From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Cc: Indan Zupancic <indan@nul.nu>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
Evgeniy Polyakov <johnpol@2ka.mipt.ru>,
Daniel Phillips <phillips@google.com>,
Rik van Riel <riel@redhat.com>,
David Miller <davem@davemloft.net>
Subject: [PATCH 3/4] nbd: deadlock prevention for NBD
Date: Fri, 25 Aug 2006 17:40:17 +0200 [thread overview]
Message-ID: <20060825154017.24271.20362.sendpatchset@twins> (raw)
In-Reply-To: <20060825153946.24271.42758.sendpatchset@twins>
Use sk_set_vmio() on the nbd socket.
Limit each request to 1 page, so that the request throttling also limits the
number of in-flight pages and force the IO scheduler to NOOP as anything else
doesn't make sense anyway.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Daniel Phillips <phillips@google.com>
---
drivers/block/nbd.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
Index: linux-2.6/drivers/block/nbd.c
===================================================================
--- linux-2.6.orig/drivers/block/nbd.c
+++ linux-2.6/drivers/block/nbd.c
@@ -135,7 +135,6 @@ static int sock_xmit(struct socket *sock
spin_unlock_irqrestore(¤t->sighand->siglock, flags);
do {
- sock->sk->sk_allocation = GFP_NOIO;
iov.iov_base = buf;
iov.iov_len = size;
msg.msg_name = NULL;
@@ -361,8 +360,16 @@ static void nbd_do_it(struct nbd_device
BUG_ON(lo->magic != LO_MAGIC);
+ sk_adjust_memalloc(0, 1);
+ if (sk_set_vmio(lo->sock->sk))
+ printk(KERN_WARNING
+ "failed to set SOCK_VMIO on NBD socket\n");
+
while ((req = nbd_read_stat(lo)) != NULL)
nbd_end_request(req);
+
+ sk_adjust_memalloc(0, -1);
+
return;
}
@@ -525,6 +533,7 @@ static int nbd_ioctl(struct inode *inode
if (S_ISSOCK(inode->i_mode)) {
lo->file = file;
lo->sock = SOCKET_I(inode);
+ lo->sock->sk->sk_allocation = GFP_NOIO;
error = 0;
} else {
fput(file);
@@ -628,11 +637,16 @@ static int __init nbd_init(void)
* every gendisk to have its very own request_queue struct.
* These structs are big so we dynamically allocate them.
*/
- disk->queue = blk_init_queue(do_nbd_request, &nbd_lock);
+ disk->queue = blk_init_queue_node_elv(do_nbd_request,
+ &nbd_lock, -1, "noop");
if (!disk->queue) {
put_disk(disk);
goto out;
}
+ blk_queue_pin_elevator(disk->queue);
+ blk_queue_max_segment_size(disk->queue, PAGE_SIZE);
+ blk_queue_max_hw_segments(disk->queue, 1);
+ blk_queue_max_phys_segments(disk->queue, 1);
}
if (register_blkdev(NBD_MAJOR, "nbd")) {
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Cc: Indan Zupancic <indan@nul.nu>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
Evgeniy Polyakov <johnpol@2ka.mipt.ru>,
Daniel Phillips <phillips@google.com>,
Rik van Riel <riel@redhat.com>,
David Miller <davem@davemloft.net>
Subject: [PATCH 3/4] nbd: deadlock prevention for NBD
Date: Fri, 25 Aug 2006 17:40:17 +0200 [thread overview]
Message-ID: <20060825154017.24271.20362.sendpatchset@twins> (raw)
In-Reply-To: <20060825153946.24271.42758.sendpatchset@twins>
Use sk_set_vmio() on the nbd socket.
Limit each request to 1 page, so that the request throttling also limits the
number of in-flight pages and force the IO scheduler to NOOP as anything else
doesn't make sense anyway.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Daniel Phillips <phillips@google.com>
---
drivers/block/nbd.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
Index: linux-2.6/drivers/block/nbd.c
===================================================================
--- linux-2.6.orig/drivers/block/nbd.c
+++ linux-2.6/drivers/block/nbd.c
@@ -135,7 +135,6 @@ static int sock_xmit(struct socket *sock
spin_unlock_irqrestore(¤t->sighand->siglock, flags);
do {
- sock->sk->sk_allocation = GFP_NOIO;
iov.iov_base = buf;
iov.iov_len = size;
msg.msg_name = NULL;
@@ -361,8 +360,16 @@ static void nbd_do_it(struct nbd_device
BUG_ON(lo->magic != LO_MAGIC);
+ sk_adjust_memalloc(0, 1);
+ if (sk_set_vmio(lo->sock->sk))
+ printk(KERN_WARNING
+ "failed to set SOCK_VMIO on NBD socket\n");
+
while ((req = nbd_read_stat(lo)) != NULL)
nbd_end_request(req);
+
+ sk_adjust_memalloc(0, -1);
+
return;
}
@@ -525,6 +533,7 @@ static int nbd_ioctl(struct inode *inode
if (S_ISSOCK(inode->i_mode)) {
lo->file = file;
lo->sock = SOCKET_I(inode);
+ lo->sock->sk->sk_allocation = GFP_NOIO;
error = 0;
} else {
fput(file);
@@ -628,11 +637,16 @@ static int __init nbd_init(void)
* every gendisk to have its very own request_queue struct.
* These structs are big so we dynamically allocate them.
*/
- disk->queue = blk_init_queue(do_nbd_request, &nbd_lock);
+ disk->queue = blk_init_queue_node_elv(do_nbd_request,
+ &nbd_lock, -1, "noop");
if (!disk->queue) {
put_disk(disk);
goto out;
}
+ blk_queue_pin_elevator(disk->queue);
+ blk_queue_max_segment_size(disk->queue, PAGE_SIZE);
+ blk_queue_max_hw_segments(disk->queue, 1);
+ blk_queue_max_phys_segments(disk->queue, 1);
}
if (register_blkdev(NBD_MAJOR, "nbd")) {
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2006-08-25 15:45 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-08-25 15:39 [PATCH 0/4] VM deadlock prevention -v5 Peter Zijlstra
2006-08-25 15:39 ` Peter Zijlstra
2006-08-25 15:39 ` [PATCH 1/4] net: VM deadlock avoidance framework Peter Zijlstra
2006-08-25 15:39 ` Peter Zijlstra
2006-08-26 2:37 ` Indan Zupancic
2006-08-26 2:37 ` Indan Zupancic
2006-08-28 10:22 ` Peter Zijlstra
2006-08-28 10:22 ` Peter Zijlstra
2006-08-28 16:03 ` Indan Zupancic
2006-08-28 16:03 ` Indan Zupancic
2006-08-28 17:32 ` Peter Zijlstra
2006-08-28 17:32 ` Peter Zijlstra
2006-08-29 0:01 ` Indan Zupancic
2006-08-29 0:01 ` Indan Zupancic
2006-08-29 9:49 ` Peter Zijlstra
2006-08-29 9:49 ` Peter Zijlstra
2006-08-29 19:53 ` Indan Zupancic
2006-08-29 19:53 ` Indan Zupancic
2006-08-25 15:40 ` [PATCH 2/4] blkdev: iosched selection for queue creation Peter Zijlstra
2006-08-25 15:40 ` Peter Zijlstra
2006-08-25 15:40 ` Peter Zijlstra [this message]
2006-08-25 15:40 ` [PATCH 3/4] nbd: deadlock prevention for NBD Peter Zijlstra
2006-08-25 15:40 ` [PATCH 4/4] nfs: deadlock prevention for NFS Peter Zijlstra
2006-08-25 15:40 ` Peter Zijlstra
2006-08-25 20:14 ` Trond Myklebust
2006-08-25 20:14 ` Trond Myklebust
2006-08-25 20:36 ` Peter Zijlstra
2006-08-25 20:36 ` Peter Zijlstra
2006-08-26 3:05 ` Indan Zupancic
2006-08-26 3:05 ` Indan Zupancic
2006-08-25 15:51 ` [PATCH 0/4] VM deadlock prevention -v5 Christoph Lameter
2006-08-25 15:51 ` Christoph Lameter
2006-08-25 15:52 ` Peter Zijlstra
2006-08-25 15:52 ` Peter Zijlstra
2006-08-25 16:04 ` Rik van Riel
2006-08-25 16:04 ` Rik van Riel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20060825154017.24271.20362.sendpatchset@twins \
--to=a.p.zijlstra@chello.nl \
--cc=davem@davemloft.net \
--cc=indan@nul.nu \
--cc=johnpol@2ka.mipt.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=netdev@vger.kernel.org \
--cc=phillips@google.com \
--cc=riel@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.