Bug 441420 – =net-analyzer/bwmon-1.3-r2 stable request

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Paweł Hajdan, Jr. (RETIRED) 2012-11-03 08:25:35 UTC]

Is it OK to stabilize =net-analyzer/bwmon-1.3-r1 ?

If so, please CC all arches which have stable keywords

for older versions of this package and add STABLEREQ keyword

to the bug.

[Comment 1 Jeroen Roovers (RETIRED) 2012-11-03 18:30:05 UTC]

Arch teams, please test and mark stable:
=net-analyzer/bwmon-1.3-r1
Stable KEYWORDS : amd64 hppa sparc x86

[Comment 2 Agostino Sarubbo 2012-11-04 20:22:37 UTC]

amd64 stable

[Comment 3 Vicente Olivert Riera (RETIRED) 2012-11-04 21:22:24 UTC]

I have added ppc arch to CC.

Tested ppc: works fine here.

[Comment 4 Vicente Olivert Riera (RETIRED) 2012-11-05 02:14:15 UTC]

I have found a problem on x86. If bwmon is compiled with -O2 optimization level, it crashes with a buffer overflow.
It works with -O0, -O1, -O3 and -Os. But, with -O3 it has a weird behaviour...

[Comment 5 Vicente Olivert Riera (RETIRED) 2012-11-05 02:15:38 UTC]

Created attachment 328396 [details]
buffer overflow

[Comment 6 Diego Elio Pettenò (RETIRED) 2012-11-05 02:16:22 UTC]

Can you get a stack trace?

[Comment 7 Vicente Olivert Riera (RETIRED) 2012-11-05 02:18:54 UTC]

Created attachment 328400 [details]
build log

[Comment 8 Vicente Olivert Riera (RETIRED) 2012-11-05 02:19:12 UTC]

Created attachment 328402 [details]
emerge info

[Comment 9 Vicente Olivert Riera (RETIRED) 2012-11-05 02:40:26 UTC]

Created attachment 328410 [details]
backtrace

[Comment 10 Diego Elio Pettenò (RETIRED) 2012-11-05 02:59:52 UTC]

Overflow fixed, there was an uninitialized buffer being used.

amd64 since you stabled a package that has a broken corner case (it's not restricted to x86, it just depends on whether the compiler is in the mood to optimize that call or not), you might want to stable -r2 in place of -r1.

[Comment 11 Sergey Popov 2012-11-05 09:15:55 UTC]

+  05 Nov 2012; Sergey Popov <[e-mail]> bwmon-1.3-r2.ebuild:
+  Add comment about overflow patch, stable on amd64 wrt bug #441420

[Comment 12 Agostino Sarubbo 2012-11-05 09:28:53 UTC]

(In reply to comment #10)
> Overflow fixed, there was an uninitialized buffer being used.
> 
> amd64 since you stabled a package that has a broken corner case (it's not
> restricted to x86, it just depends on whether the compiler is in the mood to
> optimize that call or not), you might want to stable -r2 in place of -r1.

I compiled with -O2 and there was not problem at runtime :/

[Comment 13 Diego Elio Pettenò (RETIRED) 2012-11-05 15:28:40 UTC]

As I said it depends on the compiler's mood.

[Comment 14 Jeroen Roovers (RETIRED) 2012-11-07 03:23:11 UTC]

Stable for HPPA.

[Comment 15 Anthony Basile 2012-11-07 13:14:36 UTC]

stable ppc

[Comment 16 Myckel Habets 2012-11-09 20:54:15 UTC]

Builds and runs fine on x86. Please mark stable for x86.

[Comment 17 Jeroen Roovers (RETIRED) 2012-11-21 04:54:35 UTC]

Stable for SPARC.

[Comment 18 Agostino Sarubbo 2012-11-21 11:16:44 UTC]

x86 stable. Last arch, closing