Hi Li On Tue, Aug 26, 2008 at 12:06 PM, Li Zefan <[EMAIL PROTECTED]> wrote: > Michael Kerrisk wrote: >> Hi Li, >> >> On Tue, Aug 26, 2008 at 10:38 AM, Li Zefan <[EMAIL PROTECTED]> wrote: >>>> on 2.6.23 kernel on Mips and i686 arch. >>>> >>> ... >>>> -bash-3.2# ./inotify02 >>>> >>>> inotify02 1 PASS : get event: wd=1 mask=40000004 cookie=0 len=0 >>>> name="" >>>> inotify02 2 PASS : get event: wd=1 mask=100 cookie=0 len=16 >>>> name="test_file1" >>>> inotify02 3 PASS : get event: wd=1 mask=20 cookie=0 len=16 >>>> name="test_file1" >>>> inotify02 4 PASS : get event: wd=1 mask=8 cookie=0 len=16 >>>> name="test_file1" >>>> inotify02 5 PASS : get event: wd=1 mask=40 cookie=23 len=16 >>>> name="test_file1" >>>> inotify02 6 PASS : get event: wd=1 mask=80 cookie=23 len=16 >>>> name="test_file2" >>>> inotify02 7 PASS : get event: wd=1 mask=800 cookie=0 len=0 name="" >>>> inotify02 8 PASS : get event: wd=1 mask=200 cookie=0 len=16 >>>> name="test_file2" >>>> inotify02 9 PASS : get event: wd=1 mask=800 cookie=0 len=0 name="" >>>> inotify02 10 FAIL : get unnecessary event: wd=1 mask=800 cookie=0 >>>> len=0name="" >>>> -bash-3.2# >>>> >>>> >>>> This inotify02 test case failed. >>>> >>>> Please give fix for the same. >>>> >>> But fix what... >>> >>> You are using 2.6.23, and the bug in inotify was fixed in 2.6.25, so the >>> above >>> is exactly the expected result. >> >> Actually, what was the bug in 2.6.24 and earlier? Maybe it is >> something that needs to be documented in man-pages. >> > > In the manpage: > > If successive output inotify events produced on the inotify file > descriptor are identical (same wd, mask, cookie, and name) then they > are coalesced into a single event. > > This is not true in older kernels due to the bug.
Bother! I remember testing this, but only in a simple fashion: queuing two eevets that were the same, and fiding that they were coalesced. So if I understand correctly (and some testing now seems to indicate that this is true), then a new inotify event was checked to see if it could be coleasced with the oldest event that was queued, rather than the most recently queued event. Right? > The bug description and fix is 1c17d18e3775485bf1e0ce79575eb637a94494a2 > > I agree it's better to document it in man-pages. Yes, I will document this. Cheers, Michael -- Michael Kerrisk Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/ man-pages online: http://www.kernel.org/doc/man-pages/online_pages.html Found a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Ltp-list mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ltp-list
