[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
R=thakis@chromium.org

Committed: https://chromium.googlesource.com/chromium/src/+/0c97b34adb439d638986d50769cba543e6ad5056

[Robert Sesek, 2016-01-27 18:44:31]

rsesek@chromium.org changed reviewers:
+ thestig@chromium.org

[Robert Sesek, 2016-01-27 18:46:46]

Description was changed from

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
==========

to

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
==========

[Robert Sesek, 2016-01-27 18:46:46]

rsesek@chromium.org changed reviewers:
+ thakis@chromium.org

[Robert Sesek, 2016-01-27 18:47:33]

+Nico as well since this is blocking M49 webview

[Nico, 2016-01-27 18:54:34]

Code lgtm, but maybe the CL description could motivate why this is changed a bit
more.

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
File base/process/process_iterator_linux.cc (right):

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
base/process/process_iterator_linux.cc:62: if (!procfs_dir_) {
Can you add a comment when this can happen? From the bug, sounds like this can
happen on Android -- and it kind of sounds like it might happen there
consistently. Is it useful to LOG(ERROR) about this on android when this happens
if it's going to happen every time there?

What do we use ProcessIterator for on Android? I thought it's mostly for task
manager and about:memory etc, which don't exist there.

[Robert Sesek, 2016-01-27 18:58:42]

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
File base/process/process_iterator_linux.cc (right):

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
base/process/process_iterator_linux.cc:62: if (!procfs_dir_) {
On 2016/01/27 18:54:34, Nico wrote:
> Can you add a comment when this can happen? From the bug, sounds like this can
> happen on Android -- and it kind of sounds like it might happen there
> consistently. Is it useful to LOG(ERROR) about this on android when this
happens
> if it's going to happen every time there?

Added a comment. This *shouldn't* happen every time, it seems like the SELinux
policies may not be sorted out right.

> What do we use ProcessIterator for on Android? I thought it's mostly for task
> manager and about:memory etc, which don't exist there.

Given what nnk@ wrote on the bug, it seems like collecting memory stats this way
isn't going to work anyways. I'll file a bug with the person who wrote
memory_details_android.cc.

[Robert Sesek, 2016-01-27 19:11:11]

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
File base/process/process_iterator_linux.cc (right):

https://codereview.chromium.org/1642663003/diff/1/base/process/process_iterat...
base/process/process_iterator_linux.cc:62: if (!procfs_dir_) {
On 2016/01/27 18:58:42, Robert Sesek wrote:
> On 2016/01/27 18:54:34, Nico wrote:
> > Can you add a comment when this can happen? From the bug, sounds like this
can
> > happen on Android -- and it kind of sounds like it might happen there
> > consistently. Is it useful to LOG(ERROR) about this on android when this
> happens
> > if it's going to happen every time there?
> 
> Added a comment. This *shouldn't* happen every time, it seems like the SELinux
> policies may not be sorted out right.
> 
> > What do we use ProcessIterator for on Android? I thought it's mostly for
task
> > manager and about:memory etc, which don't exist there.
> 
> Given what nnk@ wrote on the bug, it seems like collecting memory stats this
way
> isn't going to work anyways. I'll file a bug with the person who wrote
> memory_details_android.cc.

Filed https://code.google.com/p/chromium/issues/detail?id=581778

[commit-bot: I haz the power, 2016-01-27 19:16:32]

Description was changed from

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
==========

to

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
R=thakis@chromium.org

Committed: https://crrev.com/0c97b34adb439d638986d50769cba543e6ad5056
Cr-Commit-Position: refs/heads/master@{#371837}
==========

[commit-bot: I haz the power, 2016-01-27 19:16:32]

Patchset 2 (id:??) landed as
https://crrev.com/0c97b34adb439d638986d50769cba543e6ad5056
Cr-Commit-Position: refs/heads/master@{#371837}

[Robert Sesek, 2016-01-27 19:17:08]

Description was changed from

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
R=thakis@chromium.org

Committed: https://crrev.com/0c97b34adb439d638986d50769cba543e6ad5056
Cr-Commit-Position: refs/heads/master@{#371837}
==========

to

==========
[Linux] Do not try to call readdir() on a null DIR* in base::ProcessIterator.

BUG=581517
R=thakis@chromium.org

Committed:
https://chromium.googlesource.com/chromium/src/+/0c97b34adb439d638986d50769cb...
==========

[Robert Sesek, 2016-01-27 19:17:10]

Committed patchset #2 (id:20001) manually as
0c97b34adb439d638986d50769cba543e6ad5056 (presubmit successful).