linux/drivers/block/aoe/aoeblk.c

328 lines
7.4 KiB
C
Raw Normal View History

/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoeblk.c
* block device routines
*/
#include <linux/hdreg.h>
#include <linux/blkdev.h>
#include <linux/backing-dev.h>
#include <linux/fs.h>
#include <linux/ioctl.h>
#include <linux/genhd.h>
#include <linux/netdevice.h>
#include "aoe.h"
static struct kmem_cache *buf_pool_cache;
static ssize_t aoedisk_show_state(struct device *dev,
struct device_attribute *attr, char *page)
{
struct gendisk *disk = dev_to_disk(dev);
struct aoedev *d = disk->private_data;
return snprintf(page, PAGE_SIZE,
"%s%s\n",
(d->flags & DEVFL_UP) ? "up" : "down",
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
(d->flags & DEVFL_KICKME) ? ",kickme" :
(d->nopen && !(d->flags & DEVFL_UP)) ? ",closewait" : "");
/* I'd rather see nopen exported so we can ditch closewait */
}
static ssize_t aoedisk_show_mac(struct device *dev,
struct device_attribute *attr, char *page)
{
struct gendisk *disk = dev_to_disk(dev);
struct aoedev *d = disk->private_data;
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
struct aoetgt *t = d->targets[0];
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
if (t == NULL)
return snprintf(page, PAGE_SIZE, "none\n");
return snprintf(page, PAGE_SIZE, "%pm\n", t->addr);
}
static ssize_t aoedisk_show_netif(struct device *dev,
struct device_attribute *attr, char *page)
{
struct gendisk *disk = dev_to_disk(dev);
struct aoedev *d = disk->private_data;
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
struct net_device *nds[8], **nd, **nnd, **ne;
struct aoetgt **t, **te;
struct aoeif *ifp, *e;
char *p;
memset(nds, 0, sizeof nds);
nd = nds;
ne = nd + ARRAY_SIZE(nds);
t = d->targets;
te = t + NTARGETS;
for (; t < te && *t; t++) {
ifp = (*t)->ifs;
e = ifp + NAOEIFS;
for (; ifp < e && ifp->nd; ifp++) {
for (nnd = nds; nnd < nd; nnd++)
if (*nnd == ifp->nd)
break;
if (nnd == nd && nd != ne)
*nd++ = ifp->nd;
}
}
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
ne = nd;
nd = nds;
if (*nd == NULL)
return snprintf(page, PAGE_SIZE, "none\n");
for (p = page; nd < ne; nd++)
p += snprintf(p, PAGE_SIZE - (p-page), "%s%s",
p == page ? "" : ",", (*nd)->name);
p += snprintf(p, PAGE_SIZE - (p-page), "\n");
return p-page;
}
/* firmware version */
static ssize_t aoedisk_show_fwver(struct device *dev,
struct device_attribute *attr, char *page)
{
struct gendisk *disk = dev_to_disk(dev);
struct aoedev *d = disk->private_data;
return snprintf(page, PAGE_SIZE, "0x%04x\n", (unsigned int) d->fw_ver);
}
static DEVICE_ATTR(state, S_IRUGO, aoedisk_show_state, NULL);
static DEVICE_ATTR(mac, S_IRUGO, aoedisk_show_mac, NULL);
static DEVICE_ATTR(netif, S_IRUGO, aoedisk_show_netif, NULL);
static struct device_attribute dev_attr_firmware_version = {
.attr = { .name = "firmware-version", .mode = S_IRUGO },
.show = aoedisk_show_fwver,
};
static struct attribute *aoe_attrs[] = {
&dev_attr_state.attr,
&dev_attr_mac.attr,
&dev_attr_netif.attr,
&dev_attr_firmware_version.attr,
NULL,
};
static const struct attribute_group attr_group = {
.attrs = aoe_attrs,
};
static int
aoedisk_add_sysfs(struct aoedev *d)
{
return sysfs_create_group(&disk_to_dev(d->gd)->kobj, &attr_group);
}
void
aoedisk_rm_sysfs(struct aoedev *d)
{
sysfs_remove_group(&disk_to_dev(d->gd)->kobj, &attr_group);
}
static int
aoeblk_open(struct block_device *bdev, fmode_t mode)
{
struct aoedev *d = bdev->bd_disk->private_data;
ulong flags;
spin_lock_irqsave(&d->lock, flags);
if (d->flags & DEVFL_UP) {
d->nopen++;
spin_unlock_irqrestore(&d->lock, flags);
return 0;
}
spin_unlock_irqrestore(&d->lock, flags);
return -ENODEV;
}
static int
aoeblk_release(struct gendisk *disk, fmode_t mode)
{
struct aoedev *d = disk->private_data;
ulong flags;
spin_lock_irqsave(&d->lock, flags);
if (--d->nopen == 0) {
spin_unlock_irqrestore(&d->lock, flags);
aoecmd_cfg(d->aoemajor, d->aoeminor);
return 0;
}
spin_unlock_irqrestore(&d->lock, flags);
return 0;
}
static int
aoeblk_make_request(struct request_queue *q, struct bio *bio)
{
struct sk_buff_head queue;
struct aoedev *d;
struct buf *buf;
ulong flags;
blk_queue_bounce(q, &bio);
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
if (bio == NULL) {
printk(KERN_ERR "aoe: bio is NULL\n");
BUG();
return 0;
}
d = bio->bi_bdev->bd_disk->private_data;
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
if (d == NULL) {
printk(KERN_ERR "aoe: bd_disk->private_data is NULL\n");
BUG();
bio_endio(bio, -ENXIO);
return 0;
} else if (bio_rw_flagged(bio, BIO_RW_BARRIER)) {
bio_endio(bio, -EOPNOTSUPP);
return 0;
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
} else if (bio->bi_io_vec == NULL) {
printk(KERN_ERR "aoe: bi_io_vec is NULL\n");
BUG();
bio_endio(bio, -ENXIO);
return 0;
}
buf = mempool_alloc(d->bufpool, GFP_NOIO);
if (buf == NULL) {
printk(KERN_INFO "aoe: buf allocation failure\n");
bio_endio(bio, -ENOMEM);
return 0;
}
memset(buf, 0, sizeof(*buf));
INIT_LIST_HEAD(&buf->bufs);
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
buf->stime = jiffies;
buf->bio = bio;
buf->resid = bio->bi_size;
buf->sector = bio->bi_sector;
buf->bv = &bio->bi_io_vec[bio->bi_idx];
buf->bv_resid = buf->bv->bv_len;
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
WARN_ON(buf->bv_resid == 0);
buf->bv_off = buf->bv->bv_offset;
spin_lock_irqsave(&d->lock, flags);
if ((d->flags & DEVFL_UP) == 0) {
printk(KERN_INFO "aoe: device %ld.%d is not up\n",
d->aoemajor, d->aoeminor);
spin_unlock_irqrestore(&d->lock, flags);
mempool_free(buf, d->bufpool);
bio_endio(bio, -ENXIO);
return 0;
}
list_add_tail(&buf->bufs, &d->bufq);
aoecmd_work(d);
__skb_queue_head_init(&queue);
skb_queue_splice_init(&d->sendq, &queue);
spin_unlock_irqrestore(&d->lock, flags);
aoenet_xmit(&queue);
return 0;
}
static int
aoeblk_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
struct aoedev *d = bdev->bd_disk->private_data;
if ((d->flags & DEVFL_UP) == 0) {
printk(KERN_ERR "aoe: disk not up\n");
return -ENODEV;
}
geo->cylinders = d->geo.cylinders;
geo->heads = d->geo.heads;
geo->sectors = d->geo.sectors;
return 0;
}
static const struct block_device_operations aoe_bdops = {
.open = aoeblk_open,
.release = aoeblk_release,
.getgeo = aoeblk_getgeo,
.owner = THIS_MODULE,
};
/* alloc_disk and add_disk can sleep */
void
aoeblk_gdalloc(void *vp)
{
struct aoedev *d = vp;
struct gendisk *gd;
ulong flags;
gd = alloc_disk(AOE_PARTITIONS);
if (gd == NULL) {
printk(KERN_ERR
"aoe: cannot allocate disk structure for %ld.%d\n",
d->aoemajor, d->aoeminor);
goto err;
}
d->bufpool = mempool_create_slab_pool(MIN_BUFS, buf_pool_cache);
if (d->bufpool == NULL) {
printk(KERN_ERR "aoe: cannot allocate bufpool for %ld.%d\n",
d->aoemajor, d->aoeminor);
goto err_disk;
}
aoe: allocate unused request_queue for sysfs Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f071524647e660706b784ebcbefbd2e44 block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by: Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 12:10:18 +00:00
d->blkq = blk_alloc_queue(GFP_KERNEL);
if (!d->blkq)
goto err_mempool;
aoe: allocate unused request_queue for sysfs Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f071524647e660706b784ebcbefbd2e44 block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by: Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 12:10:18 +00:00
blk_queue_make_request(d->blkq, aoeblk_make_request);
d->blkq->backing_dev_info.name = "aoe";
aoe: allocate unused request_queue for sysfs Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f071524647e660706b784ebcbefbd2e44 block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by: Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 12:10:18 +00:00
if (bdi_init(&d->blkq->backing_dev_info))
goto err_blkq;
spin_lock_irqsave(&d->lock, flags);
gd->major = AOE_MAJOR;
gd->first_minor = d->sysminor * AOE_PARTITIONS;
gd->fops = &aoe_bdops;
gd->private_data = d;
set_capacity(gd, d->ssize);
aoe: handle multiple network paths to AoE device A remote AoE device is something can process ATA commands and is identified by an AoE shelf number and an AoE slot number. Such a device might have more than one network interface, and it might be reachable by more than one local network interface. This patch tracks the available network paths available to each AoE device, allowing them to be used more efficiently. Andrew Morton asked about the call to msleep_interruptible in the revalidate function. Yes, if a signal is pending, then msleep_interruptible will not return 0. That means we will not loop but will call aoenet_xmit with a NULL skb, which is a noop. If the system is too low on memory or the aoe driver is too low on frames, then the user can hit control-C to interrupt the attempt to do a revalidate. I have added a comment to the code summarizing that. Andrew Morton asked whether the allocation performed inside addtgt could use a more relaxed allocation like GFP_KERNEL, but addtgt is called when the aoedev lock has been locked with spin_lock_irqsave. It would be nice to allocate the memory under fewer restrictions, but targets are only added when the device is being discovered, and if the target can't be added right now, we can try again in a minute when then next AoE config query broadcast goes out. Andrew Morton pointed out that the "too many targets" message could be printed for failing GFP_ATOMIC allocations. The last patch in this series makes the messages more specific. Signed-off-by: Ed L. Cashin <ecashin@coraid.com> Cc: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-08 12:20:00 +00:00
snprintf(gd->disk_name, sizeof gd->disk_name, "etherd/e%ld.%d",
d->aoemajor, d->aoeminor);
aoe: allocate unused request_queue for sysfs Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f071524647e660706b784ebcbefbd2e44 block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by: Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 12:10:18 +00:00
gd->queue = d->blkq;
d->gd = gd;
d->flags &= ~DEVFL_GDALLOC;
d->flags |= DEVFL_UP;
spin_unlock_irqrestore(&d->lock, flags);
add_disk(gd);
aoedisk_add_sysfs(d);
return;
aoe: allocate unused request_queue for sysfs Andy Whitcroft reported an oops in aoe triggered by use of an incorrectly initialised request_queue object: [ 2645.959090] kobject '<NULL>' (ffff880059ca22c0): tried to add an uninitialized object, something is seriously wrong. [ 2645.959104] Pid: 6, comm: events/0 Not tainted 2.6.31-5-generic #24-Ubuntu [ 2645.959107] Call Trace: [ 2645.959139] [<ffffffff8126ca2f>] kobject_add+0x5f/0x70 [ 2645.959151] [<ffffffff8125b4ab>] blk_register_queue+0x8b/0xf0 [ 2645.959155] [<ffffffff8126043f>] add_disk+0x8f/0x160 [ 2645.959161] [<ffffffffa01673c4>] aoeblk_gdalloc+0x164/0x1c0 [aoe] The request queue of an aoe device is not used but can be allocated in code that does not sleep. Bruno bisected this regression down to cd43e26f071524647e660706b784ebcbefbd2e44 block: Expose stacked device queues in sysfs "This seems to generate /sys/block/$device/queue and its contents for everyone who is using queues, not just for those queues that have a non-NULL queue->request_fn." Addresses http://bugs.launchpad.net/bugs/410198 Addresses http://bugzilla.kernel.org/show_bug.cgi?id=13942 Note that embedding a queue inside another object has always been an illegal construct, since the queues are reference counted and must persist until the last reference is dropped. So aoe was always buggy in this respect (Jens). Signed-off-by: Ed Cashin <ecashin@coraid.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Bruno Premont <bonbons@linux-vserver.org> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 12:10:18 +00:00
err_blkq:
blk_cleanup_queue(d->blkq);
d->blkq = NULL;
err_mempool:
mempool_destroy(d->bufpool);
err_disk:
put_disk(gd);
err:
spin_lock_irqsave(&d->lock, flags);
d->flags &= ~DEVFL_GDALLOC;
spin_unlock_irqrestore(&d->lock, flags);
}
void
aoeblk_exit(void)
{
kmem_cache_destroy(buf_pool_cache);
}
int __init
aoeblk_init(void)
{
buf_pool_cache = kmem_cache_create("aoe_bufs",
sizeof(struct buf),
0, 0, NULL);
if (buf_pool_cache == NULL)
return -ENOMEM;
return 0;
}