safe/jmp/linux-2.6
14 years agonetfilter: xtables: compat out of scope fix
Alexey Dobriyan [Mon, 8 Feb 2010 14:44:07 +0000 (15:44 +0100)]
netfilter: xtables: compat out of scope fix

As per C99 6.2.4(2) when temporary table data goes out of scope,
the behaviour is undefined:

if (compat) {
struct foo tmp;
...
private = &tmp;
}
[dereference private]

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: nf_conntrack: restrict runtime expect hashsize modifications
Alexey Dobriyan [Thu, 4 Feb 2010 17:24:06 +0000 (18:24 +0100)]
netfilter: nf_conntrack: restrict runtime expect hashsize modifications

Expectation hashtable size was simply glued to a variable with no code
to rehash expectations, so it was a bug to allow writing to it.
Make "expect_hashsize" readonly.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: nf_conntrack: per netns nf_conntrack_cachep
Eric Dumazet [Thu, 4 Feb 2010 13:54:05 +0000 (14:54 +0100)]
netfilter: nf_conntrack: per netns nf_conntrack_cachep

nf_conntrack_cachep is currently shared by all netns instances, but
because of SLAB_DESTROY_BY_RCU special semantics, this is wrong.

If we use a shared slab cache, one object can instantly flight between
one hash table (netns ONE) to another one (netns TWO), and concurrent
reader (doing a lookup in netns ONE, 'finding' an object of netns TWO)
can be fooled without notice, because no RCU grace period has to be
observed between object freeing and its reuse.

We dont have this problem with UDP/TCP slab caches because TCP/UDP
hashtables are global to the machine (and each object has a pointer to
its netns).

If we use per netns conntrack hash tables, we also *must* use per netns
conntrack slab caches, to guarantee an object can not escape from one
namespace to another one.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
[Patrick: added unique slab name allocation]
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: nf_conntrack: fix memory corruption with multiple namespaces
Patrick McHardy [Wed, 3 Feb 2010 11:58:06 +0000 (12:58 +0100)]
netfilter: nf_conntrack: fix memory corruption with multiple namespaces

As discovered by Jon Masters <jonathan@jonmasters.org>, the "untracked"
conntrack, which is located in the data section, might be accidentally
freed when a new namespace is instantiated while the untracked conntrack
is attached to a skb because the reference count it re-initialized.

The best fix would be to use a seperate untracked conntrack per
namespace since it includes a namespace pointer. Unfortunately this is
not possible without larger changes since the namespace is not easily
available everywhere we need it. For now move the untracked conntrack
initialization to the init_net setup function to make sure the reference
count is not re-initialized and handle cleanup in the init_net cleanup
function to make sure namespaces can exit properly while the untracked
conntrack is in use in other namespaces.

Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: ctnetlink: fix expectation mask dump
Patrick McHardy [Tue, 26 Jan 2010 16:04:02 +0000 (17:04 +0100)]
netfilter: ctnetlink: fix expectation mask dump

The protocol number is not initialized, so userspace can't interpret
the layer 4 data properly.

Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: nf_conntrack_sip: fix off-by-one in compact header parsing
Patrick McHardy [Tue, 19 Jan 2010 18:06:59 +0000 (19:06 +0100)]
netfilter: nf_conntrack_sip: fix off-by-one in compact header parsing

In a string like "v:SIP/2.0..." it was checking for !isalpha('S') when it
meant to be inspecting the ':'.

Patch by Greg Alexander <greqcs@galexander.org>

Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: ebtables: enforce CAP_NET_ADMIN
Florian Westphal [Fri, 8 Jan 2010 16:31:24 +0000 (17:31 +0100)]
netfilter: ebtables: enforce CAP_NET_ADMIN

normal users are currently allowed to set/modify ebtables rules.
Restrict it to processes with CAP_NET_ADMIN.

Note that this cannot be reproduced with unmodified ebtables binary
because it uses SOCK_RAW.

Signed-off-by: Florian Westphal <fwestphal@astaro.com>
Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: nf_ct_ftp: fix out of bounds read in update_nl_seq()
Patrick McHardy [Thu, 7 Jan 2010 17:33:18 +0000 (18:33 +0100)]
netfilter: nf_ct_ftp: fix out of bounds read in update_nl_seq()

As noticed by Dan Carpenter <error27@gmail.com>, update_nl_seq()
currently contains an out of bounds read of the seq_aft_nl array
when looking for the oldest sequence number position.

Fix it to only compare valid positions.

Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoipvs: Add boundary check on ioctl arguments
Arjan van de Ven [Mon, 4 Jan 2010 15:37:12 +0000 (16:37 +0100)]
ipvs: Add boundary check on ioctl arguments

The ipvs code has a nifty system for doing the size of ioctl command
copies; it defines an array with values into which it indexes the cmd
to find the right length.

Unfortunately, the ipvs code forgot to check if the cmd was in the
range that the array provides, allowing for an index outside of the
array, which then gives a "garbage" result into the length, which
then gets used for copying into a stack buffer.

Fix this by adding sanity checks on these as well as the copy size.

[ horms@verge.net.au: adjusted limit to IP_VS_SO_GET_MAX ]
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoipvs: ip_vs_wrr.c: use lib/gcd.c
Florian Fainelli [Tue, 22 Dec 2009 08:42:06 +0000 (09:42 +0100)]
ipvs: ip_vs_wrr.c: use lib/gcd.c

Remove the private version of the greatest common divider to use
lib/gcd.c, the latter also implementing the a < b case.

[akpm@linux-foundation.org: repair neighboring whitespace because the diff looked odd]
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Cc: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Cc: Takashi Iwai <tiwai@suse.de>
Acked-by: Simon Horman <horms@verge.net.au>
Cc: Julius Volz <juliusv@google.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoipvs: zero usvc and udest
Simon Horman [Tue, 15 Dec 2009 16:01:25 +0000 (17:01 +0100)]
ipvs: zero usvc and udest

Make sure that any otherwise uninitialised fields of usvc are zero.

This has been obvserved to cause a problem whereby the port of
fwmark services may end up as a non-zero value which causes
scheduling of a destination server to fail for persisitent services.

As observed by Deon van der Merwe <dvdm@truteq.co.za>.
This fix suggested by Julian Anastasov <ja@ssi.bg>.

For good measure also zero udest.

Cc: Deon van der Merwe <dvdm@truteq.co.za>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: fix crashes in bridge netfilter caused by fragment jumps
Patrick McHardy [Tue, 15 Dec 2009 15:59:59 +0000 (16:59 +0100)]
netfilter: fix crashes in bridge netfilter caused by fragment jumps

When fragments from bridge netfilter are passed to IPv4 or IPv6 conntrack
and a reassembly queue with the same fragment key already exists from
reassembling a similar packet received on a different device (f.i. with
multicasted fragments), the reassembled packet might continue on a different
codepath than where the head fragment originated. This can cause crashes
in bridge netfilter when a fragment received on a non-bridge device (and
thus with skb->nf_bridge == NULL) continues through the bridge netfilter
code.

Add a new reassembly identifier for packets originating from bridge
netfilter and use it to put those packets in insolated queues.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14805

Reported-and-Tested-by: Chong Qiao <qiaochong@loongson.cn>
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoipv6: reassembly: use seperate reassembly queues for conntrack and local delivery
Patrick McHardy [Tue, 15 Dec 2009 15:59:18 +0000 (16:59 +0100)]
ipv6: reassembly: use seperate reassembly queues for conntrack and local delivery

Currently the same reassembly queue might be used for packets reassembled
by conntrack in different positions in the stack (PREROUTING/LOCAL_OUT),
as well as local delivery. This can cause "packet jumps" when the fragment
completing a reassembled packet is queued from a different position in the
stack than the previous ones.

Add a "user" identifier to the reassembly queue key to seperate the queues
of each caller, similar to what we do for IPv4.

Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoipvs: fix synchronization on connection close
Xiaotian Feng [Mon, 14 Dec 2009 15:38:21 +0000 (16:38 +0100)]
ipvs: fix synchronization on connection close

commit 9d3a0de makes slaves expire as they would do on the master
with much shorter timeouts. But it introduces another problem:
When we close a connection, on master server the connection became
CLOSE_WAIT/TIME_WAIT, it was synced to slaves, but if master is
finished within it's timeouts (CLOSE), it will not be synced to
slaves. Then slaves will be kept on CLOSE_WAIT/TIME_WAIT until
timeout reaches. Thus we should also sync with CLOSE.

Cc: Wensong Zhang <wensong@linux-vs.org>
Cc: Simon Horman <horms@verge.net.au>
Cc: Julian Anastasov <ja@ssi.bg>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Xiaotian Feng <dfeng@redhat.com>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agonetfilter: xtables: document minimal required version
Jan Engelhardt [Mon, 14 Dec 2009 13:52:10 +0000 (14:52 +0100)]
netfilter: xtables: document minimal required version

For both .33 and .32-stable.

Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Cc: stable@kernel.org
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Fri, 11 Dec 2009 23:59:23 +0000 (15:59 -0800)]
Merge branch 'next' of git://git./linux/kernel/git/davej/cpufreq

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq:
  [ACPI/CPUFREQ] Introduce bios_limit per cpu cpufreq sysfs interface
  [CPUFREQ] make internal cpufreq_add_dev_* static
  [CPUFREQ] use an enum for speedstep processor identification
  [CPUFREQ] Document units for transition latency
  [CPUFREQ] Use global sysfs cpufreq structure for conservative governor tunings
  [CPUFREQ] Documentation: ABI: /sys/devices/system/cpu/cpu#/cpufreq/
  [CPUFREQ] powernow-k6: set transition latency value so ondemand governor can be used
  [CPUFREQ] cpumask: don't put a cpumask on the stack in x86...cpufreq/powernow-k8.c

14 years agoStaging: batman: fix debug Kconfig option
Greg Kroah-Hartman [Fri, 11 Dec 2009 23:40:17 +0000 (15:40 -0800)]
Staging: batman: fix debug Kconfig option

The debug batman option needs to depend on the correct
config option.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[ "No means no!"  - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6
Linus Torvalds [Fri, 11 Dec 2009 23:34:40 +0000 (15:34 -0800)]
Merge git://git./linux/kernel/git/gregkh/tty-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6: (58 commits)
  tty: split the lock up a bit further
  tty: Move the leader test in disassociate
  tty: Push the bkl down a bit in the hangup code
  tty: Push the lock down further into the ldisc code
  tty: push the BKL down into the handlers a bit
  tty: moxa: split open lock
  tty: moxa: Kill the use of lock_kernel
  tty: moxa: Fix modem op locking
  tty: moxa: Kill off the throttle method
  tty: moxa: Locking clean up
  tty: moxa: rework the locking a bit
  tty: moxa: Use more tty_port ops
  tty: isicom: fix deadlock on shutdown
  tty: mxser: Use the new locking rules to fix setserial properly
  tty: mxser: use the tty_port_open method
  tty: isicom: sort out the board init logic
  tty: isicom: switch to the new tty_port_open helper
  tty: tty_port: Add a kref object to the tty port
  tty: istallion: tty port open/close methods
  tty: stallion: Convert to the tty_port_open/close methods
  ...

14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux...
Linus Torvalds [Fri, 11 Dec 2009 23:31:13 +0000 (15:31 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs-2.6

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6: (21 commits)
  ext3: PTR_ERR return of wrong pointer in setup_new_group_blocks()
  ext3: Fix data / filesystem corruption when write fails to copy data
  ext4: Support for 64-bit quota format
  ext3: Support for vfsv1 quota format
  quota: Implement quota format with 64-bit space and inode limits
  quota: Move definition of QFMT_OCFS2 to linux/quota.h
  ext2: fix comment in ext2_find_entry about return values
  ext3: Unify log messages in ext3
  ext2: clear uptodate flag on super block I/O error
  ext2: Unify log messages in ext2
  ext3: make "norecovery" an alias for "noload"
  ext3: Don't update the superblock in ext3_statfs()
  ext3: journal all modifications in ext3_xattr_set_handle
  ext2: Explicitly assign values to on-disk enum of filetypes
  quota: Fix WARN_ON in lookup_one_len
  const: struct quota_format_ops
  ubifs: remove manual O_SYNC handling
  afs: remove manual O_SYNC handling
  kill wait_on_page_writeback_range
  vfs: Implement proper O_SYNC semantics
  ...

14 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Fri, 11 Dec 2009 23:30:29 +0000 (15:30 -0800)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: Fix error return for fallocate() on XFS
  xfs: cleanup dmapi macros in the umount path
  xfs: remove incorrect sparse annotation for xfs_iget_cache_miss
  xfs: kill the STATIC_INLINE macro
  xfs: uninline xfs_get_extsz_hint
  xfs: rename xfs_attr_fetch to xfs_attr_get_int
  xfs: simplify xfs_buf_get / xfs_buf_read interfaces
  xfs: remove IO_ISAIO
  xfs: Wrapped journal record corruption on read at recovery
  xfs: cleanup data end I/O handlers
  xfs: use WRITE_SYNC_PLUG for synchronous writeout
  xfs: reset the i_iolock lock class in the reclaim path
  xfs: I/O completion handlers must use NOFS allocations
  xfs: fix mmap_sem/iolock inversion in xfs_free_eofblocks
  xfs: simplify inode teardown

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6
Linus Torvalds [Fri, 11 Dec 2009 23:25:56 +0000 (15:25 -0800)]
Merge git://git./linux/kernel/git/gregkh/staging-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (235 commits)
  Staging: IIO: add selection of IIO_SW_RING to LIS3L02DQ as needed
  Staging: IIO: Add tsl2560-2 support to tsl2563 driver.
  Staging: IIO: Remove tsl2561 driver. Support merged with tsl2563.
  Staging: wlags49_h2: fix up signal levels
  + drivers-staging-wlags49_h2-remove-cvs-metadata.patch added to -mm tree
  Staging: samsung-laptop: add TODO file
  Staging: samsung-laptop: remove old kernel code
  Staging: add Samsung Laptop driver
  staging: batman-adv meshing protocol
  Staging: rtl8192u: depends on USB
  Staging: rtl8192u: remove dead code
  Staging: rtl8192u: remove bad whitespaces
  Staging: rtl8192u: make it compile
  Staging: Added Realtek rtl8192u driver to staging
  Staging: dream: add gpio and pmem support
  Staging: dream: add TODO file
  Staging: android: delete android drivers
  Staging: et131x: clean up the avail fields in the rx registers
  Staging: et131x: Clean up number fields
  Staging: et131x: kill RX_DMA_MAX_PKT_TIME
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6
Linus Torvalds [Fri, 11 Dec 2009 23:24:56 +0000 (15:24 -0800)]
Merge git://git./linux/kernel/git/gregkh/driver-core-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6: (27 commits)
  Driver core: fix race in dev_driver_string
  Driver Core: Early platform driver buffer
  sysfs: sysfs_setattr remove unnecessary permission check.
  sysfs: Factor out sysfs_rename from sysfs_rename_dir and sysfs_move_dir
  sysfs: Propagate renames to the vfs on demand
  sysfs: Gut sysfs_addrm_start and sysfs_addrm_finish
  sysfs: In sysfs_chmod_file lazily propagate the mode change.
  sysfs: Implement sysfs_getattr & sysfs_permission
  sysfs: Nicely indent sysfs_symlink_inode_operations
  sysfs: Update s_iattr on link and unlink.
  sysfs: Fix locking and factor out sysfs_sd_setattr
  sysfs: Simplify iattr time assignments
  sysfs: Simplify sysfs_chmod_file semantics
  sysfs: Use dentry_ops instead of directly playing with the dcache
  sysfs: Rename sysfs_d_iput to sysfs_dentry_iput
  sysfs: Update sysfs_setxattr so it updates secdata under the sysfs_mutex
  debugfs: fix create mutex racy fops and private data
  Driver core: Don't remove kobjects in device_shutdown.
  firmware_class: make request_firmware_nowait more useful
  Driver-Core: devtmpfs - set root directory mode to 0755
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Fri, 11 Dec 2009 23:22:55 +0000 (15:22 -0800)]
Merge git://git./linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (122 commits)
  USB: mos7840: add device IDs for B&B electronics devices
  USB: ftdi_sio: add USB device ID's for B&B Electronics line
  USB: musb: musb_host: fix sparse warning
  USB: musb: musb_gadget: fix sparse warning
  USB: musb: omap2430: fix sparse warning
  USB: core: message: fix sparse warning
  USB: core: hub: fix sparse warning
  USB: core: fix sparse warning for static function
  USB: Added USB_ETH_RNDIS to use instead of CONFIG_USB_ETH_RNDIS
  USB: Check bandwidth when switching alt settings.
  USB: Refactor code to find alternate interface settings.
  USB: xhci: Fix command completion after a drop endpoint.
  USB: xhci: Make reverting an alt setting "unfailable".
  USB: usbtmc: Use usb_clear_halt() instead of custom code.
  USB: xhci: Add correct email and files to MAINTAINERS entry.
  USB: ehci-omap.c: introduce missing kfree
  USB: xhci-mem.c: introduce missing kfree
  USB: add remove_id sysfs attr for usb drivers
  USB: g_multi kconfig: fix depends and help text
  USB: option: add pid for ZTE
  ...

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Fri, 11 Dec 2009 23:22:27 +0000 (15:22 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  firewire: ohci: handle receive packets with a data length of zero

14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel...
Linus Torvalds [Fri, 11 Dec 2009 23:19:56 +0000 (15:19 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jwessel/linux-2.6-kgdb

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/linux-2.6-kgdb:
  kgdb: Always process the whole breakpoint list on activate or deactivate
  kgdb: continue and warn on signal passing from gdb
  kgdb,x86: do not set kgdb_single_step on x86
  kgdb: allow for cpu switch when single stepping
  kgdb,i386: Fix corner case access to ss with NMI watch dog exception
  kgdb: Replace strstr() by strchr() for single-character needles
  kgdbts: Read buffer overflow
  kgdb: Read buffer overflow
  kgdb,x86: remove redundant test

14 years agotty: split the lock up a bit further
Alan Cox [Mon, 30 Nov 2009 13:18:51 +0000 (13:18 +0000)]
tty: split the lock up a bit further

The tty count sanity check may need the BKL, that isn't clear. However it
is clear that the count use of the lock is internal and independant of the
bigger use of the lock.

Furthermore the file list locking is also separately locked already

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: Move the leader test in disassociate
Alan Cox [Mon, 30 Nov 2009 13:18:45 +0000 (13:18 +0000)]
tty: Move the leader test in disassociate

There are two call points, both want to check that tty->signal->leader is
set. Move the test into disassociate_ctty() as that will make locking
changes easier in a bit

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: Push the bkl down a bit in the hangup code
Alan Cox [Mon, 30 Nov 2009 13:18:40 +0000 (13:18 +0000)]
tty: Push the bkl down a bit in the hangup code

We know that the redirect field is handled via its own locking in all
places

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: Push the lock down further into the ldisc code
Alan Cox [Mon, 30 Nov 2009 13:18:35 +0000 (13:18 +0000)]
tty: Push the lock down further into the ldisc code

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: push the BKL down into the handlers a bit
Alan Cox [Mon, 30 Nov 2009 13:18:29 +0000 (13:18 +0000)]
tty: push the BKL down into the handlers a bit

Start trying to untangle the remaining BKL mess

Updated to fix missing unlock_kernel noted by Dan Carpenter

Signed-off-by: Alan "I must be out of my tree" Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: split open lock
Alan Cox [Mon, 30 Nov 2009 13:18:24 +0000 (13:18 +0000)]
tty: moxa: split open lock

moxa_openlock is used for several situations where we want to handle the
case of an ioctl that crosses many ports (not just the open tty), and also
cases where an open races a deinit (eg a pci unplug) and we hangup a port
before we can cope with that.

The non open race cases can use the moxa_lock spinlock. This simplifies sorting
out the remaining mess.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: Kill the use of lock_kernel
Alan Cox [Mon, 30 Nov 2009 13:18:18 +0000 (13:18 +0000)]
tty: moxa: Kill the use of lock_kernel

It isn't needed here any more

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: Fix modem op locking
Alan Cox [Mon, 30 Nov 2009 13:18:13 +0000 (13:18 +0000)]
tty: moxa: Fix modem op locking

This is overkill and mostly not needed

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: Kill off the throttle method
Alan Cox [Mon, 30 Nov 2009 13:18:08 +0000 (13:18 +0000)]
tty: moxa: Kill off the throttle method

The tty flag can be tested so the shadow flag isn't needed

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: Locking clean up
Alan Cox [Mon, 30 Nov 2009 13:18:02 +0000 (13:18 +0000)]
tty: moxa: Locking clean up

- The open lock is needed to fix up the case of a board reset occuring during
  tty open but too early for a sane hangup response.
- The lock can however got for other cases
- Use the port mutex for get/setserial
- Fix up the confused lack of locking on the THROTTLE and other bits in the
  private flags. Just use set/test/clear bit and it covers the cases we need

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: rework the locking a bit
Alan Cox [Mon, 30 Nov 2009 13:17:57 +0000 (13:17 +0000)]
tty: moxa: rework the locking a bit

Introduce a lock for moxafunc() to protect the cases where were get collisions
between two function requests at the same time.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: moxa: Use more tty_port ops
Alan Cox [Mon, 30 Nov 2009 13:17:51 +0000 (13:17 +0000)]
tty: moxa: Use more tty_port ops

Rework a few bits of this into tty_port format

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: isicom: fix deadlock on shutdown
Alan Cox [Mon, 30 Nov 2009 13:17:46 +0000 (13:17 +0000)]
tty: isicom: fix deadlock on shutdown

Alexander Strakh <strakh@ispras.ru> reported

KERNEL_VERSION: 2.6.31
DESCRIBE:
Driver drivers/char/isicom.c might sleep in atomic  context, because it
calls
tty_port_xmit_buf under spin_lock.

./drivers/char/isicom.c:
1307 static void isicom_hangup(struct tty_struct *tty)
1308 {
...
1315         spin_lock_irqsave(&port->card->card_lock, flags);
1316         isicom_shutdown_port(port);
...

Path to might_sleep macro from isicom_hangup:
1. isicom_hangup calls spin_lock_irqsave (drivers/char/isicom.c:1315) and
then
calls isicom_shutdown_port.
2. isiscom_shutdown_port calls tty_port_free_xmit_buf at
drivers/char/isicom.c:906
3. tty_port_free_xmit_buf calls mutex_lock at drivers/char/tty_port:48

Found by Linux Driver Verification Project.

Reported-by: Alexander Strakh <strakh@ispras.ru>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: mxser: Use the new locking rules to fix setserial properly
Alan Cox [Mon, 30 Nov 2009 13:17:41 +0000 (13:17 +0000)]
tty: mxser: Use the new locking rules to fix setserial properly

Propogate the init/shutdown mutex through the setserial logic. Use the proper
locks for the various bits still using the BKL. Kill the BKL in this driver.

Updated to fix the bug noted by Dan Carpenter

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: mxser: use the tty_port_open method
Alan Cox [Mon, 30 Nov 2009 13:17:35 +0000 (13:17 +0000)]
tty: mxser: use the tty_port_open method

At first this looks a fairly trivial conversion but we can't quite push
everything into the right format yet. The open side is easy but care is needed
over the setserial methods. Fix up the locking now that we've adopted the
port->mutex locking rule for the initialization.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: isicom: sort out the board init logic
Alan Cox [Mon, 30 Nov 2009 13:17:30 +0000 (13:17 +0000)]
tty: isicom: sort out the board init logic

Split this into two flags - INIT meaning the board is set up and ACTIVE
meaning the board has ports open. Remove the broken HUPCL casing and push
the counts somewhere sensible.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: isicom: switch to the new tty_port_open helper
Alan Cox [Mon, 30 Nov 2009 13:17:24 +0000 (13:17 +0000)]
tty: isicom: switch to the new tty_port_open helper

Trivial conversion in this case so might as well do it while testing the
port_open design is right

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: tty_port: Add a kref object to the tty port
Alan Cox [Mon, 30 Nov 2009 13:17:14 +0000 (13:17 +0000)]
tty: tty_port: Add a kref object to the tty port

Users of tty port need a way to refcount ports when hotplugging is
involved.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: istallion: tty port open/close methods
Alan Cox [Mon, 30 Nov 2009 13:17:08 +0000 (13:17 +0000)]
tty: istallion: tty port open/close methods

Slice/dice/repeat as with the stallion driver this is just code shuffling
and removal

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: stallion: Convert to the tty_port_open/close methods
Alan Cox [Mon, 30 Nov 2009 13:17:03 +0000 (13:17 +0000)]
tty: stallion: Convert to the tty_port_open/close methods

The driver is already structured this way so just slice and dice

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: tty_port: Move the IO_ERROR clear
Alan Cox [Mon, 30 Nov 2009 13:16:57 +0000 (13:16 +0000)]
tty: tty_port: Move the IO_ERROR clear

Some devices want to set IO_ERROR in their activate methods so that you can
be handed a 'dead' port for operations like setserial. Thus we need to
clear the flag before activate so that activate can choose to set the flag
and still return 0.

This is fine as the file handle/tty are not accessible to the user yet.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: tty_port: Add IO_ERROR bit handling
Alan Cox [Mon, 30 Nov 2009 13:16:52 +0000 (13:16 +0000)]
tty: tty_port: Add IO_ERROR bit handling

To propogate tty_port_open/close to a few other devices we need to start
handling the IO_ERROR flag on the tty. We can do this pretty trivially.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: riscom8: switch to the tty_port_open API
Alan Cox [Mon, 30 Nov 2009 13:16:47 +0000 (13:16 +0000)]
tty: riscom8: switch to the tty_port_open API

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: tty_port: Change the buffer allocator locking
Alan Cox [Mon, 30 Nov 2009 13:16:41 +0000 (13:16 +0000)]
tty: tty_port: Change the buffer allocator locking

We want to be able to do this without regard for the activate/own open
method being used which causes a problem using port->mutex. Add another
mutex for now. Once everything uses port_open to do buffer allocs we can
kill it back off

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: sdio_uart: Fix the locking on "func" for new code
Alan Cox [Mon, 30 Nov 2009 13:16:30 +0000 (13:16 +0000)]
tty: sdio_uart: Fix the locking on "func" for new code

The new dtr_rts function didn't take the port->func lock as it should
so add use of the lock there.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: sdio_uart: add modem functionality
Alan Cox [Mon, 30 Nov 2009 13:16:36 +0000 (13:16 +0000)]
tty: sdio_uart: add modem functionality

Add the POSIX block for carrier

Linux TIOCMIWAIT functionality is still lacking from the driver.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: sdio_uart: Style fixes
Alan Cox [Mon, 30 Nov 2009 13:16:25 +0000 (13:16 +0000)]
tty: sdio_uart: Style fixes

Running the current code through checkpatch shows a few bits of noise
mostly but not entirely from before the changes.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: sdio_uart: Fix termios handling
Alan Cox [Mon, 30 Nov 2009 13:16:14 +0000 (13:16 +0000)]
tty: sdio_uart: Fix termios handling

Switching between two non standard baud rates fails because of the cflag
test. Do as we did elsewhere and just kill the "optimisation".

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: sdio_uart: Switch to the open/close helpers
Alan Cox [Mon, 30 Nov 2009 13:16:09 +0000 (13:16 +0000)]
tty: sdio_uart: Switch to the open/close helpers

Gets us proper tty semantics, removes some code and fixes up a few corner
case races (hangup during open etc)

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosdio_uart: Move the open lock
Alan Cox [Thu, 5 Nov 2009 13:28:38 +0000 (13:28 +0000)]
sdio_uart: Move the open lock

When we move to the tty_port logic the port mutex will protect open v close
v hangup. Move to this first in the existing open code so we have a bisection
point.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosdio_uart: refcount the tty objects
Alan Cox [Thu, 5 Nov 2009 13:28:29 +0000 (13:28 +0000)]
sdio_uart: refcount the tty objects

The tty can go away underneath us, so we must refcount it. Do the naïve
implementation initially. We will worry about startup shortly.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosdio_uart: Fix oops caused by the previous changeset
Nicolas Pitre [Thu, 5 Nov 2009 13:28:17 +0000 (13:28 +0000)]
sdio_uart: Fix oops caused by the previous changeset

Now... testing reveals that the very first patch "sdio_uart: use
tty_port" causes a segmentation fault in sdio_uart_open():

Unable to handle kernel NULL pointer dereference at virtual address 00000084
pgd = dfb44000 [00000084] *pgd=1fb99031, *pte=00000000, *ppte=00000000
Internal error: Oops: 17 [#1] PREEMPT
last sysfs file:
/sys/devices/platform/mvsdio/mmc_host/mmc0/mmc0:f111/uevent
Modules linked in:
CPU: 0    Not tainted  (2.6.32-rc5-next-20091102-00001-gb36eae9 #10)
PC is at sdio_uart_open+0x204/0x2cc
[...]

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosdio_uart: use tty_port
Alan Cox [Thu, 5 Nov 2009 13:28:06 +0000 (13:28 +0000)]
sdio_uart: use tty_port

Add a tty_port object to the sdio uart. For the moment just begin using the
tty field of the port, as this is the critical one to clean up.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty_port: Move hupcl handling
Alan Cox [Thu, 5 Nov 2009 13:27:57 +0000 (13:27 +0000)]
tty_port: Move hupcl handling

Move the HUCPL handling from the end of close_port_start to the beginning
of close_port_end. What this actually does is change the ordering from

port shutdown
port->dtr_rts

to

port->dtr_rts
port shutdown

Some hardware drops the physical connection on shutdown so we must perform
the port operations before the shutdown.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodevpts_get_tty() should validate inode
Sukadev Bhattiprolu [Wed, 18 Nov 2009 02:35:43 +0000 (18:35 -0800)]
devpts_get_tty() should validate inode

devpts_get_tty() assumes that the inode passed in is associated with a valid
pty.  But if the only reference to the pty is via a bind-mount, the inode
passed to devpts_get_tty() while valid, would refer to a pty that no longer
exists.

With a lot of debug effort, Grzegorz Nosek developed a small program (see
below) to reproduce a crash on recent kernels. This crash is a regression
introduced by the commit:

commit 527b3e4773628b30d03323a2cb5fb0d84441990f
Author: Sukadev Bhattiprolu <sukadev@us.ibm.com>
Date:   Mon Oct 13 10:43:08 2008 +0100

To fix, ensure that the dentry associated with the inode has not yet been
deleted/unhashed by devpts_pty_kill().

See also:
https://lists.linux-foundation.org/pipermail/containers/2009-July/019273.html

tty-bug.c:

#define _GNU_SOURCE
#include <fcntl.h>
#include <sched.h>
#include <stdlib.h>
#include <sys/mount.h>
#include <sys/signal.h>
#include <unistd.h>
#include <stdio.h>

#include <linux/fs.h>

void dummy(int sig)
{
}

static int child(void *unused)
{
int fd;

signal(SIGINT, dummy); signal(SIGHUP, dummy);
pause(); /* cheesy synchronisation to wait for /dev/pts/0 to appear */

mount("/dev/pts/0", "/dev/console", NULL, MS_BIND, NULL);
sleep(2);

fd = open("/dev/console", O_RDWR);
dup(0); dup(0);
write(1, "Hello world!\n", sizeof("Hello world!\n")-1);
return 0;
}

int main(void)
{
pid_t pid;
char *stack;

stack = malloc(16384);
pid = clone(child, stack+16384, CLONE_NEWNS|SIGCHLD, NULL);

open("/dev/ptmx", O_RDWR|O_NOCTTY|O_NONBLOCK);

unlockpt(fd); grantpt(fd);

sleep(2);
kill(pid, SIGHUP);
sleep(1);
return 0; /* exit before child opens /dev/console */
}

Reported-by: Grzegorz Nosek <root@localdomain.pl>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Tested-by: Serge Hallyn <serue@us.ibm.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoSerial: Do not read IIR in serial8250_start_tx when UART_BUG_TXEN
Ian Jackson [Wed, 18 Nov 2009 10:08:11 +0000 (11:08 +0100)]
Serial: Do not read IIR in serial8250_start_tx when UART_BUG_TXEN

Do not read IIR in serial8250_start_tx when UART_BUG_TXEN

Reading the IIR clears some oustanding interrupts so it is not safe.
Instead, simply transmit immediately if the buffer is empty without
regard to IIR.

Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Jiri Kosina <jkosina@suse.cz>
Cc: Alan Cox <alan@linux.intel.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: docs: serial/tty, add to ldisc methods
Tilman Schmidt [Thu, 5 Nov 2009 00:04:52 +0000 (16:04 -0800)]
tty: docs: serial/tty, add to ldisc methods

A small addition to the ldisc method descriptions.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoSerial: pxa: work around Errata #75
Uwe Kleine-König [Fri, 6 Nov 2009 20:40:46 +0000 (21:40 +0100)]
Serial: pxa: work around Errata #75

Intel(R) PXA27x Processor Family Specification Update (Nov 2005)
says:

  E75. UART: Baud rate may not be programmed correctly on
       back-to-back writes.

  Problem:
  When programming the Divisor Latch registers, Low and High (DLL and
  DLH), with back-to-back writes, the second register write may not
  take effect. The result is an incorrect baud rate.

  Workaround:
  After programming the first Divisor Latch register, read and verify
  it before programming the second Divisor Latch register.

This was hit when changing the baud rate from 115200 to 9600 while
receiving characters at 9600 Bd.

And fixed indention of some comments nearby.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
Cc: Eric Miao <eric.y.miao@gmail.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: Mike Rapoport <mike@compulab.co.il>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoserial, 8250: calculate irqflags bitmask before loop
André Goddard Rosa [Sun, 25 Oct 2009 14:01:34 +0000 (12:01 -0200)]
serial, 8250: calculate irqflags bitmask before loop

Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoserial: cascade needless conditionals
André Goddard Rosa [Sun, 25 Oct 2009 13:18:26 +0000 (11:18 -0200)]
serial: cascade needless conditionals

Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoserial: fix NULL pointer dereference
André Goddard Rosa [Sun, 25 Oct 2009 13:16:32 +0000 (11:16 -0200)]
serial: fix NULL pointer dereference

If kzalloc() or alloc_tty_driver() fails, we call:
    put_tty_driver(normal = NULL).

Then:
    put_tty_driver -> tty_driver_kref_put -> kref_put(&NULL->kref, ...)

Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoopticon: Fix resume logic
Alan Cox [Tue, 6 Oct 2009 15:06:57 +0000 (16:06 +0100)]
opticon: Fix resume logic

Opticon now takes the right mutex to check the port status but the status
check is done wrongly for the modern serial code, so fix it.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb_serial: Kill port mutex
Alan Cox [Tue, 6 Oct 2009 15:06:46 +0000 (16:06 +0100)]
usb_serial: Kill port mutex

The tty port has a port mutex used for all the port related locking so we
don't need the one in the USB serial layer any more.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb_serial: Use the shutdown() operation
Alan Cox [Tue, 6 Oct 2009 15:06:36 +0000 (16:06 +0100)]
usb_serial: Use the shutdown() operation

As Alan Stern pointed out - now we have tty_port_open the shutdown method
and locking allow us to whack the other bits into the full helper methods
and provide a shutdown op which the tty port code will synchronize with
setup for us.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty_port: coding style cleaning pass
Alan Cox [Tue, 6 Oct 2009 15:06:21 +0000 (16:06 +0100)]
tty_port: coding style cleaning pass

Mind the hoover wire...

Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty_port: add "tty_port_open" helper
Alan Cox [Tue, 6 Oct 2009 15:06:11 +0000 (16:06 +0100)]
tty_port: add "tty_port_open" helper

For the moment this just moves the USB logic over and fixes the 'what if
we open and hangup at the same time' race noticed by Oliver Neukum.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: stallion: kill BKL ioctl
Alan Cox [Tue, 13 Oct 2009 15:34:15 +0000 (16:34 +0100)]
tty: stallion: kill BKL ioctl

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: istallion: Kill off the BKL ioctl
Alan Cox [Tue, 13 Oct 2009 15:34:06 +0000 (16:34 +0100)]
tty: istallion: Kill off the BKL ioctl

Fairly trivial as the BKL push down into the methods has already been done.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: esp: remove broken driver
Alan Cox [Fri, 9 Oct 2009 11:56:41 +0000 (12:56 +0100)]
tty: esp: remove broken driver

The ESP driver has been marked broken for years. It's an old ISA device
that clearly nobody cares about any more. Remove it

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotty: const: constify remaining tty_operations
Alexey Dobriyan [Fri, 2 Oct 2009 20:12:06 +0000 (00:12 +0400)]
tty: const: constify remaining tty_operations

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: adding EEH handlers
Breno Leitao [Wed, 14 Oct 2009 17:57:51 +0000 (14:57 -0300)]
jsm: adding EEH handlers

Adding EEH handlers for the serial jsm driver. This patch adds
the PCI error handlers and also register them to be called when
a error is detected.

Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com>
Acked-by: Scott Kilau <scottk@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: removing the field jsm_board->intr_count
Breno Leitão [Tue, 29 Sep 2009 18:16:59 +0000 (15:16 -0300)]
jsm: removing the field jsm_board->intr_count

Currently there is a field in the jsm_board structure to cont
the number of interrupt that the card recevived, but it's not
working properly when the IRQ line is shared, and also nowhere
else this field is used. So, This patch is removing it.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <Scott.Kilau@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: Removing unused jsm_channel->ch_wopen field
Breno Leitão [Tue, 29 Sep 2009 18:16:58 +0000 (15:16 -0300)]
jsm: Removing unused jsm_channel->ch_wopen field

Currently the jsm_channel->ch_wopen field is defined and never
used. So, this patch removes it.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <Scott.Kilau@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: Remove ch_cpstime field
Breno Leitão [Tue, 29 Sep 2009 18:16:57 +0000 (15:16 -0300)]
jsm: Remove ch_cpstime field

Currently the field jsm_channel->ch_cpstime is defined but never
used, so this patch removes it.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <Scott.Kilau@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: removing ch_old_baud field
Breno Leitão [Tue, 29 Sep 2009 18:16:56 +0000 (15:16 -0300)]
jsm: removing ch_old_baud field

Currently the field jsm_channel->ch_old_baud is not used, just
assigned in a lot of places but never used. This patches removes
this field.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <scottk@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: remove the ch_custom_speed field
Breno Leitão [Tue, 29 Sep 2009 18:16:55 +0000 (15:16 -0300)]
jsm: remove the ch_custom_speed field

Currently the ch_custom_speed field exists but is never used,
so, this patch removes it.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <scottk@digi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: Rewriting a bad log message
Breno Leitão [Thu, 24 Sep 2009 19:58:23 +0000 (16:58 -0300)]
jsm: Rewriting a bad log message

Actually jsm displays "Device Added" 8 times (for a 8 port device).
This silly patch just makes things more informative, showing
the port (instead of the device) that was added.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <scottk@digi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojsm: IRQ handlers doesn't need to have IRQ_DISABLED enabled
Breno Leitão [Thu, 24 Sep 2009 19:58:22 +0000 (16:58 -0300)]
jsm: IRQ handlers doesn't need to have IRQ_DISABLED enabled

Currently jsm is showing the following message when loaded:

IRQ 432/JSM: IRQF_DISABLED is not guaranteed on shared IRQs

It's because the request_irq() is called using IRQF_DISABLED
and IRQF_SHARED.
Actually there is no need to use IRQF_DISABLED in this driver.

Signed-off-by: Breno Leitão <leitao@linux.vnet.ibm.com>
Cc: Scott Kilau <scottk@digi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoMerge branch 'drm-nouveau-pony' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 11 Dec 2009 22:32:49 +0000 (14:32 -0800)]
Merge branch 'drm-nouveau-pony' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-nouveau-pony' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/nouveau: Add DRM driver for NVIDIA GPUs

14 years agoMerge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Fri, 11 Dec 2009 22:32:30 +0000 (14:32 -0800)]
Merge branch 'drm-linus' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/ttm: export some functions useful to drivers using ttm
  drm/radeon/kms/avivo: fix typo in new_pll module description
  drm/radeon/kms: Convert radeon to new ttm_bo_init
  drm/ttm: Convert ttm_buffer_object_init to use ttm_placement

14 years agoxfs: Fix error return for fallocate() on XFS
Jason Gunthorpe [Tue, 24 Nov 2009 21:52:53 +0000 (21:52 +0000)]
xfs: Fix error return for fallocate() on XFS

Noticed that through glibc fallocate would return 28 rather than -1
and errno = 28 for ENOSPC. The xfs routines uses XFS_ERROR format
positive return error codes while the syscalls use negative return
codes.  Fixup the two cases in xfs_vn_fallocate syscall to convert to
negative.

Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: cleanup dmapi macros in the umount path
Christoph Hellwig [Sat, 14 Nov 2009 16:17:24 +0000 (16:17 +0000)]
xfs: cleanup dmapi macros in the umount path

Stop the flag saving as we never mangle those in the unmount path, and
hide all the weird arguents to the dmapi code inside the
XFS_SEND_PREUNMOUNT / XFS_SEND_UNMOUNT macros.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: remove incorrect sparse annotation for xfs_iget_cache_miss
Christoph Hellwig [Sat, 14 Nov 2009 16:17:23 +0000 (16:17 +0000)]
xfs: remove incorrect sparse annotation for xfs_iget_cache_miss

xfs_iget_cache_miss does not get called with the pag_ici_lock held, so
the __releases annotation is incorrect.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: kill the STATIC_INLINE macro
Christoph Hellwig [Sat, 14 Nov 2009 16:17:22 +0000 (16:17 +0000)]
xfs: kill the STATIC_INLINE macro

Remove our own STATIC_INLINE macro.  For small function inside
implementation files just use STATIC and let gcc inline it, and for
those in headers do the normal static inline - they are all small
enough to be inlined for debug builds, too.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: uninline xfs_get_extsz_hint
Christoph Hellwig [Sat, 14 Nov 2009 16:17:21 +0000 (16:17 +0000)]
xfs: uninline xfs_get_extsz_hint

This function is too large to efficiently be inlined.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: rename xfs_attr_fetch to xfs_attr_get_int
Christoph Hellwig [Sat, 14 Nov 2009 16:17:20 +0000 (16:17 +0000)]
xfs: rename xfs_attr_fetch to xfs_attr_get_int

Using a totally different name for the low-level get operation does
not fit the _int convention used in the rest of the attr code, so
rename it.

While we're at it also fix the prototype to use the normal convention
and mark it static as it's never used outside of xfs_attr.c.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: simplify xfs_buf_get / xfs_buf_read interfaces
Christoph Hellwig [Tue, 24 Nov 2009 18:02:23 +0000 (18:02 +0000)]
xfs: simplify xfs_buf_get / xfs_buf_read interfaces

Currently the low-level buffer cache interfaces are highly confusing
as we have a _flags variant of each that does actually respect the
flags, and one without _flags which has a flags argument that gets
ignored and overriden with a default set.  Given that very few places
use the default arguments get rid of the duplication and convert all
callers to pass the flags explicitly.  Also remove the now confusing
_flags postfix.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: remove IO_ISAIO
Christoph Hellwig [Sat, 14 Nov 2009 16:17:18 +0000 (16:17 +0000)]
xfs: remove IO_ISAIO

We set the IO_ISAIO flag for all read/write I/O since early Linux
2.6.x.  Remove it as it has lost it's purpose long ago.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: Wrapped journal record corruption on read at recovery
Andy Poling [Tue, 3 Nov 2009 17:26:47 +0000 (17:26 +0000)]
xfs: Wrapped journal record corruption on read at recovery

Summary of problem:

If a journal record wraps at the physical end of the journal, it has to be
read in two parts in xlog_do_recovery_pass(): a read at the physical end and a
read at the physical beginning.  If xlog_bread() has to re-align the first
read, the second read request does not take that re-alignment into account.
If the first read was re-aligned, the second read over-writes the end of the
data from the first read, effectively corrupting it.  This can happen either
when reading the record header or reading the record data.

The first sanity check in xlog_recover_process_data() is to check for a valid
clientid, so that is the error reported.

Summary of fix:

If there was a first read at the physical end, XFS_BUF_PTR() returns where the
data was requested to begin.  Conversely, because it is the result of
xlog_align(), offset indicates where the requested data for the first read
actually begins - whether or not xlog_bread() has re-aligned it.

Using offset as the base for the calculation of where to place the second read
data ensures that it will be correctly placed immediately following the data
from the first read instead of sometimes over-writing the end of it.

The attached patch has resolved the reported problem of occasional inability
to recover the journal (reporting "bad clientid").

Signed-off-by: Andy Poling <andy@realbig.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: cleanup data end I/O handlers
Christoph Hellwig [Fri, 30 Oct 2009 09:11:47 +0000 (09:11 +0000)]
xfs: cleanup data end I/O handlers

Currently we have different end I/O handlers for read vs the different
types of write I/O.  But they are all very similar so we could just
use one with a few conditionals and reduce code size a lot.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: use WRITE_SYNC_PLUG for synchronous writeout
Christoph Hellwig [Fri, 30 Oct 2009 09:09:15 +0000 (09:09 +0000)]
xfs: use WRITE_SYNC_PLUG for synchronous writeout

The VM and I/O schedulers now expect us to use WRITE_SYNC_PLUG for
synchronous writeout.  Right now I can't see any changes in performance
numbers with this, but we're getting some beating for not using it,
and the knowledge definitely could help the block code to make better
decisions.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: reset the i_iolock lock class in the reclaim path
Christoph Hellwig [Mon, 19 Oct 2009 04:05:26 +0000 (04:05 +0000)]
xfs: reset the i_iolock lock class in the reclaim path

The iolock is used for protecting reads, writes and block truncates
against each other.  We have two classes of callers, the first one is
induced by a file operation and requires a reference to the inode be
held and not dropped after the operation is done:

 - xfs_vm_vmap, xfs_vn_fallocate, xfs_read, xfs_write, xfs_splice_read,
   xfs_splice_write and xfs_setattr are all implementations of VFS
   methods that require a live inode
 - xfs_getbmap and xfs_swap_extents are ioctl subcommand for which the
   same is true
 - xfs_truncate_file is only called on quota inodes just returned from
   xfs_iget
 - xfs_sync_inode_data does the lock just after an igrab()
 - xfs_filestream_associate and xfs_filestream_new_ag take the iolock
   on the parent inode of an inode which by VFS rules must be referenced

And we have various calls to truncate blocks past EOF or the whole
file when dropping the last reference to an inode.  Unfortunately
lockdep complains when we do memory allocations that can recurse into
the filesystem in the first class because the second class happens to
take the same lock.  To avoid this re-init the iolock in the beginning
of xfs_fs_clear_inode to get a new lock class.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: I/O completion handlers must use NOFS allocations
Christoph Hellwig [Mon, 19 Oct 2009 04:00:03 +0000 (04:00 +0000)]
xfs: I/O completion handlers must use NOFS allocations

When completing I/O requests we must not allow the memory allocator to
recurse into the filesystem, as we might deadlock on waiting for the
I/O completion otherwise.  The only thing currently allocating normal
GFP_KERNEL memory is the allocation of the transaction structure for
the unwritten extent conversion.  Add a memflags argument to
_xfs_trans_alloc to allow controlling the allocator behaviour.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Thomas Neumann <tneumann@users.sourceforge.net>
Tested-by: Thomas Neumann <tneumann@users.sourceforge.net>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: fix mmap_sem/iolock inversion in xfs_free_eofblocks
Christoph Hellwig [Mon, 19 Oct 2009 04:03:46 +0000 (04:03 +0000)]
xfs: fix mmap_sem/iolock inversion in xfs_free_eofblocks

When xfs_free_eofblocks is called from ->release the VM might already
hold the mmap_sem, but in the write path we take the iolock before
taking the mmap_sem in the generic write code.

Switch xfs_free_eofblocks to only trylock the iolock if called from
->release and skip trimming the prellocated blocks in that case.
We'll still free them later on the final iput.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: simplify inode teardown
Christoph Hellwig [Tue, 29 Sep 2009 13:48:56 +0000 (13:48 +0000)]
xfs: simplify inode teardown

Currently the reclaim code for the case where we don't reclaim the
final reclaim is overly complicated.  We know that the inode is clean
but instead of just directly reclaiming the clean inode we go through
the whole process of marking the inode reclaimable just to directly
reclaim it from the calling context.  Besides being overly complicated
this introduces a race where iget could recycle an inode between
marked reclaimable and actually being reclaimed leading to panics.

This patch gets rid of the existing reclaim path, and replaces it with
a simple call to xfs_ireclaim if the inode was clean.  While we're at
it we also use the slightly more lax xfs_inode_clean check we'd use
later to determine if we need to flush the inode here.

Finally get rid of xfs_reclaim function and place the remaining small
bits of reclaim code directly into xfs_fs_destroy_inode.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Patrick Schreurs <patrick@news-service.com>
Reported-by: Tommy van Leeuwen <tommy@news-service.com>
Tested-by: Patrick Schreurs <patrick@news-service.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Alex Elder <aelder@sgi.com>