These are the fixes steeleye makes to use nbd for block device replication in
a production environment (they've been field tested for at least six months
now on various 2.4 kernels):
- make nbd daemon killable
- make signals clean up the queues correctly (otherwise the queue will be in
an
insane state if the deamon exits normally with a signal pending)
- increase the device size to 2TB
We've OK'd them with Pavel Machek, if you could apply them.
Thanks,
James Bottomley
You can import this changeset into BK by piping this whole message to:
'| bk receive [path to repository]' or apply the patch as usual.
===================================================================
ChangeSet@1.856, 2002-11-01 09:59:57-05:00, Paul.Clements@steeleye.com
Minor fixes and cleanups to nbd:
make nbd daemon killable
make signals clean up the queues correctly (otherwise the queue will be in an
insane state if the deamon exits normally with a signal pending)
increase the device size to 2TB
nbd.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---
-
1 files changed, 65 insertions(+), 4 deletions(-)
diff -Nru a/drivers/block/nbd.c b/drivers/block/nbd.c
--- a/drivers/block/nbd.c Fri Nov 1 10:02:08 2002
+++ b/drivers/block/nbd.c Fri Nov 1 10:02:08 2002
@@ -26,6 +26,8 @@
* reduce number of partial TCP segments sent. <steve@chygwyn.com>
* 01-12-6 Fix deadlock condition by making queue locks independant of
* the transmit lock. <steve@chygwyn.com>
+ * 02-10-11 Allow hung xmit to be aborted via SIGKILL & various fixes.
+ * <Paul.Clements@SteelEye.com> <James.Bottomley@SteelEye.com>
*
* possible FIXME: make set_sock / set_blksize / set_size / do_it one syscall
* why not: would need verify_area and friends, would share yet another
@@ -91,10 +93,12 @@
oldfs = get_fs();
set_fs(get_ds());
-
+ /* Allow interception of SIGKILL only
+ * Don't allow other signals to interrupt the transmission */
spin_lock_irqsave(¤t->sig->siglock, flags);
oldset = current->blocked;
sigfillset(¤t->blocked);
+ sigdelsetmask(¤t->blocked, sigmask(SIGKILL));
recalc_sigpending();
spin_unlock_irqrestore(¤t->sig->siglock, flags);
@@ -117,6 +121,17 @@
else
result = sock_recvmsg(sock, &msg, size, 0);
+ if (signal_pending(current)) {
+ siginfo_t info;
+ spin_lock_irqsave(¤t->sig->siglock, flags);
+ printk(KERN_WARNING "NBD (pid %d: %s) got signal %d\n",
+ current->pid, current->comm,
+ dequeue_signal(¤t->blocked, &info));
+ spin_unlock_irqrestore(¤t->sig->siglock, flags);
+ result = -EINTR;
+ break;
+ }
+
if (result <= 0) {
#ifdef PARANOIA
printk(KERN_ERR "NBD: %s - sock=%ld at buf=%ld, size=%d returned %d.\n",
@@ -156,6 +171,10 @@
down(&lo->tx_lock);
+ if (!sock || !lo->sock) {
+ FAIL("Attempted sendmsg to closed socket\n");
+ }
+
result = nbd_xmit(1, sock, (char *) &request, sizeof(request), nbd_cmd(req)
== NBD_CMD_WRITE ? MSG_MORE : 0);
if (result <= 0)
FAIL("Sendmsg failed for control.");
@@ -343,11 +362,31 @@
spin_unlock_irq(q->queue_lock);
spin_lock(&lo->queue_lock);
+
+ if (!lo->file) {
+ spin_unlock(&lo->queue_lock);
+ printk(KERN_ERR "nbd: failed between accept and semaphore, file lost\n");
+ req->errors++;
+ nbd_end_request(req);
+ spin_lock_irq(q->queue_lock);
+ continue;
+ }
+
list_add(&req->queuelist, &lo->queue_head);
spin_unlock(&lo->queue_lock);
nbd_send_req(lo, req);
+ if (req->errors) {
+ printk(KERN_ERR "nbd: nbd_send_req failed\n");
+ spin_lock(&lo->queue_lock);
+ list_del(&req->queuelist);
+ spin_unlock(&lo->queue_lock);
+ nbd_end_request(req);
+ spin_lock_irq(q->queue_lock);
+ continue;
+ }
+
spin_lock_irq(q->queue_lock);
continue;
@@ -390,12 +429,14 @@
printk(KERN_ERR "nbd: Some requests are in progress -> can not turn
off.\n");
return -EBUSY;
}
- spin_unlock(&lo->queue_lock);
file = lo->file;
- if (!file)
+ if (!file) {
+ spin_unlock(&lo->queue_lock);
return -EINVAL;
+ }
lo->file = NULL;
lo->sock = NULL;
+ spin_unlock(&lo->queue_lock);
fput(file);
return 0;
case NBD_SET_SOCK:
@@ -439,6 +480,26 @@
if (!lo->file)
return -EINVAL;
nbd_do_it(lo);
+ /* on return tidy up in case we have a signal */
+ /* Forcibly shutdown the socket causing all listeners
+ * to error
+ *
+ * FIXME: This code is duplicated from sys_shutdown, but
+ * there should be a more generic interface rather than
+ * calling socket ops directly here */
+ down(&lo->tx_lock);
+ printk(KERN_WARNING "nbd: shutting down socket\n");
+ lo->sock->ops->shutdown(lo->sock, SEND_SHUTDOWN|RCV_SHUTDOWN);
+ lo->sock = NULL;
+ up(&lo->tx_lock);
+ spin_lock(&lo->queue_lock);
+ file = lo->file;
+ lo->file = NULL;
+ spin_unlock(&lo->queue_lock);
+ nbd_clear_que(lo);
+ printk(KERN_WARNING "nbd: queue cleared\n");
+ if (file)
+ fput(file);
return lo->harderror;
case NBD_CLEAR_QUE:
nbd_clear_que(lo);
@@ -517,7 +578,7 @@
init_MUTEX(&nbd_dev[i].tx_lock);
nbd_dev[i].blksize = 1024;
nbd_dev[i].blksize_bits = 10;
- nbd_dev[i].bytesize = 0x7ffffc00; /* 2GB */
+ nbd_dev[i].bytesize = ((u64)0x7ffffc00) << 10; /* 2TB */
disk->major = MAJOR_NR;
disk->first_minor = i;
disk->fops = &nbd_fops;
===================================================================
This BitKeeper patch contains the following changesets:
+
## Wrapped with gzip_uu ##
begin 664 bkpatch2828
M'XL(`'"7PCT``]U76U/;.!1^CG_%*9W2I"6)[-BY<9D"H6T6FG8";'>F[604
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M3&CU2FG&8G;###*\=*%,G`/XM>X<.B&\XT*F,.+73`$5$80QHR*;*E0%&/XV
M+L%O0B^8&4)$62(%7/`XIL.8W9I5?"QHK'(`R*:@)PPN,Y8A<"C3E(4ZOH&B
M1'$ZXXJMYV&&:#!DP`6:@'A<*"H045.-PI%=&3%J%+-KKA6@R0F-$6[&]03H
M0C=@5B(NQJ7<*B["E-&%HHA=\=`8^2<SGGEG!\XQU&LMC,*'-6.<\K_\.`ZA
MQ-E[)"]1R@UQJ\-8AA>&U)7P5H9\)`D.O49C7J\W&ZV1'S:B5J-)O=9W*/%-
M4-<EKNL'K<"?8_8;+5L3#RQ^O#I^VFSGBJ?R58*@E:G**BS*/BTU??FVW<1M
M$<\+B#>ODV8ML.7BDWO5XO]`M=0#*/O_2;G\/PLA)\I[**<S^T5B?WB(,S]1
M'UVO"9X#+S"/99>471?VXUC.8)*),5PG7!LKT&4ZE*EF$5QQ"J?=-\?=DQ/8
MA"N:<IFIO#]5#`S`SMW,GYK,'^69WX.=WVB"*P^DUC)!0MR==CHM'URGB[^>
M4ZB^6-C"A69IR*::8V#E:*5?BOC&*:#2CA3/-5"[V.9ME60TWNY.LZFV\<TI
MQI4R6"^JJ*N!&@NX/F*Q8CJAZJ*X&6;(`Z'+>S:X+-HR@'9JH;M4VG:ZKML"
M%S<7,/7%7.-@D=KB`J%4@K]P@<'G8B0'&LQCVXJF7`P,_("GEPK/VEMJ<;G]
M,=-;,(KI6)7LIFF*[EP4CX_ZO<''_7ZOVWL#&[V##A2G/()G41N>J1*,I5Y2
M[5GT66QLF:V%%3HNW8+5"`.?;(%=$3'+\T&^]Z$P;!KS2Z6U`YE8NI`RA?7\
M0U[@TBS&\Q_*1]W>6=_*AE@#%^;M;^<SAC9H@N_8P#Y1N!WF<W@22X3#01[3
MU_O=D^+&OM8LF1IF*HQ\HL8FY6$LE9$8HS7Z;[1:V)H?@$OP+<^911SQF"VS
MM/:HN&GF\G"8\?WP'_7[L&&.8!A1A(BP2O2,,6P-H>&J/:T52^AT@E%!]W$-
MH%U+>TP4+LM[2$V9JI<OK031!NC%(#5Y4+J(S])]KA0O[QL62J&YR-@J@+6`
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M%)?R+3@]ZG4&IV_/SSKO/_;F_</?5X,[`-B'>N<G)T:43>];]WWBV[+>A64C
M6>(NQ$O<Q_AOZ&_N(^D`)]"#1\*2WT;LAG6%&H9;@AMZCZ:9SD?;3B?PB"F%
M_)$KPVO&)_ZE,KS1S%XU=J%8S.I^B5PW1O@)"2G!S@YVD&U`<N(MQ.1O]4]=
7.&'AA<J2W8#YK<"K4><?L7%3&44.````
`
end
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Nov 07 2002 - 22:00:20 EST