Update glibc revision

Update glibc revision

This brings in:
8ec02f0e Enable process and thread CPU time clocks

This change also enables untrusted tests for process and thread CPU
time clocks to check the end-to-end integration.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=2806
TEST= small_tests large_tests trybots
R=mcgrathr@chromium.org

Committed: https://src.chromium.org/viewvc/native_client?view=rev&revision=12098

[Roland McGrath, 2013-08-28 16:53:06]

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
File tests/clock/clock_test.c (right):

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
tests/clock/clock_test.c:88: if (-EINTR == err) {
Return value is an errno code, not a negated errno code.
EINTR is never possible in NaCl anyway, but it's fine to keep the code if it's
made correct.

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
tests/clock/clock_test.c:150: if (0 != (err = gettimeofday(&t_now_tv, NULL))) {
The return value is -1 and the error is in errno.

[Petr Hosek, 2013-08-28 19:31:11]

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
File tests/clock/clock_test.c (right):

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
tests/clock/clock_test.c:88: if (-EINTR == err) {
On 2013/08/28 16:53:06, Roland McGrath wrote:
> Return value is an errno code, not a negated errno code.
> EINTR is never possible in NaCl anyway, but it's fine to keep the code if it's
> made correct.

Oh, I completely forgot about those return values, fixed everywhere.

https://chromiumcodereview.appspot.com/23464007/diff/27001/tests/clock/clock_...
tests/clock/clock_test.c:150: if (0 != (err = gettimeofday(&t_now_tv, NULL))) {
On 2013/08/28 16:53:06, Roland McGrath wrote:
> The return value is -1 and the error is in errno.

ditto

[Roland McGrath, 2013-08-28 19:54:34]

lgtm with nits

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
File tests/clock/clock_test.c (right):

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:28: * On an unloaded i7, 1ms with a 1.25 fuzziness
factor and a 100,000
This comment belongs down a paragraph, right before the constants it's actually
talking about.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:45: * what is considered accurate.
Make these static.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:59: int ClockMonotonicAccuracyTest(uint64_t
sleep_nanos) {
Make the local functions static.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:100: elapsed_nanos = (t_end.tv_sec - t_start.tv_sec) *
NANOS_PER_UNIT
Style guide says operator at eol, not next bol.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:108: printf("actual elapsed sleep: %20"PRIu64" nS\n",
elapsed_nanos);
This shouldn't report as "actual" a value that includes slop.
At least print the slop separately.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:113: || elapsed_upper_bound < elapsed_nanos) {
operator line break again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:153: + t_now_tv.tv_usec * NANOS_PER_MICRO;
operator line break again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:184: for (i = 1; i < info->cycles; i++) {
This should have a comment on its intent.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:277: + (t_thread_end.tv_nsec - t_thread_start.tv_nsec);
again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:281: + (t_process_end.tv_nsec -
t_process_start.tv_nsec);
again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:293: + info[i].thread_time.tv_nsec;
again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:295: + info[i].process_time.tv_nsec;
again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:303: + child_thread_elapsed * g_fuzzy_factor +
g_syscall_overhead);
again

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:312: || elapsed_upper_bound < process_elapsed) {
again

[Petr Hosek, 2013-08-28 22:48:57]

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
File tests/clock/clock_test.c (right):

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:28: * On an unloaded i7, 1ms with a 1.25 fuzziness
factor and a 100,000
On 2013/08/28 19:54:34, Roland McGrath wrote:
> This comment belongs down a paragraph, right before the constants it's
actually
> talking about.

Moved down.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:45: * what is considered accurate.
On 2013/08/28 19:54:34, Roland McGrath wrote:
> Make these static.

Done.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:59: int ClockMonotonicAccuracyTest(uint64_t
sleep_nanos) {
On 2013/08/28 19:54:34, Roland McGrath wrote:
> Make the local functions static.

Done.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:100: elapsed_nanos = (t_end.tv_sec - t_start.tv_sec) *
NANOS_PER_UNIT
On 2013/08/28 19:54:34, Roland McGrath wrote:
> Style guide says operator at eol, not next bol.

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:108: printf("actual elapsed sleep: %20"PRIu64" nS\n",
elapsed_nanos);
On 2013/08/28 19:54:34, Roland McGrath wrote:
> This shouldn't report as "actual" a value that includes slop.
> At least print the slop separately.

Removed the "actual".

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:113: || elapsed_upper_bound < elapsed_nanos) {
On 2013/08/28 19:54:34, Roland McGrath wrote:
> operator line break again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:153: + t_now_tv.tv_usec * NANOS_PER_MICRO;
On 2013/08/28 19:54:34, Roland McGrath wrote:
> operator line break again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:184: for (i = 1; i < info->cycles; i++) {
On 2013/08/28 19:54:34, Roland McGrath wrote:
> This should have a comment on its intent.

Added.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:277: + (t_thread_end.tv_nsec - t_thread_start.tv_nsec);
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:281: + (t_process_end.tv_nsec -
t_process_start.tv_nsec);
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:293: + info[i].thread_time.tv_nsec;
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:295: + info[i].process_time.tv_nsec;
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:303: + child_thread_elapsed * g_fuzzy_factor +
g_syscall_overhead);
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

https://chromiumcodereview.appspot.com/23464007/diff/83001/tests/clock/clock_...
tests/clock/clock_test.c:312: || elapsed_upper_bound < process_elapsed) {
On 2013/08/28 19:54:34, Roland McGrath wrote:
> again

Fixed.

[bsy, 2013-08-29 16:32:57]

i'm happy if ronald^Wroland is happy.

[Petr Hosek, 2013-09-02 02:20:58]

Committed patchset #4 manually as r12098 (presubmit successful).