Re: lsbsi PPC test failures

On Wed, Jan 15, Wichmann, Mats D wrote:

> 30 Failures that need examining for PPC32
> 
> /tset/LSB.os/devclass/grantpt_L/T.grantpt_L 1 FAIL
> 200|734 1 00:59:45|TP Start
> 520|734 1 00006906 1 1|incorrect mode on /dev/pts/6: expected 0620 got
> 020600
> 220|734 1 1 00:59:45|FAIL

This ones means, that /dev/pts is not mounted with the correct options.
Add "mode=0620" as option when you mount /dev/pts.

> /tset/LSB.os/jump/getcontext/T.getcontext 1 FAIL
> 200|805 1 01:21:49|TP Start
> 520|805 1 00013528 1 1|getcontext returned -1 instead of 0
> 220|805 1 1 01:21:49|FAIL

As far as I know, getcontext and co. are not implemented
for PPC32 yet (at least our current distribution does not
have them, too). This would explain all *context errors.

Don't know about the mmap fails.

> /tset/LSB.os/procprim/exec_L/T.execl_L 5 UNREPORTED
> 200|877 5 01:34:14|TP Start
> 220|877 5 7 01:34:15|NORESULT (auto-generated by TCC)

This was a kernel, glibc or testsuite bug. I need to ask
our PPC developers tomorrow.
As far as I know they solved it.

> /tset/POSIX.os/devclass/tcgetattr/T.tcgetattr 3 FAIL
> 200|279 3 20:44:27|TP Start
> 520|279 3 00003273 1 1|tcgetattr() did not return actual input baud rate
> when input = output
> 520|279 3 00003273 1 2|expected 13, observed 0
> 220|279 3 1 20:44:27|FAIL

If I remember correct, this is a problem with the kernel on
some PPC hardware.

> /tset/POSIX.os/devclass/tcsetattr/T.tcsetattr 3 FAIL
> 200|282 3 20:45:51|TP Start
> 520|282 3 00003299 1 1|tcsetattr(TCSAFLUSH) discarded output
> 220|282 3 1 20:46:03|FAIL

Sounds like the kernel patch we also need for i386 is missing.

Don't know about the other one.

-- 
Thorsten Kukuk       http://www.suse.de/~kukuk/        [email protected]
SuSE Linux AG        Deutschherrnstr. 15-19        D-90429 Nuernberg
--------------------------------------------------------------------    
Key fingerprint = A368 676B 5E1B 3E46 CFCE  2D97 F8FD 4E23 56C6 FB4B