PCHECK when closing UDP sockets only when errno=EBADF.

PCHECK when closing UDP sockets only when errno=EBADF.

This is an attempto to resolve bug 461246. Issues there is that close()
fails sometimes on OSX for unknown reason. My guess is that it close()
fails with EIO error for some reason, but EIO may be ignored in this
case as the sockets gets closed anyway.

BUG=461246
Committed: https://crrev.com/494d408657f49528aa43827bc3602df1331ebfe3
Cr-Commit-Position: refs/heads/master@{#321483}

[Sergey Ulanov, 2015-03-19 21:43:34]

sergeyu@chromium.org changed reviewers:
+ rvargas@chromium.org

[rvargas (doing something else), 2015-03-19 22:30:21]

On 2015/03/19 21:43:35, Sergey Ulanov wrote:

But the previous change was in response to a flood of EBADF. Are things
different now?

[rvargas (doing something else), 2015-03-19 22:34:05]

https://codereview.chromium.org/1025533002/diff/1/net/udp/udp_socket_libevent.cc
File net/udp/udp_socket_libevent.cc (right):

https://codereview.chromium.org/1025533002/diff/1/net/udp/udp_socket_libevent...
net/udp/udp_socket_libevent.cc:130: PCHECK(errno != EBADF);
How about filtering by EIO instead and leaving errno on the stack (although that
is almost pointless as getting it out of a dump on mac is going to be a pain).

[Sergey Ulanov, 2015-03-19 22:39:37]

On 2015/03/19 22:30:21, rvargas wrote:
> On 2015/03/19 21:43:35, Sergey Ulanov wrote:
> 
> But the previous change was in response to a flood of EBADF. Are things
> different now?

That EBADF was in a different place. Crashes we are getting now may or may not
be EBADF. So here I'm trying figure out if we are getting EIO here.

[Sergey Ulanov, 2015-03-19 22:53:36]

https://codereview.chromium.org/1025533002/diff/1/net/udp/udp_socket_libevent.cc
File net/udp/udp_socket_libevent.cc (right):

https://codereview.chromium.org/1025533002/diff/1/net/udp/udp_socket_libevent...
net/udp/udp_socket_libevent.cc:130: PCHECK(errno != EBADF);
On 2015/03/19 22:34:05, rvargas wrote:
> How about filtering by EIO instead and leaving errno on the stack (although
that
> is almost pointless as getting it out of a dump on mac is going to be a pain).

Done.

[rvargas (doing something else), 2015-03-19 23:06:34]

lgtm

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
File net/udp/udp_socket_libevent.cc (right):

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
net/udp/udp_socket_libevent.cc:130: int last_error = errno;
You may be interested in base::debug::Alias

[rvargas (doing something else), 2015-03-19 23:07:18]

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
File net/udp/udp_socket_libevent.cc (right):

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
net/udp/udp_socket_libevent.cc:130: int last_error = errno;
On 2015/03/19 23:06:33, rvargas wrote:
> You may be interested in base::debug::Alias

nah, that's win only.

[Sergey Ulanov, 2015-03-19 23:52:03]

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
File net/udp/udp_socket_libevent.cc (right):

https://codereview.chromium.org/1025533002/diff/40001/net/udp/udp_socket_libe...
net/udp/udp_socket_libevent.cc:130: int last_error = errno;
On 2015/03/19 23:07:18, rvargas wrote:
> On 2015/03/19 23:06:33, rvargas wrote:
> > You may be interested in base::debug::Alias
> 
> nah, that's win only.

I think it may be useful on Mac too. Comments don't say anything about it being
windows-specific.

[Sergey Ulanov, 2015-03-19 23:52:20]

The CQ bit was checked by sergeyu@chromium.org

[Sergey Ulanov, 2015-03-19 23:52:21]

The patchset sent to the CQ was uploaded after l-g-t-m from rvargas@chromium.org
Link to the patchset: https://codereview.chromium.org/1025533002/#ps60001
(title: " ")

[commit-bot: I haz the power, 2015-03-19 23:53:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1025533002/60001

[commit-bot: I haz the power, 2015-03-20 01:12:04]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-03-20 01:12:41]

Patchset 4 (id:??) landed as
https://crrev.com/494d408657f49528aa43827bc3602df1331ebfe3
Cr-Commit-Position: refs/heads/master@{#321483}