Bug 159314 – sandbox access violation /usr/src/linux/-.gcda with external kernel modules for amd64 >=kernel 2.6.19

• Status: RESOLVED DUPLICATE of bug 135745
• Product: Gentoo Linux

[Description Martin Väth 2006-12-28 14:49:30 UTC]

Although I suspect that this is an analogous problem to bug #149307,
I was told there that it is not directly related:

Every attempt on amd64 with some kernel >= 2.6.19 (even the newest ones, containing patches for #149307) to emerge some ebuild containing kernel modules leads to the same type of sandbox access violation:

--------------------------- ACCESS VIOLATION SUMMARY ----------------
LOG FILE = "/var/log/sandbox/[... snipped ...]"

open_wr:   /usr/src/linux-2.6.19-hardened-r3/-.gcda
open_wr:   /usr/src/linux-2.6.19-hardened-r3/-.gcda
---------------------------------------------------------------------

The problem does not occur on x86 (and of course also not with
FEATURES=-usersandbox).

[Comment 1 Jakub Moc (RETIRED) 2006-12-28 15:04:51 UTC]

*** This bug has been marked as a duplicate of 156807 ***

[Comment 2 Matthias Schwarzott 2006-12-28 15:25:04 UTC]

Reopen to reverse duplicate/original.
To trace that specific sandbox violation better.

[Comment 3 Matthias Schwarzott 2006-12-28 15:26:27 UTC]

*** Bug 156807 has been marked as a duplicate of this bug. ***

[Comment 4 Daniel Drake (RETIRED) 2006-12-30 18:03:09 UTC]

Please post "emerge --info" and attach the full session of compiling an affected package

[Comment 5 Daniel Drake (RETIRED) 2006-12-30 18:03:29 UTC]

and kernel .config

[Comment 6 Martin Väth 2006-12-31 02:35:01 UTC]

Sorry, I cannot reach my amd64 system for the next few days (unexpected connection shutdown). However, in the compilation logs, IIRC, there never was something unexpected (in particular, no visible command concerning "gcda"), only the violation message at the end of the installation(?) stage.
Maybe some of the "CC" people will attach the information?
Otherwise, you will have to wait for some days, sorry.

[Comment 7 Matthias Schwarzott 2006-12-31 05:05:30 UTC]

As long as I understand gcc-behaviour, it just tries to open sourcefile.gcda (to check its existance).
And in some weird place of kernel-makefiles gcc gets source over stdin "-" and also tries to open gcda file for that.

Related bug (duplicate?) is: Bug #135745.

[Comment 8 Thomas Sachau 2006-12-31 09:20:59 UTC]

Created attachment 105052 [details]
output of emerge --info

[Comment 9 Thomas Sachau 2006-12-31 09:21:48 UTC]

Created attachment 105053 [details]
.config (kernel 2.6.19-r2)

[Comment 10 Thomas Sachau 2006-12-31 09:22:32 UTC]

Created attachment 105054 [details]
output of emerge alsa-driver

[Comment 11 Daniel Drake (RETIRED) 2007-01-02 20:59:48 UTC]

(In reply to comment #7)
> As long as I understand gcc-behaviour, it just tries to open sourcefile.gcda
> (to check its existance).

It shouldn't be opening it for writing then... any comments from gcc maintainers?

> And in some weird place of kernel-makefiles gcc gets source over stdin "-" and
> also tries to open gcda file for that.
> 
> Related bug (duplicate?) is: Bug #135745.
> 

[Comment 12 SpanKY 2007-01-03 01:48:27 UTC]

*** This bug has been marked as a duplicate of 135745 ***

[Comment 13 Jakub Moc (RETIRED) 2007-01-10 21:39:24 UTC]

*** Bug 161414 has been marked as a duplicate of this bug. ***