Commit Graph

992 Commits

Author SHA1 Message Date
Alexey Dobriyan 99b7623380 proc 2/2: remove struct proc_dir_entry::owner
Setting ->owner as done currently (pde->owner = THIS_MODULE) is racy
as correctly noted at bug #12454. Someone can lookup entry with NULL
->owner, thus not pinning enything, and release it later resulting
in module refcount underflow.

We can keep ->owner and supply it at registration time like ->proc_fops
and ->data.

But this leaves ->owner as easy-manipulative field (just one C assignment)
and somebody will forget to unpin previous/pin current module when
switching ->owner. ->proc_fops is declared as "const" which should give
some thoughts.

->read_proc/->write_proc were just fixed to not require ->owner for
protection.

rmmod'ed directories will be empty and return "." and ".." -- no harm.
And directories with tricky enough readdir and lookup shouldn't be modular.
We definitely don't want such modular code.

Removing ->owner will also make PDE smaller.

So, let's nuke it.

Kudos to Jeff Layton for reminding about this, let's say, oversight.

http://bugzilla.kernel.org/show_bug.cgi?id=12454

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
2009-03-31 01:14:44 +04:00
Linus Torvalds 4496d937a5 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: irq_node.handler() should return irqreturn_t
  m68k: section mismatch fixes: Atari SCSI
  m68k: section mismatch fixes: DMAsound for Atari
  MAINTAINERS: Replace dead link to m68k CVS repository by link to new git repository
  m68k: mac - Add SWIM floppy support
  m68k: mac - Add a new entry in mac_model to identify the floppy controller type.
  m68k: Add install target
2009-03-26 16:15:31 -07:00
Linus Torvalds 86d9c07017 Merge branch 'for-2.6.30' of git://git.kernel.dk/linux-2.6-block
* 'for-2.6.30' of git://git.kernel.dk/linux-2.6-block:
  Get rid of pdflush_operation() in emergency sync and remount
  btrfs: get rid of current_is_pdflush() in btrfs_btree_balance_dirty
  Move the default_backing_dev_info out of readahead.c and into backing-dev.c
  block: Repeated lines in switching-sched.txt
  bsg: Remove bogus check against request_queue->max_sectors
  block: WARN in __blk_put_request() for potential bio leak
  loop: fix circular locking in loop_clr_fd()
  loop: support barrier writes
  bsg: add support for tail queuing
  cpqarray: enable bus mastering
  block: genhd.h cleanup patch
  block: add private bio_set for bio integrity allocations
  block: genhd.h comment needs updating
  block: get rid of unused blkdev_free_rq() define
  block: remove various blk_queue_*() setting functions in blk_init_queue_node()
  cciss: add BUILD_BUG_ON() for catching bad CommandList_struct alignment
  block: don't create bio_vec slabs of less than the inline number
  block: cleanup bio_alloc_bioset()
2009-03-26 16:03:04 -07:00
David S. Miller 08abe18af1 Merge branch 'master' of /home/davem/src/GIT/linux-2.6/
Conflicts:
	drivers/net/wimax/i2400m/usb-notif.c
2009-03-26 15:23:24 -07:00
Laurent Vivier 8852ecd974 m68k: mac - Add SWIM floppy support
It allows to read data from a floppy, but not to write to, and to eject the
floppy (useful on our Mac without eject button).

Signed-off-by: Laurent Vivier <Laurent@lvivier.info>
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
2009-03-26 21:15:27 +01:00
Linus Torvalds 61a091827e Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (97 commits)
  USB: qcserial: add device id for HP devices
  USB: isp1760: Add a delay before reading the SKIPMAP registers in isp1760-hcd.c
  USB: allow malformed LANGID descriptors
  USB: pxa27x_udc: typo fixes and code cleanups
  USB: gadget: gadget zero uses new suspend/resume hooks
  USB: gadget: composite device-level suspend/resume hooks
  USB: r8a66597-hcd: suspend/resume support
  USB: more u32 conversion after transfer_buffer_length and actual_length
  USB: Fix cp2101 USB serial device driver termios functions for console use
  USB: CP2101 New Device ID
  USB: ipaq: handle 4 endpoint devices
  USB: S3C: Move usb-control.h to platform include
  USB: ohci-hcd: Add ARCH_S3C24XX to the ohci-s3c2410.c glue
  USB: pedantic: spelling correction in comment for ch9.h
  USB: host: fix sparse warning: Using plain integer as NULL pointer
  USB: ohci-s3c2410: fix name of bus clock
  USB: ohci-s3c2410: remove <mach/hardware.h> include
  USB: serial: rename cp2101 driver to cp210x
  USB: CP2101 Reduce Error Logging
  USB: CP2101 Support AN205 baud rates
  ...
2009-03-26 11:17:39 -07:00
Nikanth Karthikesan f028f3b2f9 loop: fix circular locking in loop_clr_fd()
With CONFIG_PROVE_LOCKING enabled

$ losetup /dev/loop0 file
$ losetup -o 32256 /dev/loop1 /dev/loop0

$ losetup -d /dev/loop1
$ losetup -d /dev/loop0

triggers a [ INFO: possible circular locking dependency detected ]

I think this warning is a false positive.

Open/close on a loop device acquires bd_mutex of the device before
acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after
acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of
a device, if backing file is a device and this is the last reference to the
file being dropped . But it is guaranteed that it is impossible to have a
circular list of backing devices.(say loop2->loop1->loop0->loop2 is not
possible), which guarantees that this can never deadlock.

So this warning should be suppressed. It is very difficult to annotate lockdep
not to warn here in the correct way. A simple way to silence lockdep could be
to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some
other real bugs.

@@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode,
 	struct loop_device *lo = bdev->bd_disk->private_data;
 	int err;

-	mutex_lock(&lo->lo_ctl_mutex);
+	mutex_lock_nested(&lo->lo_ctl_mutex, 1);
 	switch (cmd) {
 	case LOOP_SET_FD:
 		err = loop_set_fd(lo, mode, bdev, arg);

Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be
a better solution.

Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex
held, so no lockdep annotation is required.

If you do not like the special handling of the lo_ctl_mutex just for the
LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside
each of the individual ioctl handlers and I could send you another patch.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-26 11:01:19 +01:00
Eric Miao 71b3e0c1ad platform: make better use of to_platform_{device,driver}() macros
This helps the code look more consistent and cleaner.

Signed-off-by: Eric Miao <eric.miao@marvell.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-24 16:38:24 -07:00
Alan Stern e6e244b6cb usb-storage: prepare for subdriver separation
This patch (as1206) is the first step in converting usb-storage's
subdrivers into separate modules.  It makes the following large-scale
changes:

	Remove a bunch of unnecessary #ifdef's from usb_usual.h.
	Not truly necessary, but it does clean things up.

	Move the USB device-ID table (which is duplicated between
	libusual and usb-storage) into its own source file,
	usual-tables.c, and arrange for this to be linked with
	either libusual or usb-storage according to whether
	USB_LIBUSUAL is configured.

	Add to usual-tables.c a new usb_usual_ignore_device()
	function to detect whether a particular device needs to be
	managed by a subdriver and not by the standard handlers
	in usb-storage.

	Export a whole bunch of functions in usb-storage, renaming
	some of them because their names don't already begin with
	"usb_stor_".  These functions will be needed by the new
	subdriver modules.

	Split usb-storage's probe routine into two functions.
	The subdrivers will call the probe1 routine, then fill in
	their transport and protocol settings, and then call the
	probe2 routine.

	Take the default cases and error checking out of
	get_transport() and get_protocol(), which run during
	probe1, and instead put a check for invalid transport
	or protocol values into the probe2 function.

	Add a new probe routine to be used for standard devices,
	i.e., those that don't need a subdriver.  This new routine
	checks whether the device should be ignored (because it
	should be handled by ub or by a subdriver), and if not,
	calls the probe1 and probe2 functions.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-24 16:20:34 -07:00
Julia Lawall db5e6df172 USB: ub: use USB API functions rather than constants
This set of patches introduces calls to the following set of functions:

usb_endpoint_dir_in(epd)
usb_endpoint_dir_out(epd)
usb_endpoint_is_bulk_in(epd)
usb_endpoint_is_bulk_out(epd)
usb_endpoint_is_int_in(epd)
usb_endpoint_is_int_out(epd)
usb_endpoint_num(epd)
usb_endpoint_type(epd)
usb_endpoint_xfer_bulk(epd)
usb_endpoint_xfer_control(epd)
usb_endpoint_xfer_int(epd)
usb_endpoint_xfer_isoc(epd)

In some cases, introducing one of these functions is not possible, and it
just replaces an explicit integer value by one of the following constants:

USB_ENDPOINT_XFER_BULK
USB_ENDPOINT_XFER_CONTROL
USB_ENDPOINT_XFER_INT
USB_ENDPOINT_XFER_ISOC

An extract of the semantic patch that makes these changes is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@r1@ struct usb_endpoint_descriptor *epd; @@

- ((epd->bmAttributes & \(USB_ENDPOINT_XFERTYPE_MASK\|3\)) ==
- \(USB_ENDPOINT_XFER_CONTROL\|0\))
+ usb_endpoint_xfer_control(epd)

@r5@ struct usb_endpoint_descriptor *epd; @@

- ((epd->bEndpointAddress & \(USB_ENDPOINT_DIR_MASK\|0x80\)) ==
-  \(USB_DIR_IN\|0x80\))
+ usb_endpoint_dir_in(epd)

@inc@
@@

#include <linux/usb.h>

@depends on !inc && (r1||r5)@
@@

+ #include <linux/usb.h>
  #include <linux/usb/...>
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-24 16:20:27 -07:00
Nikanth Karthikesan 68db1961bb loop: support barrier writes
Honour barrier requests in the loop back block device driver.
In case of barrier bios, flush the backing file once before processing the
barrier and once after to guarantee ordering. In case of filesystems that
does not support fsync, barrier bios would be failed with -EOPNOTSUPP.

Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 12:35:18 +01:00
Dave Jones 0061d38642 cpqarray: enable bus mastering
We've been carrying this patch for the last 3 years in Fedora,
long past time we got it upstream...

Call pci_set_master to enable bus-mastering if the BIOS hasn't
done it already.

Signed-off-by: Kyle McMartin <kyle@redhat.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 12:35:17 +01:00
Jens Axboe 10cbda97e7 cciss: add BUILD_BUG_ON() for catching bad CommandList_struct alignment
The hardware requires 64-bit alignment of commands, so add a build bug
check for that. The recent commit 8a3173de4a
didn't change the size of the command, but other additions/changes may and
thus break badly at runtime.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 12:35:16 +01:00
Benjamin Herrenschmidt c71327ad9f Merge commit 'gcl/merge' into merge 2009-03-18 13:16:30 +11:00
Grant Likely bfbd442f69 Fix Xilinx SystemACE driver to handle empty CF slot
The SystemACE driver does not handle an empty CF slot gracefully. An
empty CF slot ends up hanging the system. This patch adds a check for
the CF state and stops trying to process requests if the slot is empty.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-14 21:06:52 +01:00
Geert Uytterhoeven f507cd2203 ps3/block: Replace mtd/ps3vram by block/ps3vram
Convert the PS3 Video RAM Storage Driver from an MTD driver to a plain block
device driver.

The ps3vram driver exposes unused video RAM on the PS3 as a block device
suitable for storage or swap.  Fast data transfer is achieved using a local
cache in system RAM and DMA transfers via the GPU.

The new driver is ca. 50% faster for reading, and ca. 10% for writing.

Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Acked-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-03-13 16:07:19 +11:00
Stephen Hemminger 7546dd97d2 net: convert usage of packet_type to read_mostly
Protocols that use packet_type can be __read_mostly section for better
locality. Elminate any unnecessary initializations of NULL.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-03-10 05:22:43 -07:00
Linus Torvalds df0b4a5080 Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (29 commits)
  p54: fix race condition in memory management
  cfg80211: test before subtraction on unsigned
  iwlwifi: fix error flow in iwl*_pci_probe
  rt2x00 : more devices to rt73usb.c
  rt2x00 : more devices to rt2500usb.c
  bonding: Fix device passed into ->ndo_neigh_setup().
  vlan: Fix vlan-in-vlan crashes.
  net: Fix missing dev->neigh_setup in register_netdevice().
  tmspci: fix request_irq race
  pkt_sched: act_police: Fix a rate estimator test.
  tg3: Fix 5906 link problems
  SCTP: change sctp_ctl_sock_init() to try IPv4 if IPv6 fails
  IPv6: add "disable" module parameter support to ipv6.ko
  sungem: another error printed one too early
  aoe: error printed 1 too early
  net pcmcia: worklimit reaches -1
  net: more timeouts that reach -1
  net: fix tokenring license
  dm9601: new vendor/product IDs
  netlink: invert error code in netlink_set_err()
  ...
2009-03-09 09:15:40 -07:00
Roel Kluin a3941ec101 loop: don't increment p->offset with (size_t) -EINVAL
Upon a 'transfer error block' size is set to -EINVAL, but this becomes positive
since size is unsigned: p->offset still gets incremented.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-05 12:04:57 +01:00
Jens Axboe 5e18cfd04f cciss: remove 30 second initial timeout on controller reset
Commit 5e4c91c84b forgot to remove the
initial sleep, get rid of it.

Thanks to Randy Dunlap <randy.dunlap@oracle.com> for spotting this error.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-05 12:04:57 +01:00
Kris Shannon a1a15ac5f9 Fix kernel NULL pointer dereference in xen-blkfront
When booting Xen Dom0 on a pre-release 3.2.1 hypervisor the system Oopses on a
"Unable to handle kernel NULL pointer dereference" in xenwatch.

From the backtrace it looks like backend_changed is calling bdget_disk
with a NULL pointer.  Checking for NULL and returning ENODEV instead
allows the kernel to boot.
2009-03-05 12:04:57 +01:00
Roel Kluin 9487311157 aoe: error printed 1 too early
with while (i-- > 0); i reaches -1 after the loop, so the test below is printed
one too early: 0 still means success.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-03-04 00:11:52 -08:00
Jens Axboe 9e973e64ac xen/blkfront: use blk_rq_map_sg to generate ring entries
On occasion, the request will apparently have more segments than we
fit into the ring. Jens says:

> The second problem is that the block layer then appears to create one
> too many segments, but from the dump it has rq->nr_phys_segments ==
> BLKIF_MAX_SEGMENTS_PER_REQUEST. I suspect the latter is due to
> xen-blkfront not handling the merging on its own. It should check that
> the new page doesn't form part of the previous page. The
> rq_for_each_segment() iterates all single bits in the request, not dma
> segments. The "easiest" way to do this is to call blk_rq_map_sg() and
> then iterate the mapped sg list. That will give you what you are
> looking for.

> Here's a test patch, compiles but otherwise untested. I spent more
> time figuring out how to enable XEN than to code it up, so YMMV!
> Probably the sg list wants to be put inside the ring and only
> initialized on allocation, then you can get rid of the sg on stack and
> sg_init_table() loop call in the function. I'll leave that, and the
> testing, to you.

[Moved sg array into info structure, and initialize once. -J]

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
2009-02-26 10:45:48 +01:00
Jens Axboe 5e4c91c84b cciss: shorten 30s timeout on controller reset
If reset_devices is set for kexec, then cciss will delay 30 seconds
since the old 5i controller _may_ need that long to recover. Replace
the long sleep with incremental sleep and tests to reduce the 30 seconds
to worst case for 5i, so that other controllers will proceed quickly.

Reviewed-by: Mike Miller <mike.miller@hp.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-02-26 10:45:48 +01:00
Geert Uytterhoeven 3d92e8f3ae m68k: atari - Rename "mfp" to "st_mfp"
http://kisskb.ellerman.id.au/kisskb/buildresult/72115/:
| net/mac80211/ieee80211_i.h:327: error: syntax error before 'volatile'
| net/mac80211/ieee80211_i.h:350: error: syntax error before '}' token
| net/mac80211/ieee80211_i.h:455: error: field 'sta' has incomplete type
| distcc[19430] ERROR: compile net/mac80211/main.c on sprygo/32 failed

This is caused by

| # define mfp ((*(volatile struct MFP*)MFP_BAS))

in arch/m68k/include/asm/atarihw.h, which conflicts with the new "mfp" enum in
net/mac80211/ieee80211_i.h.

Rename "mfp" to "st_mfp", as it's a way too generic name for a global #define.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-02-22 09:23:02 -08:00
Linus Torvalds ba95fd47d1 Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  block: fix deadlock in blk_abort_queue() for drivers that readd to timeout list
  block: fix booting from partitioned md array
  block: revert part of 18ce3751cc
  cciss: PCI power management reset for kexec
  paride/pg.c: xs(): &&/|| confusion
  fs/bio: bio_alloc_bioset: pass right object ptr to mempool_free
  block: fix bad definition of BIO_RW_SYNC
  bsg: Fix sense buffer bug in SG_IO
2009-02-18 18:33:04 -08:00
Philippe De Muyter 5a74db06cc floppy: request and release only the ports we actually use
The floppy driver requests an I/O port it doesn't need, and sometimes this
causes a conflict with a motherboard device reported by PNPBIOS.

This patch makes the floppy driver request and release only the ports it
actually uses.  It also factors out the request/release stuff and the
io-ports list so they're all in one place now.

The current floppy driver uses only these ports:

    0x3f2 (FD_DOR)
    0x3f4 (FD_STATUS)
    0x3f5 (FD_DATA)
    0x3f7 (FD_DCR/FD_DIR)

but it requests 0x3f2-0x3f5 and 0x3f7, which includes the unused port
0x3f3.

Some BIOSes report 0x3f3 as a motherboard resource.  The PNP system driver
reserves that, which causes a conflict when the floppy driver requests
0x3f2-0x3f5 later.

Philippe reported that this conflict broke the floppy driver between
2.6.11 and 2.6.22.  His PNPBIOS reports these devices:

    $ cat 00:07/id 00:07/resources	# motherboard device
    PNP0c02
    state = active
    io 0x80-0x80
    io 0x10-0x1f
    io 0x22-0x3f
    io 0x44-0x5f
    io 0x90-0x9f
    io 0xa2-0xbf
    io 0x3f0-0x3f1
    io 0x3f3-0x3f3

    $ cat 00:03/id 00:03/resources	# floppy device
    PNP0700
    state = active
    io 0x3f4-0x3f5
    io 0x3f2-0x3f2

Reference:
    http://lkml.org/lkml/2009/1/31/162

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Reported-by: Philippe De Muyter <phdm@macqel.be>
Tested-by: Philippe De Muyter <phdm@macqel.be>
Cc: Adam M Belay <abelay@mit.edu>
Cc: Robert Hancock <hancockrwd@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-02-18 15:37:55 -08:00
Ed Cashin b6d6c51758 aoe: ignore vendor extension AoE responses
The Welland ME-747K-SI AoE target generates unsolicited AoE responses that
are marked as vendor extensions.  Instead of ignoring these packets, the
aoe driver was generating kernel messages for each unrecognized response
received.  This patch corrects the behavior.

Signed-off-by: Ed Cashin <ecashin@coraid.com>
Reported-by: <karaluh@karaluh.pl>
Tested-by: <karaluh@karaluh.pl>
Cc: <stable@kernel.org>
Cc: Alex Buell <alex.buell@munted.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-02-18 15:37:53 -08:00
Chip Coldwell 82eb03cfd8 cciss: PCI power management reset for kexec
The kexec kernel resets the CCISS hardware in three steps:

1. Use PCI power management states to reset the controller in the
   kexec kernel.

2. Clear the MSI/MSI-X bits in PCI configuration space so that MSI
   initialization in the kexec kernel doesn't fail.

3. Use the CCISS "No-op" message to determine when the controller
   firmware has recovered from the PCI PM reset.

[akpm@linux-foundation.org: cleanups]
Signed-off-by: Mike Miller <mike.miller@hp.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-02-18 10:32:01 +01:00
Roel Kluin c8cbec6bdf paride/pg.c: xs(): &&/|| confusion
&&/|| confusion

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-02-18 10:32:01 +01:00
Paul Clements 4d48a542b4 nbd: fix I/O hang on disconnected nbds
Fix a problem that causes I/O to a disconnected (or partially initialized)
nbd device to hang indefinitely.  To reproduce:

# ioctl NBD_SET_SIZE_BLOCKS /dev/nbd23 514048
# dd if=/dev/nbd23 of=/dev/null bs=4096 count=1

...hangs...

This can also occur when an nbd device loses its nbd-client/server
connection.  Although we clear the queue of any outstanding I/Os after the
client/server connection fails, any additional I/Os that get queued later
will hang.

This bug may also be the problem reported in this bug report:
http://bugzilla.kernel.org/show_bug.cgi?id=12277

Testing would need to be performed to determine if the two issues are the
same.

This problem was introduced by the new request handling thread code ("NBD:
allow nbd to be used locally", 3/2008), which entered into mainline around
2.6.25.

The fix, which is fairly simple, is to restore the check for lo->sock
being NULL in do_nbd_request.  This causes I/O to an uninitialized nbd to
immediately fail with an I/O error, as it did prior to the introduction of
this bug.

Signed-off-by: Paul Clements <paul.clements@steeleye.com>
Reported-by: Jon Nelson <jnelson-kernel-bugzilla@jamponi.net>
Acked-by: Pavel Machek <pavel@ucw.cz>
Cc: <stable@kernel.org>		[2.6.26.x, 2.6.27.x, 2.6.28.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-02-11 14:25:37 -08:00
Stephen Rothwell e377c6e24d powerpc/ps3: Printing fixups for l64 to ll64 conversion drivers/block
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-16 16:15:13 +11:00
Pavel Machek c91192d66d nbd: do not allow two clients at the same time
Two nbd-clients at same time are bad idea, and cause WARN_ON from nbd in
2.6.28-rc7 from sysfs_add_one.  This simply prevents that from happening.

To reproduce:

 cat /dev/zero | head -c 10000000 > /tmp/delme.fstest.fs
 nbd-server 9100 -l /anyone.can.connect > /tmp/delme.fstest.fs &
 sleep 1
 nbd-client localhost 9100 /dev/nd0 &
 nbd-client localhost 9100 /dev/nd0 &

Signed-off-by: Pavel Machek <pavel@suse.cz>
Acked-by: Paul Clements <paul.clements@steeleye.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-15 16:39:38 -08:00
Benjamin Herrenschmidt bd1f7936ab Merge commit 'gcl/gcl-next' into next 2009-01-13 13:59:11 +11:00
Andreas Bombe 6d0be946e1 m68k: amiflop - Get rid of sleep_on calls
Apart from sleep_on() calls that could be easily converted to
wait_event() and completion calls amiflop also used a flag in ms_delay()
and ms_isr() as a custom mutex for ms_delay() without a need for
explicit unlocking.  I converted that to a standard mutex.

The replacement for the unconditional sleep_on() in fd_motor_on() is a
complete_all() together with a INIT_COMPLETION() before the mod_timer()
call.  It appears to me that fd_motor_on() might be called concurrently
and fd_select() does not guarantee mutual exclusivity in the case the
same drive gets selected again.

Signed-off-by: Andreas Bombe <aeb@debian.org>
Acked-by: Jörg Dorchain <joerg@dorchain.net>
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
2009-01-12 20:56:33 +01:00
Yuri Tikhonov f5020384e4 powerpc/xsysace: add compatible string for non-ipcore instance
Add "xlnx,sysace" compatible string to the of_platform binding
table.  Platforms which have the SysACE chip on board (e.g.
Katmai) instead of via a Xilinx generated IP core will use
this value in their device tree.

Signed-off-by: Yuri Tikhonov <yur@emcraft.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
2009-01-09 15:49:06 -07:00
Linus Torvalds 9e42d0cf50 Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Work around branch tracer warning.
  sparc64: Fix unsigned long long warnings in drivers.
  sparc64: Use unsigned long long for u64.
  sparc: refactor code in fault_32.c
  sparc64: refactor code in init_64.c
  sparc64: refactor code in viohs.c
  sparc: make proces_ver_nack a bit more readable
2009-01-07 17:23:53 -08:00
Alan Stern 011b15df46 USB: change interface to usb_lock_device_for_reset()
This patch (as1161) changes the interface to
usb_lock_device_for_reset().  The existing interface is apparently not
very clear, judging from the fact that several of its callers don't
use it correctly.  The new interface always returns 0 for success and
it always requires the caller to unlock the device afterward.

The new routine will not return immediately if it is called while the
driver's probe method is running.  Instead it will wait until the
probe is over and the device has been unlocked.  This shouldn't cause
any problems; I don't know of any cases where drivers call
usb_lock_device_for_reset() during probe.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-01-07 09:59:52 -08:00
Sam Ravnborg 3f4528d6e9 sparc64: Fix unsigned long long warnings in drivers.
Fix warnings caused by the unsigned long long usage in sparc
specific drivers.

The drivers were considered sparc specific more or less from the
filename alone.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-06 13:20:38 -08:00
Linus Torvalds ab70537c32 Merge git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus
* git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus:
  lguest: struct device - replace bus_id with dev_name()
  lguest: move the initial guest page table creation code to the host
  kvm-s390: implement config_changed for virtio on s390
  virtio_console: support console resizing
  virtio: add PCI device release() function
  virtio_blk: fix type warning
  virtio: block: dynamic maximum segments
  virtio: set max_segment_size and max_sectors to infinite.
  virtio: avoid implicit use of Linux page size in balloon interface
  virtio: hand virtio ring alignment as argument to vring_new_virtqueue
  virtio: use KVM_S390_VIRTIO_RING_ALIGN instead of relying on pagesize
  virtio: use LGUEST_VRING_ALIGN instead of relying on pagesize
  virtio: Don't use PAGE_SIZE for vring alignment in virtio_pci.
  virtio: rename 'pagesize' arg to vring_init/vring_size
  virtio: Don't use PAGE_SIZE in virtio_pci.c
  virtio: struct device - replace bus_id with dev_name(), dev_set_name()
  virtio-pci queue allocation not page-aligned
2008-12-30 17:37:25 -08:00
Randy Dunlap b194aee956 virtio_blk: fix type warning
Fix parameter type warning:

linux-next-20081126/drivers/block/virtio_blk.c:307: warning: large integer implicitly truncated to unsigned type

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2008-12-30 09:26:06 +10:30
Rusty Russell 0864b79a15 virtio: block: dynamic maximum segments
Enhance the driver to handle whatever maximum segment number the host
tells us to handle.  Do to this, we need to allocate the scatterlist
dynamically.

We set max_phys_segments and max_hw_segments to the same value (1 if
the host doesn't tell us, since that's safest and all known hosts do
tell us).

Note that kmalloc'ing the structure for large sg_elems might be
problematic: the fix for this is sg_table, but that requires more
work.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2008-12-30 09:26:05 +10:30
Rusty Russell 4b7f7e2049 virtio: set max_segment_size and max_sectors to infinite.
Setting max_segment_size allows more than 64k per sg element, unless
the host specified a limit.  Setting max_sectors indicates that our
max_hw_segments is the only limit.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2008-12-30 09:26:05 +10:30
Stephen M. Cameron a0ea862291 cciss: simplify parameters to deregister_disk function
Simplify parameters to deregister_disk function.

Signed-off-by: Stephen M. Cameron <scameron@beardog.cca.cpqcorp.net>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:29:52 +01:00
Milan Broz 8ae30b8958 loop: Do not call loop_unplug for not configured loop device.
In loop_unplug() function is expected that mapping is set
and lo->lo_backing_file is not NULL.

Unfortunately loop_set_fd() set the request queue unplug function,
but loop_clr_fd() doesn't clear that.

Loop device allows open of non-configured loop in some situations.
If the unplug on request queue is called, loop module oopses because
of missing lo_backing_file.

Simple reproducer:
	losetup /dev/loop0 /xxx
	losetup -d /dev/loop0
	dmsetup create x --table "0 1 linear /dev/loop0 0"

 EIP is at loop_unplug+0x1d/0x3b
 ...
  Call Trace:
   blk_unplug+0x57/0x5e
   dm_table_unplug_all+0x34/0x77 [dm_mod]
   destroy_inode+0x27/0x38
   generic_delete_inode+0xd5/0xd9
   iput+0x4b/0x4e
   dm_resume+0xca/0xfe [dm_mod]
   dev_suspend+0x143/0x165 [dm_mod]
   dm_ctl_ioctl+0x18e/0x1cf [dm_mod]
   dev_suspend+0x0/0x165 [dm_mod]
   dm_ctl_ioctl+0x0/0x1cf [dm_mod]
   vfs_ioctl+0x22/0x69
   do_vfs_ioctl+0x39d/0x3c7
   trace_hardirqs_on+0xb/0xd
   remove_vma+0x50/0x56
   do_munmap+0x21c/0x237
   sys_ioctl+0x2c/0x45
   sysenter_do_call+0x12/0x31

Several reports here
http://www.kerneloops.org/search.php?search=loop_unplug

Fix it by simply clear unplug function together with
removing of backing file.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:29:52 +01:00
Milan Broz 14f2793958 loop: Flush possible running bios when loop device is released.
When there are still queued bios and reference count
drops to zero, loop device must flush all queued bios.

Otherwise it can lead to situation that caller
closes the device, but some bios are still running
and endio() function call later OOpses when uses
unallocated mempool.

This happens for example when running dm-crypt over loop,
here is typical oops backtrace:

 Oops: 0000 [#1] PREEMPT SMP
 EIP is at mempool_free+0x12/0x6b
...
 crypt_dec_pending+0x50/0x54 [dm_crypt]
 crypt_endio+0x9f/0xa7 [dm_crypt]
 crypt_endio+0x0/0xa7 [dm_crypt]
 bio_endio+0x2b/0x2e
 loop_thread+0x37a/0x3b1
 do_lo_send_aops+0x0/0x165
 autoremove_wake_function+0x0/0x33
 loop_thread+0x0/0x3b1
 kthread+0x3b/0x61
 kthread+0x0/0x61
 kernel_thread_helper+0x7/0x10

(But crash is reproducible with different dm targets
running over loop device too.)

Patch fixes it by flushing the bios in release call,
reusing the flush mechanism for switching backing store.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:29:52 +01:00
Jens Axboe 31dcfab0ae nbd: tell the block layer that it is not a rotational device
Then we can get rid of that manual elevator type fiddling.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:29:50 +01:00
Jens Axboe b374d18a4b block: get rid of elevator_t typedef
Just use struct elevator_queue everywhere instead.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:29:50 +01:00
Jens Axboe 8a3173de4a cciss: switch to using hlist for command list management
This both cleans up the code and also helps detect the spurious case
of a command attempted being removed from a queue it doesn't belong
to.

Acked-by: Mike Miller <mike.miller@hp.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:28:43 +01:00
Fernando Luis Vázquez Cao 66d352e1e4 xen-blkfront: set queue paravirt flag
Xen's blkfront sets noop as the default I/O scheduler at initialization
time to avoid elevator overheads such as idling, but with the advent of
basic disk profiling capabilities this is not necessary anymore. We
should just tell the block layer that we are a paravirt front-end driver
and the elevator will automatically make the necessary adjustments.

Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-12-29 08:28:41 +01:00