Chris Nehren | 22 May 2013 23:57
Favicon

Possible bug (either in code or my understanding) about bart(1M)

I've been setting up bart auditing for a client. It works mostly well,
except I've reached a point of confusion between the docs and the code.
Specifically, consider this short bart_rules(4) file:

/etc
CHECK ALL
IGNORE dirmtime

/etc motd
IGNORE all

According to my understanding of bart_rules(4), /etc/motd should be
ignored. The example at the end of bart_rules(4) illustrates things
reasonably well. Yet if I create a control manifest, update /etc/motd
(say, e.g., via chef), create a test manifest, and compare, I see that
/etc/motd appears in the `bart compare` output. 

This seems to be contradictory to what the documentation suggests, or
I'm misunderstanding the docs.

I dug into the code for bart(1M) for this, and cases like these are
handled by rules.c:77-79 in exclude_fname(). This returns EXCLUDE_PRUNE
to create.c:456 in eval_file(), which sets ftwx->quit = FTW_PRUNE.

It looks like eval_file() should check whether ftype == 'F' when it
receives EXCLUDE_PRUNE from exclude_fname() and then treat this as
EXCLUDE_SKIP--or, better, that exclude_fname() should do this work
itself.

Thoughts? Comments? Tomatoes? Am I missing something? I want to believe
(Continue reading)

Geoff Nordli | 22 May 2013 20:11

Supermicro X9 series Mobo with SAS and 10GB ethernet

Anyone running one of these:

http://www.supermicro.com/products/motherboard/xeon/c600/x9drh-7tf.cfm

 From seeing the discussion a few months back Saso was about to buy one.

Any links to the instructions on reflashing the LSI 2208 to IT mode.

thanks,

Geoff

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Gabriele Bulfon | 22 May 2013 15:27
Favicon

sse caps

Hi,

I'm converting a linux/bsd library with a function checking for cpu sse caps.
What is the correct way to do this on illumos?
Here is the original piece of code checking for these caps:

static void check_os_katmai_support( void )
{
// printf ("omg\n");
#if defined(VISUAL_ARCH_X86)
#if defined(VISUAL_OS_FREEBSD)
int has_sse=0, ret;
visual_size_t len=sizeof(has_sse);

ret = sysctlbyname("hw.instruction_sse", &has_sse, &len, NULL, 0);
if (ret || !has_sse)
__lv_cpu_caps.hasSSE=0;

#elif defined(VISUAL_OS_NETBSD) || defined(VISUAL_OS_OPENBSD)
int has_sse, has_sse2, ret, mib[2];
visual_size_t varlen;

mib[0] = CTL_MACHDEP;
mib[1] = CPU_SSE;
varlen = sizeof(has_sse);

ret = sysctl(mib, 2, &has_sse, &varlen, NULL, 0);
if (ret < 0 || !has_sse) {
__lv_cpu_caps.hasSSE=0;
} else {
__lv_cpu_caps.hasSSE=1;
}

mib[1] = CPU_SSE2;
varlen = sizeof(has_sse2);
ret = sysctl(mib, 2, &has_sse2, &varlen, NULL, 0);
if (ret < 0 || !has_sse2) {
__lv_cpu_caps.hasSSE2=0;
} else {
__lv_cpu_caps.hasSSE2=1;
}
__lv_cpu_caps.hasSSE = 0; /* FIXME ?!?!? */

#elif defined(VISUAL_OS_WIN32)
LPTOP_LEVEL_EXCEPTION_FILTER exc_fil;
if ( __lv_cpu_caps.hasSSE ) {
exc_fil = SetUnhandledExceptionFilter(win32_sig_handler_sse);
__asm __volatile ("xorps %xmm0, %xmm0");
SetUnhandledExceptionFilter(exc_fil);
}
#elif defined(VISUAL_OS_LINUX)
// printf ("omg1\n");
// printf ("omg2\n");
struct sigaction saved_sigill;
struct sigaction saved_sigfpe;

/* Save the original signal handlers.
*/
sigaction( SIGILL, NULL, &saved_sigill );
sigaction( SIGFPE, NULL, &saved_sigfpe );

signal( SIGILL, (void (*)(int))sigill_handler_sse );
signal( SIGFPE, (void (*)(int))sigfpe_handler_sse );

/* Emulate test for OSFXSR in CR4. The OS will set this bit if it
* supports the extended FPU save and restore required for SSE. If
* we execute an SSE instruction on a PIII and get a SIGILL, the OS
* doesn't support Streaming SIMD Exceptions, even if the processor
* does.
*/
if ( __lv_cpu_caps.hasSSE ) {
__asm __volatile ("xorps %xmm1, %xmm0");
}

/* Emulate test for OSXMMEXCPT in CR4. The OS will set this bit if
* it supports unmasked SIMD FPU exceptions. If we unmask the
* exceptions, do a SIMD divide-by-zero and get a SIGILL, the OS
* doesn't support unmasked SIMD FPU exceptions. If we get a SIGFPE
* as expected, we're okay but we need to clean up after it.
*
* Are we being too stringent in our requirement that the OS support
* unmasked exceptions? Certain RedHat 2.2 kernels enable SSE by
* setting CR4.OSFXSR but don't support unmasked exceptions. Win98
* doesn't even support them. We at least know the user-space SSE
* support is good in kernels that do support unmasked exceptions,
* and therefore to be safe I'm going to leave this test in here.
*/
if ( __lv_cpu_caps.hasSSE ) {
// test_os_katmai_exception_support();
}

/* Restore the original signal handlers.
*/
sigaction( SIGILL, &saved_sigill, NULL );
sigaction( SIGFPE, &saved_sigfpe, NULL );

#else
// printf ("hier dan3\n");
/* We can't use POSIX signal handling to test the availability of
* SSE, so we disable it by default.
*/
__lv_cpu_caps.hasSSE=0;
#endif /* __linux__ */
// printf ("hier dan\n");
#endif
// printf ("hier dan ha\n");
}



illumos-discuss | Archives | Modify Your Subscription
Peter Tribble | 20 May 2013 23:36
Picon
Gravatar

Tribblix Milestone 6

I've been plugging away at Tribblix, and just cut another prerelease

http://www.tribblix.org/download.html

All very much experimental, of course. As well as a bunch
of application updates, the key feature of this release is
sparse-root zones, very much like Solaris 10. (And whole
root as well.) Some of the innards discussed here:

http://ptribble.blogspot.co.uk/2013/05/sparse-root-zones-in-tribblix.html
illumos-discuss | Archives | Modify Your Subscription
Paul B. Henson | 15 May 2013 23:11
Picon
Favicon

Re: [developer] illumos dev environment

On 5/11/2013 12:28 AM, Marc Lobelle wrote:

> Be carefull, a desktop and even a notebook environment are of utmost
> importance if illumos wants to survive in the long term. People using
> LINUX or Windows on their everyday machine (often the laptop) will
> prefer having the same environment on their servers.

I wouldn't necessarily agree. I actually run linux on my own 
desktop/notebook, but still want to/plan to deploy an illumos based OS 
where it provides the best feature set, typically for storage, and 
depending on your needs for virtualization. And with 
vmware/virtualbox/kvm, it's fairly trivial to bring up a virtual illumos 
to work with whether you use windows or linux on your client hardware.

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

John McEntee | 15 May 2013 16:41
Favicon

Failure to delete empty dirrectories via CIFS

I have just come across an issue using the inbuilt CIFS. Occasionally when a user deletes a directory the
server logs the error

messages:May 15 14:43:26 ringwood smbsrv: [ID 560541 kern.warning] WARNING: File DELETE could not be
removed, rc=17

And the directory "reappears when they refresh their view", but no error message to the user.

This has already be reported as bug ID 1663 https://www.illumos.org/issues/1663

Does anyone have any further information. It currently seems to only effect 1 user out of 140. Or he may be the
only one who has noticed. Will this be looked into or should I just plan to switch away from openindiana.

Thanks

John

_______________________________________________________________________

The contents of this e-mail and any attachment(s) are strictly confidential and are solely for the
person(s) at the e-mail address(es) above. If you are not an addressee, you may not disclose, distribute,
copy or use this e-mail, and we request that you send an e-mail to admin <at> stirling-dynamics.com and delete
this e-mail.  Stirling Dynamics Ltd. accepts no legal liability for the contents of this e-mail including
any errors, interception or interference, as internet communications are not secure.  Any views or
opinions presented are solely those of the author and do not necessarily represent those of Stirling
Dynamics Ltd. Registered In England No. 2092114 Registered Office: 26 Regent Street, Clifton, Bristol.
BS8 4HG
VAT no. GB 464 6551 29
_______________________________________________________________________

This e-mail has been scanned for all viruses MessageLabs.

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Jay Kay | 9 May 2013 16:09
Picon
Favicon

AD domain-joined server not listed in Windows' network neighbourhood

Hi,

I am using OmniOS-bloody but figured this might be a general illumos thing.
I have a server joined to an AD domain (2012), and it is working fine so far.
Dynamic DNS works, a computer object was created in AD and I can
list and access cifs shares by using the "\\<hostname>\<sharename>" notation
from Windows.
However the server or any of its shares are not listed in the network neighborhood in windows.
I am unsure if this is is a Windows or illumos configuration issue.
Or if it is actually expected behaviour.

Other domain members (windows servers) are listed in the network neighborhood.
All hosts are on the same subnet, WINS is not configured.
Unfortunately there is not much information around on this topic.
WRT http://wiki.illumos.org/display/illumos/CIFS+Service+Troubleshooting
I have a masterbrowser (Windows DC) inplace and it is in the same subnet as the
OmniOS server.

Any help is appreciated.

Jay

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Paul B. Henson | 8 May 2013 04:37
Picon
Favicon

atime=off on rpool?

I was thinking of disabling atime on my rpool to cut down on SSD writes,
are there any parts of the OS that will break without atime enabled?
Seems like a FAQ but I couldn't find anything definitive.

Linux has a mode called relatime which is defined as:

"relative atime only updates the atime if the previous atime is older
than the mtime or ctime. Like noatime, but useful for applications like
mutt that need to know when a file has been read since it was last
modified."

Would there be any interest in adding something similar to zfs like
"atime=relative"?

Thanks...

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Pedro Giffuni | 5 May 2013 18:29
Picon
Favicon

clang static analyzer scans

Hello;

Just thought I'd mention that FreeBSD runs regularly the clang static 
analyzer and the results are public:

http://scan.freebsd.your.org/freebsd-head/

Interesting subdirs for illumos developers usually start with cddl.*

For DTrace:
http://scan.freebsd.your.org/freebsd-head/cddl.lib.libctf/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.ctfconvert/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.ctfdump/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.ctfmerge/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.sbin.dtrace/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.sbin.lockstat/

For ZFS:
http://scan.freebsd.your.org/freebsd-head/cddl.lib.libavl/
http://scan.freebsd.your.org/freebsd-head/cddl.lib.libnvpair/
http://scan.freebsd.your.org/freebsd-head/cddl.lib.libuutil/
http://scan.freebsd.your.org/freebsd-head/cddl.lib.libzfs/
http://scan.freebsd.your.org/freebsd-head/cddl.sbin.zpool/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.sgsmsg/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.zinject/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.bin.ztest/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.sbin.zdb/
http://scan.freebsd.your.org/freebsd-head/cddl.usr.sbin.zhack/

It will probably report lots of false positives though ;).

Regards,

Pedro.

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Sašo Kiselkov | 30 Apr 2013 12:16
Picon

Trouble getting Supermicro board to see all its PCI-e slots

I've got a bunch of new Supermicro X9DRW-7TPF+ boards, with on-board
10-gigy sweetness by Intel. The motherboard is loaded with two Intel
E5-2630 CPUs with 128GB of RAM. The boards come in the proprietary WIO
form factor from Supermicro, so to get to the PCI-e slots, I'm using the
appropriate riser cards, to get to them. The risers are:

(looking from the front of the server)
Left-hand side:	 RSC-R2UW-4E8
                 four x8 PCI-e slots (two to CPU1, two to CPU2)
Right-hand side: RSC-R2UW-2E8R
                 two x8 PCI-e slots (two to CPU2)

However, and here's the kicker, I can't get OmniOS bloody to see
anything plugged into any slot other than those that lead to CPU1 (I'm
trying a pair of LSI SAS HBAs). And yet the BIOS sees everything just
fine and I can access the cards, the boot ROMs get installed properly
and I can see all of the attached disks. Allocation of device data above
4G is disabled in the BIOS. The "lspci" utility from
pkg:/system/pciutils doesn't see them either.

Strangely, however, prtdiag sees the slots as populated:

# prtdiag -v
...
==== Upgradeable Slots ====================================

ID  Status    Type             Description
--- --------- ---------------- ----------------------------
1   available Unknown          CPU1 Slot1
2   available Unknown          CPU1 Slot2
3   in use    Unknown          CPU2 Slot3
4   in use    Unknown          CPU2 Slot4
5   available Unknown          CPU2 Slot5
6   available Unknown          CPU2 Slot6

(the cards work in slots 1 & 2, but nowhere else)

Motherboard manual:
http://www.supermicro.com/manuals/motherboard/C606_602/MNL-1420.pdf

Can anybody help please? I'm seriously out of ideas here...

Here's the dmesg on anything PCI related:

# dmesg  | grep -i pci
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci8086,3c04 <at> 2, pcieb1
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pcieb1 is
/pci <at> 0,0/pci8086,3c04 <at> 2
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci8086,3c06 <at> 2,2, pcieb2
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pcieb2 is
/pci <at> 0,0/pci8086,3c06 <at> 2,2
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci8086,3c07 <at> 2,3, pcieb3
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pcieb3 is
/pci <at> 0,0/pci8086,3c07 <at> 2,3
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci8086,1d3e <at> 11, pcieb6
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pcieb6 is
/pci <at> 0,0/pci8086,1d3e <at> 11
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci8086,244e <at> 1e, pci_pci0
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pci_pci0 is
/pci <at> 0,0/pci8086,244e <at> 1e
Apr 30 05:12:32 head1 pcieb: [ID 586369 kern.info] PCIE-device:
pci10e3,8113 <at> 0, pcieb7
Apr 30 05:12:32 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci10e3,8113 <at> 0, pcieb7
Apr 30 05:12:32 head1 genunix: [ID 936769 kern.info] pcieb7 is
/pci <at> 0,0/pci8086,3c06 <at> 2,2/pci10e3,8113 <at> 0
Apr 30 05:12:33 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci15d9,720 <at> 1a, ehci0
Apr 30 05:12:33 head1 genunix: [ID 936769 kern.info] ehci0 is
/pci <at> 0,0/pci15d9,720 <at> 1a
Apr 30 05:12:34 head1 npe: [ID 236367 kern.info] PCI Express-device:
pci15d9,720 <at> 1d, ehci1
Apr 30 05:12:34 head1 genunix: [ID 936769 kern.info] ehci1 is
/pci <at> 0,0/pci15d9,720 <at> 1d
Apr 30 05:12:34 head1 genunix: [ID 936769 kern.info] mouse80420 is
/pci <at> 0,0/isa <at> 1f/i8042 <at> 1,60/mouse <at> 1
Apr 30 05:12:37 head1 genunix: [ID 936769 kern.info] hubd0 is
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1
Apr 30 05:12:37 head1 genunix: [ID 408114 kern.info]
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1 (hubd0) online
Apr 30 05:12:37 head1 genunix: [ID 936769 kern.info] hubd1 is
/pci <at> 0,0/pci15d9,720 <at> 1d/hub <at> 1
Apr 30 05:12:37 head1 genunix: [ID 408114 kern.info]
/pci <at> 0,0/pci15d9,720 <at> 1d/hub <at> 1 (hubd1) online
Apr 30 05:12:38 head1 genunix: [ID 936769 kern.info] usb_mid0 is
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3
Apr 30 05:12:38 head1 genunix: [ID 408114 kern.info]
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3 (usb_mid0) online
Apr 30 05:12:38 head1 genunix: [ID 936769 kern.info] hid0 is
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3/mouse <at> 0
Apr 30 05:12:38 head1 genunix: [ID 408114 kern.info]
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3/mouse <at> 0 (hid0) online
Apr 30 05:12:38 head1 genunix: [ID 936769 kern.info] hid1 is
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3/keyboard <at> 1
Apr 30 05:12:38 head1 genunix: [ID 408114 kern.info]
/pci <at> 0,0/pci15d9,720 <at> 1a/hub <at> 1/device <at> 3/keyboard <at> 1 (hid1) online
Apr 30 05:12:41 head1 genunix: [ID 936769 kern.info] asy0 is
/pci <at> 0,0/isa <at> 1f/asy <at> 1,3f8
Apr 30 05:12:41 head1 genunix: [ID 936769 kern.info] asy1 is
/pci <at> 0,0/isa <at> 1f/asy <at> 1,2f8

--
Saso

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com

Franz Schober | 26 Apr 2013 09:07
Picon
Gravatar

Emulex 10GbE NIC not working in SmartOS and OI

Hi,

I want to use an Emulex 10GbE OCe11102 in a HP DL360 G6 server, but the 
driver oce is not initializing,
on the console the output

oce[0]: MBOX Command Failed with Status: 1 0
oce[0]: Interface creation failed: 0x5
oce[0]: Failed to setup adapter

is shown, same for oce[1].

On my Supermicro X9-QR7-TF+ the card and driver work as expected,
the card also works in the HP server booting VMWare ESXi 5.1.

Do you have any suggestions ?

Thank you very much for your help,
Franz Schober

-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175548-9b54c406
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175548&id_secret=21175548-1d01cae1
Powered by Listbox: http://www.listbox.com


Gmane