are SMB printing and Pre...
-
- Several people have pointed out that it would be utterly reasonable to EOF the traditional SVr4 print support, as lon...
-
- Currently the AES_[CG]CM modes are only supported within the kernel. Userland utilities (such as IKE, SSH, SSL/TLS) ...
2011-07-18
-
- Perl should be in the perl packages in sfw. Others could come from BSD, yet others have to be rewritten.
The las... -
05:52 PM Bug #1227 (Closed): Upstream man pages are missing
- A number of man pages are not delivered in pkg:/system/manual that are delivered by S11X.
Some of these look like ...
2011-07-17
-
- If built with -xstrconst or with gcc without -fwritable-strings lp will overwrite constant strings when:
1. Checki... -
- catman(1) uses the fact section name " catman " during its runs, but also sets the first byte to NUL, overwriting con...
-
- crypt(1) will modify constant data if built with -xstrconst or gcc without -fwritable-strings. It forces the key int...
-
- The comments about exec and symlinks were separate.
My concern with links is whether it would in any way affect wh... -
- -exec explicitly disables FTW_CHDIR...
2011-07-16
-
- You'd need, at the very least, to be careful about:
* The behaviour of -exec (which you'd change if you didn't mak... -
- Some numbers:...
-
- Resolved in r13402 commit:b8164899269e
-
- Resolved in r13403 commit:ed8234f32fb8
-
- Resolved in r13404 commit:4a1868d4ae91
2011-07-15
-
- It's usefulness is doubtful and leads to subtle problems with -acl (missing starting paths), as well as other extensi...
-
- I've added -iname as well as -print0, -useracl, -groupacl (returning true if the file/directory has non-trivial ACEs ...
-
- I can handle this issue and #676 if import of FreeBSD's xargs will be accepted (current code is hardly readable, for ...
-
- ...
-
- Gordon Ross wrote:
> Linux and Mac OS use --quiet for this. Should we too?
Why we should look at Linux.. What I ... -
- Igor Kozhukhov wrote:
> have been open bugs: #1210, #1211, #1212
> this bug have dependency to these bugs
-
- have been open bugs: 1210, 1211, 1212
this bug have dependency to these bugs
-
- While attempting to track down an issue (which turned out to just be an unannounced flag-day), I discovered the hard ...
2011-07-14
-
- When I wrote the code to fix the pages, Garrett was very much against flowing either the license or Sun copyright on ...
-
- I believe Rich Lowe has an automated way to do this.
(I already set him as a "watcher" of this:)
-
- Is it ok to just copy/paste (scripted, where it's possible) the CDDL license text (not touching copyright stuff, of c...
-
- The long lines in man pages are annoying. In particular, they make webrevs that include man page fixes not usable in...
-
- Linux and Mac OS use --quiet for this. Should we too?
-
- This means that we'll need one common locale package (locale-common?) which will be required by SUNWcs(?) and will co...
-
- Once "copy" (or some sort of include mechanism) is in place, an "extra credit" piece of follow-on work for this might...
-
- The XPG4 localedef input grammar has a "copy" statement. (See:
http://pubs.opengroup.org/onlinepubs/9699919799/base... -
- I copied 992 to create this. I guess that made it "resolved", which was unintentional.
-
- While reviewing the fix for issue 992
https://www.illumos.org/issues/992
I compared the compiled upper/lower mappin... -
- ...
-
- I would find this useful.
-
- The idea is to make chmod handle empty access mask as wildcard when removing user/group ACEs, e.g.:...
-
- provide $(ROOT)/usr/include by CPPFLAGS to usr/src/cmd/lms
-
- provide $(ROOT)/usr/include to 'INC=' through Makefile.master for usr/src/cmd/perl
-
- provide $(ROOT)/usr/include by CPPFLAGS to usr/src/lib/libsqlite
-
-
-
-
- It is supported, but is missing from manpage (FNM_IGNORECASE).
2011-07-13
-
- Last attempt at being stubborn :-)
Albert Lee wrote:
> I cite zero(7D) and null(7D) as supporting evidence this s... -
- I cite zero(7D) and null(7D) as supporting evidence this should be in section 7D.
intro(7) doesn't describe anything... -
- ~# beadm create nightly@snap
Created successfully
~# beadm list -s nightly
BE/Snapshot Space Polic...
2011-07-12
-
08:26 PM Bug #1197: Hang after resilver finished with mpt
- Just checked SMART info on the drives as well - nothing...
-
12:06 PM Bug #1202 (New): "zdb -DD" histogram block count should be decimal-K/M based, not binary
- When I look at the "zdb -DD" output for dedup stats of my pool, I see that large counts of blocks are rolled into 102...
-
- Putting prototypes in function bodies is certainly tasteless, but declaring them _static_ is bogus: if the compiler s...
-
05:16 AM Bug #1200: various modules declare static functions within function bodies
- I believe placing function prototypes in the body of functions is technically legal C, but strongly discouraged as ba...
2011-07-11
-
- It is, apparently, non-sensical to place a static function prototype within the body of a function, though Studio and...
-
- efe makes efe_restart() static in its prototype but not its definition.
(It's not clear to me why efe is so packed... -
- [Copied my note 8 from 1030]
It does seem that the "%x" format is more commonly defined as
some permutation of "%... -
- [Copied Note 7 by Gary Mills from 1030]
I would say that the bug is in strftime. It should never
produce a condensed... -
- Here are some more observations from
Gary Mills (sent to developers@illumos.org)
I kept wondering why I've never... -
- LANG=en_US.UTF-8 date '+%x'
has unexpected form in the first 9 days of the month:... -
- I have the obvious fix for this on my gcc branch.
-
- ...
-
- Also, iostat -en doesn't show anything alarming in particular, and none for the failed drives
-
- Just rebooted the box - couldn't get anything more out of it anyway. After the reboot, it shows a few more drives hav...
-
- Hi all
I just had a machine finish resilver after a drive (well, two actually) died. After resilver was finished, ... -
- changeset: 13397:edfb98e65b97
tag: tip
user: Garrett D'Amore <garrett@nexenta.com>
date: M...
2011-07-10
-
- It'd be super if man -k cooperated with man -s such that, for eg, 'man -s 3ucb -k sprintf' did the obvious thing, and...
2011-07-09
-
- Garrett D'Amore wrote:
> THIS IS A TERRIBLE IDEA.
>
> Zpool force avoids a bunch of checks. It should not be use... -
- Well, my understanding is that 4 is 'file formats', fd doesn't describe any format, it talks about special files in /...
2011-07-08
-
- I guess I'll better throw in the towel here. You're right, it'll break scripting.
roy -
- THIS IS A TERRIBLE IDEA.
Zpool force avoids a bunch of checks. It should not be used just because you want to use... -
- Well on second thought do we really need this as a feature.
Reason being there are a lot of scripts that that curren... -
- We will need the 'force' flag then so zpool/zfs commands would still be usable for scripting.
-
- I have the fix for this.. So assigned it to myself.
-
- I think the idea is that its describing files and thus in 4. We mostly use 7 itself for non-driver kernel modules it...
-
- fd manpage is describing special files in /dev and should live in section 7. While here, add links for stdin, stdout ...
-
- Albert Lee wrote:
> Yuri Pankov wrote:
> > fd(4) looks like a better candidate for stdin, stdout, stderr manpages.
... -
- Yuri Pankov wrote:
> On the related note, I have a strong feeling that fd manpage should live in section 7 (special ... -
- Yuri Pankov wrote:
> fd(4) looks like a better candidate for stdin, stdout, stderr manpages.
For _/dev/{stdin,std... -
- On the related note, I have a strong feeling that fd manpage should live in section 7 (special files) and could menti...
-
- fd(4) looks like a better candidate for stdin, stdout, stderr manpages.
-
- We have an _stdio(3C)_ man page, but none for _stdin/stdout/stderr(3C)_ - the global variables it actually describes....
-
- Same as #1190.
-
- As discussed with Rich, there's no point in building and installing those packages (it should also make it easier to ...
2011-07-07
-
- ...
-
- Those are the only manpages still living outside usr/src/man.
-
12:20 PM Bug #1118: /dev-il illumos version will break illumos onu
- Albert Lee wrote:
> 'depend' relationships are >=. A piece of software (SFEllvm, for example) with
> @depend fmri=s... -
- Andrzej Szeszo wrote:
> Rich Lowe wrote:
> > Andrzej Szeszo wrote:
> > > Not having branch version number would ma... -
- Rich Lowe wrote:
> Andrzej Szeszo wrote:
> > Not having branch version number would make life easier for people bui... -
- We should integrate the mharsch arcstat (the original arcstat.pl, updated for L2ARC amongst other things). It's in J...
-
- Joyent provide a 'vfsstat', built on statistics provided by #1183. We should integrate it along with #1183
-
- Joyent introduced a 'ziostat' built on the kstats provided by #1183, we should take it along with #1183
-
- We should integrate the changes made by Jerry Jelinek and Bill Pijewski of Joyent, and described here by the latter h...
2011-07-06
-
- It appears that GNU iconv has aliases for latinN mapping to ISO-8859-N, whereas we only have simple "8859", and "8859...
-
- Rich Lowe wrote:
> Andrzej Szeszo wrote:
> > Not having branch version number would make life easier for people bui... -
- Andrzej Szeszo wrote:
> Not having branch version number would make life easier for people building their own packag... -
- I was only experimenting with things. I am not sure personally what would be the best approach to versioning. I kind ...
-
- These 3 manpages refer to non-existing commands - compatibility/ucb doesn't deliver plot, tplot was removed in #364.
2011-07-05
-
- I can't speak as to what OI want to do, but it's very likely that incorporations _don't_ give you the behaviour you w...
-
08:52 AM Bug #1118: /dev-il illumos version will break illumos onu
- I think Andrzej's idea has merit and is worth considering, but post-oi151-stable.
For oi_151-stable, I don't mind ...
2011-07-03
-
- Resolved in r13395 commit:29f0883d3bc0
-
-
- Sigh. My fault. Thanks for fixing this.
Please send the webrev to developers@ for review, (which should be trivi... -
- Webrev can be found here: http://cr.illumos.org/view/nktd7byq/
-
- In usr/src/head/paths.h we currently have:...
-
- I was thinking about relying on timestamps in case of illumos packages when doing upgrades and keeping individual pac...
-
- Dropping the branch versions would make this harder, not easier. (we'd go from having one number to keep >= to havin...
-
- Dropping branch version numbers is another option worth considering going forward. Current pkg5 tools don't have any ...
2011-07-02
-
02:56 PM Bug #1176 (Resolved): NWAM handling of multiple DHCP-assigned DNS domains broken
- The dhcp server at work gives me three domain names: foo.bar.com, bar.com and baz.com. On OpenSolaris b134 that resul...
-
05:36 AM Bug #1166: useradd warns on names with more than 8 chars
- from the comments in <limits.h>...
2011-07-01
-
11:55 PM Bug #1175 (Feedback): ZFS rpool boot fails on oi148
- console output gives:...
-
- Pretty straight forward, take a thinkpad t520 with a quad core processor, boot off the oi 148 live CD. The kernel spi...
-
- Actually, these are just regular multibyte nops, not the "hinting nops" to which intel refer (which are, as stated, t...
-
- 8 chars is POSIX standard (POSIX_LOGIN_NAME_MAX in
http://www.opengroup.org/onlinepubs/007908799/xsh/limits.h.html) -
- an hg export of the changes, in case anyone wants to try this out.
-
- Here's an implementation of this proposed feature:
http://www.yalms.org/cr/libtermcap/
-
- While looking at issue 415 https://www.illumos.org/issues/415 it became apparent that issues like 415 could benefit ...
2011-06-30
-
- ...
-
03:43 PM Bug #832: need Intel 82579 Gigabit Ethernet PHY support in e1000g
- oracle recently fixed 6944824 Intel e1000g NIC support for Lewisville (82579LM and 82579LF)
Not sure if this will he... -
- As I can see we have '-_cc' for SunC and '-_gcc' for GCC to cw wrapper - for using different compilers.
they are fl... -
- useradd have warning with name more 8 chars
--------------
for example:
# useradd test567890
UX: useradd: test567... -
- Changing from:
if (!(is_z_a_power_of_x_le_y(2, 128, num)))
to:
if (!(ISP2(num) && (num >= 1) && (num <= 128)...
2011-06-29
-
10:20 PM Bug #1148 (New): Serious performance regression using CompactFlash as rpool ( cmdk, zfs? )
- Thanks for the benchmarks. I'm assigning this to "OI Illumos" for further investigation since the problem appears cl...
-
11:48 AM Bug #1148: Serious performance regression using CompactFlash as rpool ( cmdk, zfs? )
- The command used was:...
-
- Bonnie from OS 134 - with size same as on OI151 ( smaller than in previous test )
-
- Here are the bonnie results. It's same physical machine.
-
- These appear to be packages that were created empty, for Sun's somewhat hair-brained "make the bits appear in a patch...
-
- Heres an updated patch which I think may be ready to submit. It's also worth noting that APT requires cfmakeraw.
-
- Hi all
Destroying a beadm will prompt the admin:... -
- Roy Sigurd Karlsbakk wrote:
> Resilver started and finished to the drive, so yes, I think it's supposed to be in use...
2011-06-28
-
- You probably want to use _strrchr_ and _strcmp_ here. The gzcat also should be optional (man doesn't depend on gzip)....
-
- Thanks, that makes the patch a bit simpler.
-
- What I mean is that you don't need the the 'gzipit', just use `gzcat -f` instead of `cat` for all files.
-
- Sure.. Another option could be "fixing" zcat to accept -f option forcing it to allow non-compressed data to pass thro...
-
- I've spoken to Garrett and it seems that we could use my patch as a stop-gap solution until your changes are ready.
-
- Just to let you know, the bonnie test on 151 is running, already it took 10 hours and still not finished. Tomorrow I ...
Also available in: Atom