We need to leak the fib, as it is still owned by the controller until the command has infact completed. The interrupt service completing the command, if it ever occurs, will cause the FIB to be returned to the pool. Sincerely -- Mark Salyzyn > -----Original Message----- > From: Christoph Hellwig [mailto:hch@xxxxxxxxxxxxx] > Sent: Friday, August 04, 2006 11:20 AM > To: Mark Haverkamp > Cc: James Bottomley; Salyzyn, Mark; linux-scsi > Subject: Re: [PATCH 1/4] aacraid: interruptible ioctl > > > > -#include <linux/delay.h> > > +#include <linux/delay.h> /* ssleep prototype */ > > do we really need that comment? > > > #include <linux/kthread.h> > > #include <asm/semaphore.h> > > #include <asm/uaccess.h> > > @@ -140,7 +140,8 @@ > > fibptr->hw_fib_pa = hw_fib_pa; > > fibptr->hw_fib = hw_fib; > > } > > - aac_fib_free(fibptr); > > + if (retval != -EINTR) > > + aac_fib_free(fibptr); > > even if we are interrupted we shouldn't leak the fib, should we? > > - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
