Setuid sandbox: exit(2) on SIGABRT

Setuid sandbox: exit(2) on SIGABRT

The setuid sandbox waits on its one child and then dies
afterwards.

When receiving SIGABRT, instead of dumping core, simply exit the
process. There is no interesting information to be gathered from
knowing that the process is inside waitid(2), one should look at
the child process instead.

This patch hopes to reduce red herrings.

BUG=334345
NOTRY=true
R=jorgelo@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=251265

[jln (very slow on Chromium), 2014-02-14 01:18:10]

Jorge, do you mind taking a look?

Hopefully this will help making things less confusing.

[Jorge Lucangeli Obes, 2014-02-14 01:39:56]

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
File sandbox/linux/suid/sandbox.c (right):

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:61: static void ExitWithErrorSignalhandler(int
signal) {
Cap in "Handler"?

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:204: // Don't "Core" on SIGABRT. SIGABRT is sent by
the Chrome OS session manager
No caps in "core"?

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:205: // when things are hanging.
This is not entirely true. SIGFPE is sent when the browser process is hung.
SIGABRT is sent when SIGTERM or SIGFPE did not manage to kill the entire process
group.

[jln (very slow on Chromium), 2014-02-14 01:45:14]

PTAL!

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
File sandbox/linux/suid/sandbox.c (right):

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:61: static void ExitWithErrorSignalhandler(int
signal) {
On 2014/02/14 01:39:57, Jorge Lucangeli Obes wrote:
> Cap in "Handler"?

Done.

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:204: // Don't "Core" on SIGABRT. SIGABRT is sent by
the Chrome OS session manager
On 2014/02/14 01:39:57, Jorge Lucangeli Obes wrote:
> No caps in "core"?

I wanted to use the official POSIX actions "Term, Ign, Core etc."

https://chromiumcodereview.appspot.com/165673002/diff/1/sandbox/linux/suid/sa...
sandbox/linux/suid/sandbox.c:205: // when things are hanging.
On 2014/02/14 01:39:57, Jorge Lucangeli Obes wrote:
> This is not entirely true. SIGFPE is sent when the browser process is hung.
> SIGABRT is sent when SIGTERM or SIGFPE did not manage to kill the entire
process
> group.

No, SIGTERM is only sent to the browser process nowadays. I don't want to add
too much details, I thought "when things are hanging" was an accurate
description for SIGABRT :)

[Jorge Lucangeli Obes, 2014-02-14 01:59:23]

lgtm

I meant that SIGABRT can be sent even when the process is not hung (e.g. we
tried to SIGTERM the browser process and it took too long to shut down, so we
SIGABRT'd the whole process).

[jln (very slow on Chromium), 2014-02-14 02:05:28]

The CQ bit was checked by jln@chromium.org

[commit-bot: I haz the power, 2014-02-14 02:06:43]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jln@chromium.org/165673002/50001

[jln (very slow on Chromium), 2014-02-14 06:55:35]

The CQ bit was unchecked by jln@chromium.org

[jln (very slow on Chromium), 2014-02-14 06:56:04]

The CQ bit was checked by jln@chromium.org

[commit-bot: I haz the power, 2014-02-14 06:56:12]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jln@chromium.org/165673002/50001

[jln (very slow on Chromium), 2014-02-14 07:23:54]

Committed patchset #2 manually as r251265 (presubmit successful).