On Tue, Jan 18, 2011 at 02:51:32PM -0800, John-Mark Gurney wrote: > Konstantin Belousov wrote this message on Tue, Jan 18, 2011 at 21:57 +0000: > > sf_buf_alloc() calls msleep(PCATCH) when SFB_CATCH flag was given, > > and for simultaneous wakeup and signal delivery, msleep() returns > > EINTR/ERESTART despite the thread was selected for wakeup_one(). As > > result, we loose a wakeup, and some other waiter will not be woken up. > > Shouldn't this behavior be documented in the man page? That even > though msleep may return a non-zero value that it could have been > really woken up? Can you provide the good wording ? And commit it then ?
pgpL1MtT3jnV7.pgp
Description: PGP signature
