OpenVZ Forum


Home » Mailing lists » Devel » LTP syscall failures
LTP syscall failures [message #13121] Mon, 21 May 2007 07:47 Go to next message
Christian Kaiser2 is currently offline  Christian Kaiser2
Messages: 27
Registered: May 2007
Junior Member
Hi,

I did some testing with the LTP Regression Test Suite
(http://ltp.sourceforge.net) inside a VE the last days. When I enable all
capabilites except "setpcap" I get one warning and three failures. When I
run the tests native (in VE0) I get no errors. I tested it with several
Kernels including the latest 2.6.18 and 2.6.20 from the git repository, all
with the same results.

I run the syscall test suite of ltp-20070430 as root with the command:
./runltp -p -q -f syscalls -l `date +"%Y-%m-%d_%H-%M-%S"`

Failed tests: syslog12, ustat01, ustat02

Output:
syslog12
syslog system call: read and/or clear kernel message ring
buffer; set console_loglevel

syslog12 5 FAIL : syslog() returned unexpected
results for
console level less than 0 ; returned 0
(expected -1), errno 0 (expected 22)
syslog12 6 FAIL : syslog() returned unexpected
results for console level
greater than 8 ; returned 0 (expected -1),
errno 0 (expected 22)

ustat0[1-2]
ustat system call: get filesystem statistics

ustat01 1 FAIL : ustat(2) failed and setthe errno to
13
: Permission denied
ustat02 1 PASS : ustat(2) expected failure; Got
errno - EINVAL :
Invalid parameter
ustat02 2 FAIL : ustat(2) failed to produce expected
error; 14, errno:
EFAULT and got 13

As you can see the syslog12 and ustat02 tests fail because they return
other error values than expected. The ustat01 test fails because of
unsufficient permissions.

Is there someone out who wants to confirm or to comment why these System
Call Tests fail?

Mit freundlichen Grüßen / Best Regards
Christian Kaiser

IBM Deutschland Entwicklung GmbH
Open Systems Firmware Development
mail: ckaiser2@de.ibm.com


IBM Deutschland Entwicklung GmbH
Vorsitzender des Aufsichtsrats: Johann Weihen
Geschäftsführung: Herbert Kircher
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
Re: LTP syscall failures [message #13124 is a reply to message #13121] Mon, 21 May 2007 09:14 Go to previous messageGo to next message
Alexey Dobriyan is currently offline  Alexey Dobriyan
Messages: 195
Registered: August 2006
Senior Member
On Mon, May 21, 2007 at 09:47:31AM +0200, Christian Kaiser2 wrote:
> I did some testing with the LTP Regression Test Suite
> (http://ltp.sourceforge.net) inside a VE the last days. When I enable all
> capabilites except "setpcap" I get one warning and three failures. When I
> run the tests native (in VE0) I get no errors. I tested it with several
> Kernels including the latest 2.6.18 and 2.6.20 from the git repository, all
> with the same results.
>
> I run the syscall test suite of ltp-20070430 as root with the command:
> ./runltp -p -q -f syscalls -l `date +"%Y-%m-%d_%H-%M-%S"`
>
> Failed tests: syslog12, ustat01, ustat02
>
> Output:
> syslog12
> syslog system call: read and/or clear kernel message ring
> buffer; set console_loglevel
>
> syslog12 5 FAIL : syslog() returned unexpected
> results for
> console level less than 0 ; returned 0
> (expected -1), errno 0 (expected 22)
> syslog12 6 FAIL : syslog() returned unexpected
> results for console level
> greater than 8 ; returned 0 (expected -1),
> errno 0 (expected 22)
>
> ustat0[1-2]
> ustat system call: get filesystem statistics
>
> ustat01 1 FAIL : ustat(2) failed and setthe errno to
> 13
> : Permission denied
> ustat02 1 PASS : ustat(2) expected failure; Got
> errno - EINVAL :
> Invalid parameter
> ustat02 2 FAIL : ustat(2) failed to produce expected
> error; 14, errno:
> EFAULT and got 13
>
> As you can see the syslog12 and ustat02 tests fail because they return
> other error values than expected. The ustat01 test fails because of
> unsufficient permissions.
>
> Is there someone out who wants to confirm or to comment why these System
> Call Tests fail?

I've reporoduces ustat unexpected EACCES problem and filed report
at http://bugzilla.openvz.org/show_bug.cgi?id=584
Re: LTP syscall failures [message #13193 is a reply to message #13121] Wed, 23 May 2007 10:21 Go to previous messageGo to next message
gblond is currently offline  gblond
Messages: 64
Registered: May 2007
Member
1) Fail with ustat() is a openvz bug and it will be fixed in the near future.
2) Fail with syslog() is a normal behavior, but it can be modified specially
for LTP.


>From: Christian Kaiser2 <CKAISER2@de.ibm.com>
>To: devel@openvz.org
>Message-ID:
><OFA30B9023.1D7CA759-ONC12572E2.0029ABDB-C12572E2.002AE3A1@de.ibm.com>
>Date: Mon, 21 May 2007 09:47:31 +0200
>Subject: [Devel] LTP syscall failures
>
>Hi,
>
>I did some testing with the LTP Regression Test Suite
>(http://ltp.sourceforge.net) inside a VE the last days. When I enable all
>capabilites except "setpcap" I get one warning and three failures. When I
>run the tests native (in VE0) I get no errors. I tested it with several
>Kernels including the latest 2.6.18 and 2.6.20 from the git repository, all
>with the same results.
>
>I run the syscall test suite of ltp-20070430 as root with the command:
>./runltp -p -q -f syscalls -l `date +"%Y-%m-%d_%H-%M-%S"`
>
>Failed tests: syslog12, ustat01, ustat02
>
>Output:
> syslog12
> syslog system call: read and/or clear kernel message ring
> buffer; set console_loglevel
>
> syslog12 5 FAIL : syslog() returned unexpected
>results for
> console level less than 0 ; returned 0
>(expected -1), errno 0 (expected 22)
> syslog12 6 FAIL : syslog() returned unexpected
>results for console level
> greater than 8 ; returned 0 (expected -1),
>errno 0 (expected 22)
>
> ustat0[1-2]
> ustat system call: get filesystem statistics
>
> ustat01 1 FAIL : ustat(2) failed and setthe errno to
>13
> : Permission denied
> ustat02 1 PASS : ustat(2) expected failure; Got
>errno - EINVAL :
> Invalid parameter
> ustat02 2 FAIL : ustat(2) failed to produce expected
>error; 14, errno:
> EFAULT and got 13
>
>As you can see the syslog12 and ustat02 tests fail because they return
>other error values than expected. The ustat01 test fails because of
>unsufficient permissions.
>
>Is there someone out who wants to confirm or to comment why these System
>Call Tests fail?
>
>Mit freundlichen Grüßen / Best Regards
>Christian Kaiser

--

Vitaliy Gusev
Re: LTP syscall failures [message #13217 is a reply to message #13193] Wed, 23 May 2007 11:51 Go to previous messageGo to next message
Christian Kaiser2 is currently offline  Christian Kaiser2
Messages: 27
Registered: May 2007
Junior Member
Vitaliy Gusev <vgusev@openvz.org> wrote on 05/23/2007 12:21:41 PM:

> 1) Fail with ustat() is a openvz bug and it will be fixed in the near
future.
> 2) Fail with syslog() is a normal behavior, but it can be modified
specially
> for LTP.
>

But how I mentioned before, the syslog12 test does not fail on my native
machine. So what do you mean exactly by "normal behaviour"?

>
> >From: Christian Kaiser2 <CKAISER2@de.ibm.com>
> >To: devel@openvz.org
> >Message-ID:
> ><OFA30B9023.1D7CA759-ONC12572E2.0029ABDB-C12572E2.002AE3A1@de.ibm.com>
> >Date: Mon, 21 May 2007 09:47:31 +0200
> >Subject: [Devel] LTP syscall failures
> >
> >Hi,
> >
> >I did some testing with the LTP Regression Test Suite
> >(http://ltp.sourceforge.net) inside a VE the last days. When I enable
all
> >capabilites except "setpcap" I get one warning and three failures. When
I
> >run the tests native (in VE0) I get no errors. I tested it with several
> >Kernels including the latest 2.6.18 and 2.6.20 from the git repository,
all
> >with the same results.
> >
> >I run the syscall test suite of ltp-20070430 as root with the command:
> >./runltp -p -q -f syscalls -l `date +"%Y-%m-%d_%H-%M-%S"`
> >
> >Failed tests: syslog12, ustat01, ustat02
> >
> >Output:
> > syslog12
> > syslog system call: read and/or clear kernel message
ring
> > buffer; set console_loglevel
> >
> > syslog12 5 FAIL : syslog() returned unexpected
> >results for
> > console level less than 0 ; returned 0
> >(expected -1), errno 0 (expected 22)
> > syslog12 6 FAIL : syslog() returned unexpected
> >results for console level
> > greater than 8 ; returned 0 (expected
-1),
> >errno 0 (expected 22)
> >
> > ustat0[1-2]
> > ustat system call: get filesystem statistics
> >
> > ustat01 1 FAIL : ustat(2) failed and setthe errno
to
> >13
> > : Permission denied
> > ustat02 1 PASS : ustat(2) expected failure; Got
> >errno - EINVAL :
> > Invalid parameter
> > ustat02 2 FAIL : ustat(2) failed to produce
expected
> >error; 14, errno:
> > EFAULT and got 13
> >
> >As you can see the syslog12 and ustat02 tests fail because they return
> >other error values than expected. The ustat01 test fails because of
> >unsufficient permissions.
> >
> >Is there someone out who wants to confirm or to comment why these System
> >Call Tests fail?
> >
> >Mit freundlichen Grüßen / Best Regards
> >Christian Kaiser
>
> --
>
> Vitaliy Gusev
>

Mit freundlichen Grüßen / Best Regards
Christian Kaiser
--
IBM Deutschland Entwicklung GmbH
Open Systems Firmware Development
mail: ckaiser2@de.ibm.com

IBM Deutschland Entwicklung GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
Re: LTP syscall failures [message #13240 is a reply to message #13217] Wed, 23 May 2007 14:20 Go to previous message
gblond is currently offline  gblond
Messages: 64
Registered: May 2007
Member
I meant that if program in VE calls syscall(type, buf, len); with type==8
(set log level) then kernel doesn't permit to do it with no error return.

> Vitaliy Gusev <vgusev@openvz.org> wrote on 05/23/2007 12:21:41 PM:
> > 1) Fail with ustat() is a openvz bug and it will be fixed in the near
>
> future.
>
> > 2) Fail with syslog() is a normal behavior, but it can be modified
>
> specially
>
> > for LTP.
>
> But how I mentioned before, the syslog12 test does not fail on my native
> machine. So what do you mean exactly by "normal behaviour"?
>
> Mit freundlichen Grüßen / Best Regards
> Christian Kaiser
> --
> IBM Deutschland Entwicklung GmbH
> Open Systems Firmware Development
> mail: ckaiser2@de.ibm.com
>
> IBM Deutschland Entwicklung GmbH
> Vorsitzender des Aufsichtsrats: Martin Jetter
> Geschäftsführung: Herbert Kircher
> Sitz der Gesellschaft: Böblingen
> Registergericht: Amtsgericht Stuttgart, HRB 243294

--

Vitaliy Gusev,
Previous Topic: [PATCH 0/15] Make common helpers for seq_files that work with list_head-s (v2)
Next Topic: [RFC][PATCH 0/3] Containers: Pagecache accounting and control subsystem (v3)
Goto Forum:
  


Current Time: Sat Nov 16 22:31:08 GMT 2024

Total time taken to generate the page: 0.03031 seconds