>-----Original Message-----
>From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
>Behalf Of Dev, Vasu
>Sent: Tuesday, August 19, 2008 5:43 PM
>To: Mike Christie
>Cc: [email protected]
>Subject: Re: [Open-FCoE] some fc exch reset fixes and a fix to my abort
>patches
>
>>-----Original Message-----
>>From: Mike Christie [mailto:[EMAIL PROTECTED]
>>Sent: Tuesday, August 19, 2008 3:24 AM
>>
>>Actually :)... forget that patch. Attached is a new one
>>fix-my-f-ups2.patch that fixes more possible bugs. I also stopped
>>abusing the ESB bits and defined internal state bits.
>>
>
>The same bug of system crash after "WARNING: at net/ipv4/tcp_timer.c:290"
>occurred after applying fix-my-f-ups2.patch. I root caused this bug to
>passing up abort response to fcp while still freeing abort response
>frame as per older abort code in fc_exch_abts_resp() caused
>double frame free for eth driver allocated skb buffer and that likely
>messed up my networking stack. This explains ssh session hanged first
>and then system hanged after network skb pool messed up. Attached patch
>fixed that double frame free in fcp abort path and with this fix
>your 14 patches + fix-my-f-ups2.patch set works fine. I'm applying
>these patches to git tree.
I didn't receive attached patch back, so likely mail list or my mail
server removed it. Though attachment on this list worked for Mike
in some cases but he also had same problem before. Anway that
attached patch is pasted below.
diff --git a/drivers/scsi/libfc/fc_exch.c b/drivers/scsi/libfc/fc_exch.c
index 6fd8030..a1ef274 100644
--- a/drivers/scsi/libfc/fc_exch.c
+++ b/drivers/scsi/libfc/fc_exch.c
@@ -1347,10 +1347,12 @@ static void fc_exch_abts_resp(struct fc_exch *ep,
struct fc_frame *fp)
if (resp)
resp(sp, fp, ex_resp_arg);
+ else
+ fc_frame_free(fp);
+
if (has_rec)
fc_exch_timer_set(ep, ep->r_a_tov);
- fc_frame_free(fp);
}
_______________________________________________
devel mailing list
[email protected]
http://www.open-fcoe.org/mailman/listinfo/devel
