linux-toradex.git/drivers/s390, branch v3.2.2

[SCSI] zfcp: return early from slave_destroy if slave_alloc returned early

2011-12-14T11:40:43+00:00

zfcp_scsi_slave_destroy erroneously always tried to finish its task
even if the corresponding previous zfcp_scsi_slave_alloc returned
early. This can lead to kernel page faults on accessing uninitialized
fields of struct zfcp_scsi_dev in zfcp_erp_lun_shutdown_wait. Take the
port field of the struct to determine if slave_alloc returned early.

This zfcp bug is exposed by 4e6c82b (in turn fixing f7c9c6b to be
compatible with 21208ae) which can call slave_destroy for a
corresponding previous slave_alloc that did not finish.

This patch is based on James Bottomley's fix suggestion in
http://www.spinics.net/lists/linux-scsi/msg55449.html.

Signed-off-by: Steffen Maier 
Cc:  #2.6.38+
Signed-off-by: James Bottomley

[S390] ap: Setup timer for sending messages after reset.

2011-12-01T12:32:18+00:00

Setup timer for processing messages in request queue after a
successful AP bus device reset.

Signed-off-by: Holger Dengler 
Signed-off-by: Martin Schwidefsky

[S390] cio: fix chsc_chp_vary

2011-12-01T12:32:18+00:00

The functions called by chsc_chp_vary operate on pointers to channel
path ids not channel path links. (This worked by chance since the id
is the first member of the link structure)

Signed-off-by: Sebastian Ott 
Signed-off-by: Martin Schwidefsky

[S390] cio: provide fake irb for transport mode IO

2011-12-01T12:32:18+00:00

If a driver wants to do command mode IO while CIO is doing
online path verification we ignore this request and provide
a fake irb when we are done and the driver can do IO again.

For transport mode IO we have no such mechanism, giving the
driver no other chance then to retry the action until we are
done. This is not very reliable.

Provide a fake irb for transport mode IO as well.

Signed-off-by: Sebastian Ott 
Signed-off-by: Martin Schwidefsky

[S390] cio: disallow driver io for known to be broken paths

2011-12-01T12:32:18+00:00

When a driver requests to do IO, we will adjust the mask of
paths to be used to exclude varied offline paths.

Drivers trying to do IO solely on paths which are online but some
way defective may lack the information to do proper error handling.

There is no reason to allow the usage of known to be broken paths.
Thus restrict the paths a ccw driver can use for IO to a subset of
the paths cio found usable (this also excludes offline paths).

Signed-off-by: Sebastian Ott 
Signed-off-by: Martin Schwidefsky

[S390] hibernate: directly trigger subchannel evaluation

2011-12-01T12:32:17+00:00

Using the generic css_schedule_eval to evaluate subchannels
while resuming from hibernation is very slow when used with
many devices. Provide a new evaluation trigger which exploits
css_sched_sch_todo and use this in the resume callback for
ccw devices.

Signed-off-by: Sebastian Ott 
Signed-off-by: Martin Schwidefsky

Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

2011-11-20T22:59:33+00:00

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (86 commits)
  ipv4: fix redirect handling
  ping: dont increment ICMP_MIB_INERRORS
  sky2: fix hang in napi_disable
  sky2: enforce minimum ring size
  bonding: Don't allow mode change via sysfs with slaves present
  f_phonet: fix page offset of first received fragment
  stmmac: fix pm functions avoiding sleep on spinlock
  stmmac: remove spin_lock in stmmac_ioctl.
  stmmac: parameters auto-tuning through HW cap reg
  stmmac: fix advertising 1000Base capabilties for non GMII iface
  stmmac: use mdelay on timeout of sw reset
  sky2: version 1.30
  sky2: used fixed RSS key
  sky2: reduce default Tx ring size
  sky2: rename up/down functions
  sky2: pci posting issues
  sky2: fix hang on shutdown (and other irq issues)
  r6040: fix check against MCRO_HASHEN bit in r6040_multicast_list
  MAINTAINERS: change email address for shemminger
  pch_gbe: Move #include of module.h
  ...

qeth: Reduce CPU consumption through less SIGA-r calls

2011-11-16T22:28:20+00:00

Patch avoids SIGA-r calls in case of SIGA-r required. It only calls
SIGA-r if a threshold of free buffer is reached. CPU consumption is
reduced as a consequence.

Signed-off-by: Einar Lueck 
Signed-off-by: Frank Blaschka 
Signed-off-by: David S. Miller

qeth: l3 fix rcu splat in xmit

2011-11-16T22:28:14+00:00

when use dst_get_neighbour to get neighbour, we need
rcu_read_lock to protect, since dst_get_neighbour uses
rcu_dereference.

Signed-off-by: David S. Miller

netiucv: reinsert dev_alloc_name for device naming

2011-11-16T22:28:10+00:00

Invocation of dev_alloc_name() is re-inserted, because the created
net_device name is used to create the device name for the iucv bus.
This device is created before the register_netdev call.

Signed-off-by: Ursula Braun 
Signed-off-by: Frank Blaschka 
Signed-off-by: David S. Miller