platform/kernel/linux-rpi.git
7 years agos390/spinlock: use atomic primitives for spinlocks
Martin Schwidefsky [Mon, 28 Nov 2016 14:50:48 +0000 (15:50 +0100)]
s390/spinlock: use atomic primitives for spinlocks

Add a couple more __atomic_xxx function to atomic_ops.h and use them
to replace the compare-and-swap inlines in the spinlock code. This
changes the type of the lock value from unsigned int to int.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpumf: simplify detection of guest samples
Martin Schwidefsky [Tue, 4 Apr 2017 10:55:11 +0000 (12:55 +0200)]
s390/cpumf: simplify detection of guest samples

There are three different code levels in regard to the identification
of guest samples. They differ in the way the LPP instruction is used.

1) Old kernels without the LPP instruction. The guest program parameter
   is always zero.
2) Newer kernels load the process pid into the program parameter with LPP.
   The guest program parameter is non-zero if the guest executes in a
   process != idle.
3) The latest kernels load ((1UL << 31) | pid) with LPP to make the value
   non-zero even for the idle task. The guest program parameter is non-zero
   if the guest is running.

All kernels load the process pid to CR4 on context switch. The CPU sampling
code uses the value in CR4 to decide between guest and host samples in case
the guest program parameter is zero. The three cases:

1) CR4==pid, gpp==0
2) CR4==pid, gpp==pid
3) CR4==pid, gpp==((1UL << 31) | pid)

The load-control instruction to load the pid into CR4 is expensive and the
goal is to remove it. To distinguish the host CR4 from the guest pid for
the idle process the maximum value 0xffff for the PASN is used.
This adds a fourth case for a guest OS with an updated kernel:

4) CR4==0xffff, gpp=((1UL << 31) | pid)

The host kernel will have CR4==0xffff and will use (gpp!=0 || CR4!==0xffff)
to identify guest samples. This works nicely with all 4 cases, the only
possible issue would be a guest with an old kernel (gpp==0) and a process
pid of 0xffff. Well, don't do that..

Suggested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pci: remove forward declaration
Sebastian Ott [Mon, 27 Mar 2017 17:22:27 +0000 (19:22 +0200)]
s390/pci: remove forward declaration

Move a struct definition to get rid of a forward declaration.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pci: increase the PCI_NR_FUNCTIONS default
Sebastian Ott [Mon, 27 Mar 2017 17:09:15 +0000 (19:09 +0200)]
s390/pci: increase the PCI_NR_FUNCTIONS default

Users complained that they are hitting the limit of 64 functions (some
physical functions can spawn lots of virtual functions). Double the
default limit. With the latest savings in static data usage this
increases the image size by only 520 bytes.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pci: reduce iomap size (even more)
Sebastian Ott [Mon, 27 Mar 2017 17:07:24 +0000 (19:07 +0200)]
s390/pci: reduce iomap size (even more)

Commit c506fff3d3a8 ("s390/pci: resize iomap") reduced the iomap
to NR_FUNCTIONS * PCI_BAR_COUNT elements. Since we only support
functions with 64bit BARs we can cut that number in half.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pci: remove duplicated define
Sebastian Ott [Mon, 27 Mar 2017 16:23:01 +0000 (18:23 +0200)]
s390/pci: remove duplicated define

Address space identifiers are already defined in <asm/pci_insn.h>.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pci: remove unused function
Sebastian Ott [Mon, 27 Mar 2017 16:22:14 +0000 (18:22 +0200)]
s390/pci: remove unused function

barsize was never used. Get rid of it.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: use 64-bit lctlg to load task pid to cr4 on context switch
Martin Schwidefsky [Mon, 3 Apr 2017 11:30:23 +0000 (13:30 +0200)]
s390: use 64-bit lctlg to load task pid to cr4 on context switch

The 32-bit lctl instruction is quite a bit slower than the 64-bit
counter part lctlg. Use the faster instruction.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agoMerge branch 'vfio-ccw-for-martin' of git://git.kernel.org/pub/scm/linux/kernel/git...
Martin Schwidefsky [Fri, 31 Mar 2017 11:01:46 +0000 (13:01 +0200)]
Merge branch 'vfio-ccw-for-martin' of git://git./linux/kernel/git/kvms390/vfio-ccw into features

Pull vfio-ccw branch to add the basic channel I/O passthrough
intrastructure based on vfio.

The focus is on supporting dasd-eckd(cu_type/dev_type = 0x3990/0x3390)
as the target device.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agoMAINTAINERS: Add vfio-ccw maintainers
Dong Jia Shi [Fri, 17 Mar 2017 03:17:43 +0000 (04:17 +0100)]
MAINTAINERS: Add vfio-ccw maintainers

Add Cornelia Huck and myself as the vfio-ccw driver maintainers.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-17-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: introduce support for ccw0
Dong Jia Shi [Fri, 17 Mar 2017 03:17:42 +0000 (04:17 +0100)]
vfio: ccw: introduce support for ccw0

Although Linux does not use format-0 channel command words (CCW0)
these are a non-optional part of the platform spec, and for the sake
of platform compliance, and possibly some non-Linux guests, we have
to support CCW0.

Making the kernel execute a format 0 channel program is too much hassle
because we would need to allocate and use memory which can be addressed
by 24 bit physical addresses (because of CCW0.cda). So we implement CCW0
support by translating the channel program into an equivalent CCW1
program instead.

Based upon an orginal patch by Kai Yue Wang.
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-16-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agodocs: add documentation for vfio-ccw
Dong Jia Shi [Fri, 17 Mar 2017 03:17:41 +0000 (04:17 +0100)]
docs: add documentation for vfio-ccw

Add file Documentation/s390/vfio-ccw.txt that includes details
of vfio-ccw.

Acked-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-15-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: introduce a finite state machine
Dong Jia Shi [Fri, 17 Mar 2017 03:17:40 +0000 (04:17 +0100)]
vfio: ccw: introduce a finite state machine

The current implementation doesn't check if the subchannel is in a
proper device state when handling an event. Let's introduce
a finite state machine to manage the state/event change.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-14-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: return I/O results asynchronously
Dong Jia Shi [Fri, 17 Mar 2017 03:17:39 +0000 (04:17 +0100)]
vfio: ccw: return I/O results asynchronously

Introduce a singlethreaded workqueue to handle the I/O interrupts.
With the work added to this queue, we store the I/O results to the
io_region of the subchannel, then signal the userspace program to
handle the results.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-13-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: realize VFIO_DEVICE_G(S)ET_IRQ_INFO ioctls
Dong Jia Shi [Fri, 17 Mar 2017 03:17:38 +0000 (04:17 +0100)]
vfio: ccw: realize VFIO_DEVICE_G(S)ET_IRQ_INFO ioctls

Realize VFIO_DEVICE_GET_IRQ_INFO ioctl to retrieve
VFIO_CCW_IO_IRQ information.

Realize VFIO_DEVICE_SET_IRQS ioctl to set an eventfd fd for
VFIO_CCW_IO_IRQ. Once a write operation to the ccw_io_region
was performed, trigger a signal on this fd.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-Id: <20170317031743.40128-12-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: realize VFIO_DEVICE_RESET ioctl
Dong Jia Shi [Fri, 17 Mar 2017 03:17:37 +0000 (04:17 +0100)]
vfio: ccw: realize VFIO_DEVICE_RESET ioctl

Introduce VFIO_DEVICE_RESET ioctl for vfio-ccw to make it possible
to hot-reset the device.

We try to achieve a reset by first disabling the subchannel and
then enabling it again: this should clear all state at the subchannel.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-11-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: realize VFIO_DEVICE_GET_REGION_INFO ioctl
Dong Jia Shi [Fri, 17 Mar 2017 03:17:36 +0000 (04:17 +0100)]
vfio: ccw: realize VFIO_DEVICE_GET_REGION_INFO ioctl

Introduce device information about vfio-ccw: VFIO_DEVICE_FLAGS_CCW.
Realize VFIO_DEVICE_GET_REGION_INFO ioctl for vfio-ccw.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-Id: <20170317031743.40128-10-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: handle ccw command request
Dong Jia Shi [Fri, 17 Mar 2017 03:17:35 +0000 (04:17 +0100)]
vfio: ccw: handle ccw command request

We implement the basic ccw command handling infrastructure
here:
1. Translate the ccw commands.
2. Issue the translated ccw commands to the device.
3. Once we get the execution result, update the guest SCSW
   with it.

Acked-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-9-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: introduce ccw_io_region
Dong Jia Shi [Fri, 17 Mar 2017 03:17:34 +0000 (04:17 +0100)]
vfio: ccw: introduce ccw_io_region

To provide user-space a set of interfaces to:
1. pass in a ccw program to perform an I/O operation.
2. read back I/O results of the completed I/O operations.
We introduce an MMIO region for the vfio-ccw device here.

This region is defined to content:
1. areas to store arguments that an ssch required.
2. areas to store the I/O results.

Using pwrite/pread to the device on this region, a user-space program
could write/read data to/from the vfio-ccw device.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-8-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: register vfio_ccw to the mediated device framework
Dong Jia Shi [Fri, 17 Mar 2017 03:17:33 +0000 (04:17 +0100)]
vfio: ccw: register vfio_ccw to the mediated device framework

To make vfio support subchannel devices, we need to leverage the
mediated device framework to create a mediated device for the
subchannel device.

This registers the subchannel device to the mediated device
framework during probe to enable mediated device creation.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-7-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: introduce channel program interfaces
Dong Jia Shi [Fri, 17 Mar 2017 03:17:32 +0000 (04:17 +0100)]
vfio: ccw: introduce channel program interfaces

Introduce ccwchain structure and helper functions that can be used to
handle a channel program issued from a virtual machine.

The following limitations apply:
1. Supports only prefetch enabled mode.
2. Supports idal(c64) ccw chaining.
3. Supports 4k idaw.
4. Supports ccw1.
5. Supports direct ccw chaining by translating them to idal ccws.

CCW translation requires to leverage the vfio_(un)pin_pages interfaces
to pin/unpin sets of mem pages frequently. Currently we have a lack of
support to do this in an efficient way. So we introduce pfn_array data
structure and helper functions to handle pin/unpin operations here.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-6-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: basic implementation for vfio_ccw driver
Dong Jia Shi [Fri, 17 Mar 2017 03:17:31 +0000 (04:17 +0100)]
vfio: ccw: basic implementation for vfio_ccw driver

To make vfio support subchannel devices, we need a css driver for
the vfio subchannels. This patch adds a basic vfio-ccw subchannel
driver for this purpose.

To enable VFIO for vfio-ccw, enable S390_CCW_IOMMU config option
and configure VFIO as required.

Acked-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-5-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agovfio: ccw: define device_api strings
Dong Jia Shi [Fri, 17 Mar 2017 03:17:30 +0000 (04:17 +0100)]
vfio: ccw: define device_api strings

Define vfio-ccw device API strings. CCW vendor driver using mediated
device framework should use this string for device_api attribute.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-Id: <20170317031743.40128-4-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agos390: cio: export more interfaces
Dong Jia Shi [Fri, 17 Mar 2017 03:17:29 +0000 (04:17 +0100)]
s390: cio: export more interfaces

Export the common I/O interfaces those are needed by an I/O
subchannel driver to actually talk to the subchannel.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Cc: Sebastian Ott <sebott@linux.vnet.ibm.com>
Cc: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Acked-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-3-bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agos390: cio: introduce cio_cancel_halt_clear
Dong Jia Shi [Fri, 17 Mar 2017 03:17:28 +0000 (04:17 +0100)]
s390: cio: introduce cio_cancel_halt_clear

For future code reuse purpose, this decouples the cio code with
the ccw device specific parts from ccw_device_cancel_halt_clear,
and makes a new common I/O interface named cio_cancel_halt_clear.

Reviewed-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Cc: Sebastian Ott <sebott@linux.vnet.ibm.com>
Cc: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Acked-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Message-Id: <20170317031743.40128-2-bjsdjshi@linux.vnet.ibm.com>
[CH: Fix typo]
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
7 years agos390/cpum_cf: make hw_perf_event_update() a void function
Hendrik Brueckner [Mon, 20 Feb 2017 15:10:38 +0000 (16:10 +0100)]
s390/cpum_cf: make hw_perf_event_update() a void function

The return code of hw_perf_event_update() is not evaluated by
its callers.  Hence, simplify the function by removing the
return code.

Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpu_mf: remove register variable in __ecctr()
Hendrik Brueckner [Mon, 20 Feb 2017 15:07:43 +0000 (16:07 +0100)]
s390/cpu_mf: remove register variable in __ecctr()

Using a register variable for r4 is not necessary.  Let the
compiler decide the register to be used.

Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpum_cf: correct variable naming (cleanup)
Hendrik Brueckner [Fri, 17 Feb 2017 11:53:55 +0000 (12:53 +0100)]
s390/cpum_cf: correct variable naming (cleanup)

Make clear that the event definitions relate to the counter
facility (cf) and not to the sampling facility (sf).

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpum_cf: add IBM z13 counter event names
Hendrik Brueckner [Fri, 17 Feb 2017 11:16:00 +0000 (12:16 +0100)]
s390/cpum_cf: add IBM z13 counter event names

Add the event names for the IBM z13/z13s specific CPU-MF counters.

Also improve the merging of the generic and model specific events
so that their sysfs attribute definitions completely reside in
memory.  Hence, flagging the generic event attribute definitions
as initdata too.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpum_cf: add support for the MT-diagnostic counter set (z13)
Hendrik Brueckner [Fri, 3 Jun 2016 14:55:03 +0000 (16:55 +0200)]
s390/cpum_cf: add support for the MT-diagnostic counter set (z13)

Complete the IBM z13 support and support counters from the
MT-diagnostic counter set.  Note that this counter set is
available only if SMT is enabled.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpum_cf: cleanup event/counter validation
Hendrik Brueckner [Mon, 13 Feb 2017 11:32:17 +0000 (12:32 +0100)]
s390/cpum_cf: cleanup event/counter validation

The validate_event() function just checked for reserved counters
in particular CPU-MF counter sets.  Because the number of counters
in counter sets vary among different hardware models, remove the
explicit check to tolerate new models.

Reserved counters are not accounted and, thus, will return zero.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpum_cf: update counter numbers to ecctr limits
Hendrik Brueckner [Mon, 13 Feb 2017 11:30:09 +0000 (12:30 +0100)]
s390/cpum_cf: update counter numbers to ecctr limits

Use the highest counter number that can be specified for the
ecctr (extract CPU counter) instruction for perf.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: use generic headers if possible
Heiko Carstens [Wed, 29 Mar 2017 08:30:48 +0000 (10:30 +0200)]
s390: use generic headers if possible

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/uapi: use generic headers if possible
Heiko Carstens [Wed, 29 Mar 2017 07:59:21 +0000 (09:59 +0200)]
s390/uapi: use generic headers if possible

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: make MAX_PHYSMEM_BITS configurable
Heiko Carstens [Mon, 27 Mar 2017 13:36:20 +0000 (15:36 +0200)]
s390: make MAX_PHYSMEM_BITS configurable

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: enable ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT
Heiko Carstens [Mon, 27 Mar 2017 10:28:25 +0000 (12:28 +0200)]
s390: enable ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT

Deferred struct page initialization works on s390. However it makes
only sense for the fake numa case, since the kthreads that initialize
struct pages are started per node. Without fake numa there is just a
single node and therefore no gain.
However there is no reason to not enable this feature. Therefore
select the config option and enable the feature in all config files.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/MAINTAINERS: add gmap.c to kvm maintainers
Christian Borntraeger [Wed, 22 Mar 2017 11:36:07 +0000 (12:36 +0100)]
s390/MAINTAINERS: add gmap.c to kvm maintainers

gmap.c deals mostly with KVM-related memory management, so a lot
of changes to this file will come via the KVM tree. Reflect this
in MAINTAINERS. Please note that there are intricate ties to
arch/s390/mm/pgtable.c. If changes are needed in both files,
this will continue to be submitted via the s390 tree (or a
topic branch if necessary).

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/kdump: Add final note
Michael Holzheu [Thu, 23 Mar 2017 20:02:54 +0000 (21:02 +0100)]
s390/kdump: Add final note

Since linux v3.14 with commit 38dfac843cb6d7be1 ("vmcore: prevent PT_NOTE
p_memsz overflow during header update") on s390 we get the following
message in the kdump kernel:

  Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x6b6b6b6b,
  n_descsz=0x6b6b6b6b

The reason for this is that we don't create a final zero note in
the ELF header which the proc/vmcore code uses to find out the end
of the notes section (see also kernel/kexec_core.c:final_note()).

It still worked on s390 by chance because we (most of the time?) have the
byte pattern 0x6b6b6b6b after the notes section which also makes the notes
parsing code stop in update_note_header_size_elf64() because 0x6b6b6b6b is
interpreded as note size:

  if ((real_sz + sz) > max_sz) {
          pr_warn("Warning: Exceeded p_memsz, dropping P ...);
          break;
  }

So fix this and add the missing final note to the ELF header.
We don't have to adjust the memory size for ELF header ("alloc_size")
because the new ELF note still fits into the 0x1000 base memory.

Cc: stable@vger.kernel.org # v4.4+
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: remove HAVE_ARCH_EARLY_PFN_TO_NID select statement
Heiko Carstens [Mon, 27 Mar 2017 06:01:29 +0000 (08:01 +0200)]
s390: remove HAVE_ARCH_EARLY_PFN_TO_NID select statement

HAVE_ARCH_EARLY_PFN_TO_NID selects a not present Kconfig
option. Therefore remove it.
Given that the first call of early_pfn_to_nid() happens after
numa_setup() finished to establish the memory to node mapping, there
is no need to implement an architecture private version of
__early_pfn_to_nid() like (only) ia64 does.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/dasd: suppress command reject error for query host access command
Stefan Haberland [Wed, 22 Mar 2017 10:06:20 +0000 (11:06 +0100)]
s390/dasd: suppress command reject error for query host access command

On some z/VM systems the query host access command is not supported for
temp disks, though the corresponding feature code is set.
This does not have any impact beside that the information is not available.
Suppress the full blown command reject error messages to not confuse the
user. The error is still logged in the s390dbf.

Signed-off-by: Stefan Haberland <sth@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/dasd: check if query host access feature is supported
Stefan Haberland [Wed, 22 Mar 2017 09:30:40 +0000 (10:30 +0100)]
s390/dasd: check if query host access feature is supported

Some storage servers might not support the query host access feature.
Check if the corresponding feature code is set.

Signed-off-by: Stefan Haberland <sth@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/mm: Remove double gaddr calculation when notifying
Janosch Frank [Wed, 8 Feb 2017 07:59:56 +0000 (08:59 +0100)]
s390/mm: Remove double gaddr calculation when notifying

ptep_notify and gmap_shadow_notify both need a guest address and
therefore retrieve them from the available virtual host address.

As they operate on the same guest address, we can calculate it once
and then pass it on. As a gmap normally has more than one shadow gmap,
we also do not recalculate for each of them any more.

Signed-off-by: Janosch Frank <frankja@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/facilities: get rid of __ASSEMBLY__ in facility header file
Heiko Carstens [Mon, 20 Mar 2017 13:39:28 +0000 (14:39 +0100)]
s390/facilities: get rid of __ASSEMBLY__ in facility header file

There is no need for the __ASSEMBLY__ ifdefery anymore since the
architecture level set code that deals with facility bits was
converted to C in the meantime.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/facilites: use stfle_fac_list array size for MAX_FACILITY_BIT
Heiko Carstens [Mon, 20 Mar 2017 13:29:50 +0000 (14:29 +0100)]
s390/facilites: use stfle_fac_list array size for MAX_FACILITY_BIT

Use the actual size of the facility list array within the lowcore
structure for the MAX_FACILITY_BIT define instead of a comment which
states what this is good for. This makes it a bit harder to break
things.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/sysinfo: provide remaining stsi information via debugfs
Heiko Carstens [Mon, 13 Mar 2017 15:15:06 +0000 (16:15 +0100)]
s390/sysinfo: provide remaining stsi information via debugfs

Provide the remaining stsi information via debugfs files. This also
might be useful for debugging purposes.

Suggested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/sysinfo,topology: provide raw stsi 15,1,x data via debugfs
Heiko Carstens [Mon, 13 Mar 2017 14:58:59 +0000 (15:58 +0100)]
s390/sysinfo,topology: provide raw stsi 15,1,x data via debugfs

Provide the raw stsi 15,1,x data contents via debugfs. This makes it
much easier to debug unexpected scheduling domains on machines that
provide cpu topology information.

Therefore this file adds a new 's390/stsi' debugfs directory with a
file for each possible topology nesting level that is allowed by the
architecture. The files will be created regardless if the machine
supports all, or any, level.  If a level is not supported, or no data
is available, user space can recognize this with a -EINVAL error code
when trying to read such data.
In addition a 'topology' symlink is created that points to the file
that contains the data that is used to create the scheduling domains.

Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/debugfs: introduce top-level 's390' directory
Heiko Carstens [Mon, 13 Mar 2017 14:58:35 +0000 (15:58 +0100)]
s390/debugfs: introduce top-level 's390' directory

Introduce a top-level 's390' directory which should be used when
adding new s390 specific debug feature files and/or directories.

This makes hopefully sure that the contents of the s390 directory will
be a bit more structured. Right now we have a couple of top-level
files where it is not easy to tell to which subsystem they belong to.

Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/pkey: Introduce new API for secure key verification
Harald Freudenberger [Wed, 15 Mar 2017 10:08:27 +0000 (11:08 +0100)]
s390/pkey: Introduce new API for secure key verification

User space needs some information about the secure key(s)
before actually invoking the pkey and/or paes funcionality.
This patch introduces a new ioctl API and in kernel API to
verify the the secure key blob and give back some
information about the key (type, bitsize, old MKVP).
Both APIs are described in detail in the header files
arch/s390/include/asm/pkey.h and arch/s390/include/uapi/asm/pkey.h.

Signed-off-by: Harald Freudenberger <freude@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/configs//zfcpdump_defconfig: Remove CONFIG_SCSI_SRP_ATTRS
Michael Holzheu [Thu, 16 Mar 2017 12:33:18 +0000 (13:33 +0100)]
s390/configs//zfcpdump_defconfig: Remove CONFIG_SCSI_SRP_ATTRS

Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/dump_stack: use control program identification string
Heiko Carstens [Sat, 11 Mar 2017 11:22:11 +0000 (12:22 +0100)]
s390/dump_stack: use control program identification string

If running within a level 3 hypervisor, the hypervisor provides a
SYSIB block which contains a control program indentifier string. Use
this string instead of the simple KVM and z/VM strings only. In case
of z/VM this provides addtional information: the z/VM version.

The new string looks similar to this:

Hardware name: IBM 2964 N96 702 (z/VM 6.4.0)

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/dump_stack: remove whitespace from arch description
Heiko Carstens [Sat, 11 Mar 2017 10:58:27 +0000 (11:58 +0100)]
s390/dump_stack: remove whitespace from arch description

The arch description provided for the "Hardware name:" contains lots
of extra whitespace due to the way the SYSIB contents are defined
(strings aren't zero terminated).
This looks a bit odd and therefore remove the extra whitespace
characters. This also gives the opportunity to add more information,
if required, without hitting the magic 80 characters per line limit.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/sysinfo: allow compiler warnings again
Heiko Carstens [Mon, 13 Mar 2017 14:45:54 +0000 (15:45 +0100)]
s390/sysinfo: allow compiler warnings again

Allow compiler warnings again for the sysinfo file. Compiler warnings
were disabled when the bogomips calculation with math-emu code was
introduced ("[S390] Calibrate delay and bogomips.").
Since that code is gone, we can enable warnings again.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/topology: fix typo in early topology code
Heiko Carstens [Mon, 13 Mar 2017 12:36:10 +0000 (13:36 +0100)]
s390/topology: fix typo in early topology code

Use MACHINE_FLAG_TOPOLOGY instead of MACHINE_HAS_TOPOLOGY when
clearing the bit that indicates if the machine provides topology
information (and if it should be used). Currently works anyway.

Fixes: 68cc795d1933 ("s390/topology: make "topology=off" parameter work")
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/topology: get rid of core mask array
Heiko Carstens [Thu, 9 Mar 2017 09:02:28 +0000 (10:02 +0100)]
s390/topology: get rid of core mask array

Use a single long value instead of a single element array to represent
the core mask. The array is a leftover from 32/31 bit code so we were
able to use bitops helper functions.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/bitops: remove outdated comment
Heiko Carstens [Thu, 9 Mar 2017 08:51:22 +0000 (09:51 +0100)]
s390/bitops: remove outdated comment

The s390 specific little endian bitop macros are gone since a long time.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: get rid of superfluous __GFP_REPEAT
Michal Hocko [Tue, 7 Mar 2017 15:48:40 +0000 (16:48 +0100)]
s390: get rid of superfluous __GFP_REPEAT

__GFP_REPEAT has a rather weak semantic but since it has been introduced
around 2.6.12 it has been ignored for low order allocations.

page_table_alloc then uses the flag for a single page allocation. This
means that this flag has never been actually useful here because it has
always been used only for PAGE_ALLOC_COSTLY requests.

An earlier attempt to remove the flag 10d58bf297e2 ("s390: get rid of
superfluous __GFP_REPEAT") has missed this one but the situation is very
same here.

Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/cpuinfo: show facilities as reported by stfle
Heiko Carstens [Sat, 25 Feb 2017 11:16:48 +0000 (12:16 +0100)]
s390/cpuinfo: show facilities as reported by stfle

Add a new line to /proc/cpuinfo which shows all available facilities
as reported by the stfle instruction:

> cat /proc/cpuinfo
...
facilities      : 0 1 2 3 4 6 7 ...
...

Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/bitops: add for_each_set_bit_inv helper
Heiko Carstens [Sat, 25 Feb 2017 11:16:31 +0000 (12:16 +0100)]
s390/bitops: add for_each_set_bit_inv helper

Same helper function like for_each_set_bit in generic code.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agoMerge branch 'guarded-storage' into 'features' to make merging with
Martin Schwidefsky [Wed, 22 Mar 2017 07:22:02 +0000 (08:22 +0100)]
Merge branch 'guarded-storage' into 'features' to make merging with
the KVM tree easier.

7 years agos390/pkey: Fix wrong handling of secure key with old MKVP
Harald Freudenberger [Wed, 15 Mar 2017 09:58:07 +0000 (10:58 +0100)]
s390/pkey: Fix wrong handling of secure key with old MKVP

When a secure key with an old Master Key Verification
Pattern was given to the pkey_findcard function, there was
no responsible card found because only the current MKVP of
each card was compared. With this fix also the old MKVP
values are considered and so a matching card able to handle
the key is reported back to the caller.

Signed-off-by: Harald Freudenberger <freude@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/smp: fix ipl from cpu with non-zero address
Heiko Carstens [Thu, 16 Mar 2017 10:02:36 +0000 (11:02 +0100)]
s390/smp: fix ipl from cpu with non-zero address

Commit af51160ebd3c ("s390/smp: initialize cpu_present_mask in
setup_arch") initializes the cpu_present_mask much earlier than
before. However the cpu detection code relies on the fact that iff
logical cpu 0 is marked present then also the corresponding physical
cpu address within the pcpu_devices array slot is valid.

Since commit 44fd22992cb7 ("[PATCH] Register the boot-cpu in the cpu
maps earlier") this assumption is not true anymore. The patch marks
logical cpu 0 as present in common code without that architecture code
had a chance to setup the logical to physical map.

With that change the cpu detection code assumes that the physical cpu
address of cpu 0 is also 0, which is not necessarily true.
Subsequently the physical cpu address of the ipl cpu will be mapped to
a different logical cpu. If that cpu is brought online later the ipl
cpu will send itself an initial cpu reset sigp signal. This in turn
completely resets the ipl cpu and the system stops working.

A dump of such a system looks like a "store status" has been
forgotten. But actually the kernel itself removed all traces which
would allow to easily tell what went wrong.

To fix this initialize the logical to physical cpu address already in
smp_setup_processor_id(). In addition remove the initialization of the
cpu_present_mask and cpu_online_mask for cpu 0, since that has already
been done. Also add a sanity check, just in case common code will be
changed again...

The problem can be easily reproduced within a z/VM guest:

> chcpu -d 0
> vmcp ipl

Fixes: af51160ebd3c ("s390/smp: initialize cpu_present_mask in setup_arch")
Reported-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390/decompressor: fix initrd corruption caused by bss clear
Marcelo Henrique Cerri [Mon, 13 Mar 2017 15:14:58 +0000 (12:14 -0300)]
s390/decompressor: fix initrd corruption caused by bss clear

Reorder the operations in decompress_kernel() to ensure initrd is moved
to a safe location before the bss section is zeroed.

During decompression bss can overlap with the initrd and this can
corrupt the initrd contents depending on the size of the compressed
kernel (which affects where the initrd is placed by the bootloader) and
the size of the bss section of the decompressor.

Also use the correct initrd size when checking for overlaps with
parmblock.

Fixes: 06c0dd72aea3 ([S390] fix boot failures with compressed kernels)
Cc: stable@vger.kernel.org
Reviewed-by: Joy Latten <joy.latten@canonical.com>
Reviewed-by: Vineetha HariPai <vineetha.hari.pai@canonical.com>
Signed-off-by: Marcelo Henrique Cerri <marcelo.cerri@canonical.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agos390: add a system call for guarded storage
Martin Schwidefsky [Tue, 26 Jan 2016 13:10:34 +0000 (14:10 +0100)]
s390: add a system call for guarded storage

This adds a new system call to enable the use of guarded storage for
user space processes. The system call takes two arguments, a command
and pointer to a guarded storage control block:

    s390_guarded_storage(int command, struct gs_cb *gs_cb);

The second argument is relevant only for the GS_SET_BC_CB command.

The commands in detail:

0 - GS_ENABLE
    Enable the guarded storage facility for the current task. The
    initial content of the guarded storage control block will be
    all zeros. After the enablement the user space code can use
    load-guarded-storage-controls instruction (LGSC) to load an
    arbitrary control block. While a task is enabled the kernel
    will save and restore the current content of the guarded
    storage registers on context switch.
1 - GS_DISABLE
    Disables the use of the guarded storage facility for the current
    task. The kernel will cease to save and restore the content of
    the guarded storage registers, the task specific content of
    these registers is lost.
2 - GS_SET_BC_CB
    Set a broadcast guarded storage control block. This is called
    per thread and stores a specific guarded storage control block
    in the task struct of the current task. This control block will
    be used for the broadcast event GS_BROADCAST.
3 - GS_CLEAR_BC_CB
    Clears the broadcast guarded storage control block. The guarded-
    storage control block is removed from the task struct that was
    established by GS_SET_BC_CB.
4 - GS_BROADCAST
    Sends a broadcast to all thread siblings of the current task.
    Every sibling that has established a broadcast guarded storage
    control block will load this control block and will be enabled
    for guarded storage. The broadcast guarded storage control block
    is used up, a second broadcast without a refresh of the stored
    control block with GS_SET_BC_CB will not have any effect.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
7 years agomm, swap: Remove WARN_ON_ONCE() in free_swap_slot()
Huang Ying [Mon, 20 Mar 2017 06:26:42 +0000 (14:26 +0800)]
mm, swap: Remove WARN_ON_ONCE() in free_swap_slot()

Before commit 452b94b8c8c7 ("mm/swap: don't BUG_ON() due to
uninitialized swap slot cache"), the following bug is reported,

  ------------[ cut here ]------------
  kernel BUG at mm/swap_slots.c:270!
  invalid opcode: 0000 [#1] SMP
  CPU: 5 PID: 1745 Comm: (sd-pam) Not tainted 4.11.0-rc1-00243-g24c534bb161b #1
  Hardware name: System manufacturer System Product Name/Z170-K, BIOS 1803 05/06/2016
  RIP: 0010:free_swap_slot+0xba/0xd0
  Call Trace:
   swap_free+0x36/0x40
   do_swap_page+0x360/0x6d0
   __handle_mm_fault+0x880/0x1080
   handle_mm_fault+0xd0/0x240
   __do_page_fault+0x232/0x4d0
   do_page_fault+0x20/0x70
   page_fault+0x22/0x30
  ---[ end trace aefc9ede53e0ab21 ]---

This is raised by the BUG_ON(!swap_slot_cache_initialized) in
free_swap_slot().  This is incorrect, because even if the swap slots
cache fails to be initialized, the swap should operate properly without
the swap slots cache.  And the use_swap_slot_cache check later in the
function will protect the uninitialized swap slots cache case.

In commit 452b94b8c8c7, the BUG_ON() is replaced by WARN_ON_ONCE().  In
the patch, the WARN_ON_ONCE() is removed too.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Tue, 21 Mar 2017 20:10:17 +0000 (13:10 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "Nine small fixes: the biggest is probably finally sorting out Kconfig
  issues with lpfc nvme.  There are some performance fixes for megaraid
  and hpsa and a static checker fix"

[ Johannes Thumshirn points out that there still seems to be more lpfc
  vs nvme config issues.  Oh well.   - Linus ]

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: lpfc: Finalize Kconfig options for nvme
  scsi: ufs: don't check unsigned type for a negative value
  scsi: hpsa: do not timeout reset operations
  scsi: hpsa: limit outstanding rescans
  scsi: hpsa: update check for logical volume status
  scsi: megaraid_sas: Driver version upgrade
  scsi: megaraid_sas: raid6 also require cpuSel check same as raid5
  scsi: megaraid_sas: add correct return type check for ldio hint logic for raid1
  scsi: megaraid_sas: enable intx only if msix request fails

7 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Tue, 21 Mar 2017 20:07:18 +0000 (13:07 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

Pull HID fixes from Jiri Kosina:

 - regression fixes for Wacom devices, from Aaron Armstrong Skomra and
   Ping Cheng

 - memory leak in hid-sony driver from Roderick Colenbrander

 - new device IDs support from Oscar Campos and Daniel Drake

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: wacom: generic: Wacom mouse is only provided for opaque tablets
  HID: corsair: Add driver Scimitar Pro RGB gaming mouse 1b1c:1b3e support to hid-corsair
  HID: corsair: support for K65-K70 Rapidfire and Scimitar Pro RGB
  HID: wacom: don't manually release resources for the EKR
  HID: wacom: Correct Intuos Pro 2 resolution
  HID: sony: Fix input device leak when connecting a DS4 twice using USB/BT
  HID: chicony: Add support for another ASUS Zen AiO keyboard

7 years agoMerge tag 'gpio-v4.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Tue, 21 Mar 2017 20:01:53 +0000 (13:01 -0700)]
Merge tag 'gpio-v4.11-2' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fixes from Linus Walleij:
 "Here is the first set of GPIO fixes for 4.11. It was delayed a bit
  beacuse I was chicken when linux-next was not rotating last week.

  This hits the ST serial driver in drivers/tty/serial and that has an
  ACK from Greg, he suggested to keep the old GPIO fwnode API around to
  smoothen things in the merge Windod and those have now served their
  purpose so we take them out and convert the last driver to the new
  API.

  Apart from that it's fixes as usual.

  Summary:

   - set the parent on the Altera A10SR driver, also fix high level
     IRQs.

   - fix error path on the mockup driver.

   - compilation noise about unused functions fixed.

   - fix missed interrupts on the MCP23S08 expander, this is also tagged
     for stable.

   - retire the interrim helpers devm_get_gpiod_from_child() used to
     smoothen merging in the merge window"

* tag 'gpio-v4.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpio:mcp23s08 Fixed missing interrupts
  serial: st-asc: Use new GPIOD API to obtain RTS pin
  gpio: altera: Use handle_level_irq when configured as a level_high
  gpio: xgene: mark PM functions as __maybe_unused
  gpio: mockup: return -EFAULT if copy_from_user() fails
  gpio: altera-a10sr: Set gpio_chip parent property

7 years agoMerge tag 'rproc-v4.11-fixes' of git://github.com/andersson/remoteproc
Linus Torvalds [Tue, 21 Mar 2017 19:54:12 +0000 (12:54 -0700)]
Merge tag 'rproc-v4.11-fixes' of git://github.com/andersson/remoteproc

Pull remoteproc fix from Bjorn Andersson:
 "This fixes a Kbuild dependency issue related to the Qualcomm
  remoteproc drivers"

* tag 'rproc-v4.11-fixes' of git://github.com/andersson/remoteproc:
  remoteproc: qcom: fix QCOM_SMD dependencies

7 years agoMerge tag 'for-f2fs-4.11-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeu...
Linus Torvalds [Tue, 21 Mar 2017 19:27:06 +0000 (12:27 -0700)]
Merge tag 'for-f2fs-4.11-rc4' of git://git./linux/kernel/git/jaegeuk/f2fs

Pull f2fs fixes from Jaegeuk Kim:

 - fix performance regression reported by lkp-rebot

 - fix potential data lost after power-cut due to SSR reallocation

* tag 'for-f2fs-4.11-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs:
  f2fs: combine nat_bits and free_nid_bitmap cache
  f2fs: skip scanning free nid bitmap of full NAT blocks
  f2fs: use __set{__clear}_bit_le
  f2fs: declare static functions
  f2fs: don't overwrite node block by SSR

7 years agoHID: wacom: generic: Wacom mouse is only provided for opaque tablets
Ping Cheng [Wed, 15 Mar 2017 00:08:16 +0000 (17:08 -0700)]
HID: wacom: generic: Wacom mouse is only provided for opaque tablets

Commit f85c9dc ("Support tool ID and additional tool types") introduced mouse
and lens cursor tools to generic codepath, which covers both display (direct)
and opaque tablets (indirect devices). However, mouse and lens cursor tools are
only provided for opaque tablets. This patch ignores mouse and lens cursor tools
if the device is a display tablet.

Signed-off-by: Ping Cheng <ping.cheng@wacom.com>
Reviewed-by: Jason Gerecke <jason.gerecke@wacom.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
7 years agoHID: corsair: Add driver Scimitar Pro RGB gaming mouse 1b1c:1b3e support to hid-corsair
Oscar Campos [Mon, 6 Mar 2017 21:02:39 +0000 (21:02 +0000)]
HID: corsair: Add driver Scimitar Pro RGB gaming mouse 1b1c:1b3e support to hid-corsair

This mouse sold by Corsair as Scimitar PRO RGB defines two consecutive
Logical Minimum items in its Application (Consumer.0001) report making
it non parseable. This patch fixes the report descriptor overriding
byte 77 in rdesc from 0x16 (Logical Minimum with 16 bits value) to 0x26
(Logical Maximum with 16 bits value).

Signed-off-by: Oscar Campos <oscar.campos@member.fsf.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
7 years agoHID: corsair: support for K65-K70 Rapidfire and Scimitar Pro RGB
Oscar Campos [Fri, 10 Feb 2017 18:23:00 +0000 (18:23 +0000)]
HID: corsair: support for K65-K70 Rapidfire and Scimitar Pro RGB

Add quirks for several corsair gaming devices to avoid long delays on
report initialization

Supported devices:

 - Corsair K65RGB Rapidfire Gaming Keyboard
 - Corsair K70RGB Rapidfire Gaming Keyboard
 - Corsair Scimitar Pro RGB Gaming Mouse

Signed-off-by: Oscar Campos <oscar.campos@member.fsf.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
7 years agoHID: wacom: don't manually release resources for the EKR
Aaron Armstrong Skomra [Mon, 6 Mar 2017 18:54:58 +0000 (10:54 -0800)]
HID: wacom: don't manually release resources for the EKR

Commit 5b779fc introduces the manual release of resources in wacom_remove() as
an addition to the driver's use of devm.  The EKR resources can only be
released through wacom_remote_destroy_one() so we skip the manual release for
it.

Fixes: 5b779fc ("HID: wacom: release the resources before leaving despite devm")
Signed-off-by: Aaron Armstrong Skomra <skomra@gmail.com>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
7 years agoHID: wacom: Correct Intuos Pro 2 resolution
Aaron Armstrong Skomra [Mon, 6 Mar 2017 18:54:57 +0000 (10:54 -0800)]
HID: wacom: Correct Intuos Pro 2 resolution

The features struct for the second gen Intuos Pro uses the wrong constant for
the resolution. This fix is for commit 4922cd2.

Fixes: 4922cd2 ("HID: wacom: Support 2nd-gen Intuos Pro's Bluetooth classic interface")
Signed-off-by: Aaron Armstrong Skomra <skomra@gmail.com>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
7 years agoremoteproc: qcom: fix QCOM_SMD dependencies
Arnd Bergmann [Mon, 13 Mar 2017 16:36:25 +0000 (17:36 +0100)]
remoteproc: qcom: fix QCOM_SMD dependencies

qcom_smd_register_edge() is provided by either QCOM_SMD or RPMSG_QCOM_SMD,
and if both of them are disabled, it does nothing.

The check for the PIL drivers however only checks for QCOM_SMD, so it breaks
with QCOM_SMD=n && RPMSG_QCOM_SMD=m:

drivers/remoteproc/built-in.o: In function `smd_subdev_remove':
qcom_wcnss_iris.c:(.text+0x231c): undefined reference to `qcom_smd_unregister_edge'
drivers/remoteproc/built-in.o: In function `smd_subdev_probe':
qcom_wcnss_iris.c:(.text+0x2344): undefined reference to `qcom_smd_register_edge'
drivers/remoteproc/built-in.o: In function `smd_subdev_probe':
qcom_q6v5_pil.c:(.text+0x3538): undefined reference to `qcom_smd_register_edge'
qcom_q6v5_pil.c:(.text+0x3538): relocation truncated to fit: R_AARCH64_CALL26 against undefined symbol `qcom_smd_register_edge'

This clarifies the Kconfig dependency.

Fixes: 4b48921a8f74 ("remoteproc: qcom: Use common SMD edge handler")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
7 years agof2fs: combine nat_bits and free_nid_bitmap cache
Chao Yu [Wed, 8 Mar 2017 12:07:49 +0000 (20:07 +0800)]
f2fs: combine nat_bits and free_nid_bitmap cache

Both nat_bits cache and free_nid_bitmap cache provide same functionality
as a intermediate cache between free nid cache and disk, but with
different granularity of indicating free nid range, and different
persistence policy. nat_bits cache provides better persistence ability,
and free_nid_bitmap provides better granularity.

In this patch we combine advantage of both caches, so finally policy of
the intermediate cache would be:
- init: load free nid status from nat_bits into free_nid_bitmap
- lookup: scan free_nid_bitmap before load NAT blocks
- update: update free_nid_bitmap in real-time
- persistence: udpate and persist nat_bits in checkpoint

This patch also resolves performance regression reported by lkp-robot.

commit:
  4ac912427c4214d8031d9ad6fbc3bc75e71512df ("f2fs: introduce free nid bitmap")
  d00030cf9cd0bb96fdccc41e33d3c91dcbb672ba ("f2fs: use __set{__clear}_bit_le")
  1382c0f3f9d3f936c8bc42ed1591cf7a593ef9f7 ("f2fs: combine nat_bits and free_nid_bitmap cache")

4ac912427c4214d8 d00030cf9cd0bb96fdccc41e33 1382c0f3f9d3f936c8bc42ed15
---------------- -------------------------- --------------------------
         %stddev     %change         %stddev     %change         %stddev
             \          |                \          |                \
     77863 ±  0%      +2.1%      79485 ±  1%     +50.8%     117404 ±  0%  aim7.jobs-per-min
    231.63 ±  0%      -2.0%     227.01 ±  1%     -33.6%     153.80 ±  0%  aim7.time.elapsed_time
    231.63 ±  0%      -2.0%     227.01 ±  1%     -33.6%     153.80 ±  0%  aim7.time.elapsed_time.max
    896604 ±  0%      -0.8%     889221 ±  3%     -20.2%     715260 ±  1%  aim7.time.involuntary_context_switches
      2394 ±  1%      +4.6%       2503 ±  1%      +3.7%       2481 ±  2%  aim7.time.maximum_resident_set_size
      6240 ±  0%      -1.5%       6145 ±  1%     -14.1%       5360 ±  1%  aim7.time.system_time
   1111357 ±  3%      +1.9%    1132509 ±  2%      -6.2%    1041932 ±  2%  aim7.time.voluntary_context_switches
...

Signed-off-by: Chao Yu <yuchao0@huawei.com>
Tested-by: Xiaolong Ye <xiaolong.ye@intel.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
7 years agof2fs: skip scanning free nid bitmap of full NAT blocks
Chao Yu [Wed, 1 Mar 2017 09:09:07 +0000 (17:09 +0800)]
f2fs: skip scanning free nid bitmap of full NAT blocks

This patch adds to account free nids for each NAT blocks, and while
scanning all free nid bitmap, do check count and skip lookuping in
full NAT block.

Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
7 years agof2fs: use __set{__clear}_bit_le
Jaegeuk Kim [Tue, 7 Mar 2017 22:11:06 +0000 (14:11 -0800)]
f2fs: use __set{__clear}_bit_le

This patch uses __set{__clear}_bit_le for highter speed.

Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
7 years agof2fs: declare static functions
Jaegeuk Kim [Fri, 10 Mar 2017 17:39:57 +0000 (09:39 -0800)]
f2fs: declare static functions

This is to avoid build warning reported by kbuild test robot.

Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
7 years agof2fs: don't overwrite node block by SSR
Jaegeuk Kim [Mon, 6 Mar 2017 19:59:56 +0000 (11:59 -0800)]
f2fs: don't overwrite node block by SSR

This patch fixes that SSR can overwrite previous warm node block consisting of
a node chain since the last checkpoint.

Fixes: 5b6c6be2d878 ("f2fs: use SSR for warm node as well")
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
7 years agoLinux 4.11-rc3
Linus Torvalds [Mon, 20 Mar 2017 02:09:39 +0000 (19:09 -0700)]
Linux 4.11-rc3

7 years agomm/swap: don't BUG_ON() due to uninitialized swap slot cache
Linus Torvalds [Mon, 20 Mar 2017 02:00:47 +0000 (19:00 -0700)]
mm/swap: don't BUG_ON() due to uninitialized swap slot cache

This BUG_ON() triggered for me once at shutdown, and I don't see a
reason for the check.  The code correctly checks whether the swap slot
cache is usable or not, so an uninitialized swap slot cache is not
actually problematic afaik.

I've temporarily just switched the BUG_ON() to a WARN_ON_ONCE(), since
I'm not sure why that seemingly pointless check was there.  I suspect
the real fix is to just remove it entirely, but for now we'll warn about
it but not bring the machine down.

Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge tag 'powerpc-4.11-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Mon, 20 Mar 2017 01:49:28 +0000 (18:49 -0700)]
Merge tag 'powerpc-4.11-5' of git://git./linux/kernel/git/powerpc/linux

Pull more powerpc fixes from Michael Ellerman:
 "A couple of minor powerpc fixes for 4.11:

   - wire up statx() syscall

   - don't print a warning on memory hotplug when HPT resizing isn't
     available

  Thanks to: David Gibson, Chandan Rajendra"

* tag 'powerpc-4.11-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/pseries: Don't give a warning when HPT resizing isn't available
  powerpc: Wire up statx() syscall

7 years agoMerge branch 'parisc-4.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller...
Linus Torvalds [Mon, 20 Mar 2017 01:11:13 +0000 (18:11 -0700)]
Merge branch 'parisc-4.11-2' of git://git./linux/kernel/git/deller/parisc-linux

Pull parisc fixes from Helge Deller:

 - Mikulas Patocka added support for R_PARISC_SECREL32 relocations in
   modules with CONFIG_MODVERSIONS.

 - Dave Anglin optimized the cache flushing for vmap ranges.

 - Arvind Yadav provided a fix for a potential NULL pointer dereference
   in the parisc perf code (and some code cleanups).

 - I wired up the new statx system call, fixed some compiler warnings
   with the access_ok() macro and fixed shutdown code to really halt a
   system at shutdown instead of crashing & rebooting.

* 'parisc-4.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Fix system shutdown halt
  parisc: perf: Fix potential NULL pointer dereference
  parisc: Avoid compiler warnings with access_ok()
  parisc: Wire up statx system call
  parisc: Optimize flush_kernel_vmap_range and invalidate_kernel_vmap_range
  parisc: support R_PARISC_SECREL32 relocation in modules

7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
Linus Torvalds [Mon, 20 Mar 2017 01:06:31 +0000 (18:06 -0700)]
Merge git://git./linux/kernel/git/nab/target-pending

Pull SCSI target fixes from Nicholas Bellinger:
 "The bulk of the changes are in qla2xxx target driver code to address
  various issues found during Cavium/QLogic's internal testing (stable
  CC's included), along with a few other stability and smaller
  miscellaneous improvements.

  There are also a couple of different patch sets from Mike Christie,
  which have been a result of his work to use target-core ALUA logic
  together with tcm-user backend driver.

  Finally, a patch to address some long standing issues with
  pass-through SCSI export of TYPE_TAPE + TYPE_MEDIUM_CHANGER devices,
  which will make folks using physical (or virtual) magnetic tape happy"

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (28 commits)
  qla2xxx: Update driver version to 9.00.00.00-k
  qla2xxx: Fix delayed response to command for loop mode/direct connect.
  qla2xxx: Change scsi host lookup method.
  qla2xxx: Add DebugFS node to display Port Database
  qla2xxx: Use IOCB interface to submit non-critical MBX.
  qla2xxx: Add async new target notification
  qla2xxx: Export DIF stats via debugfs
  qla2xxx: Improve T10-DIF/PI handling in driver.
  qla2xxx: Allow relogin to proceed if remote login did not finish
  qla2xxx: Fix sess_lock & hardware_lock lock order problem.
  qla2xxx: Fix inadequate lock protection for ABTS.
  qla2xxx: Fix request queue corruption.
  qla2xxx: Fix memory leak for abts processing
  qla2xxx: Allow vref count to timeout on vport delete.
  tcmu: Convert cmd_time_out into backend device attribute
  tcmu: make cmd timeout configurable
  tcmu: add helper to check if dev was configured
  target: fix race during implicit transition work flushes
  target: allow userspace to set state to transitioning
  target: fix ALUA transition timeout handling
  ...

7 years agoMerge branch 'libnvdimm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdim...
Linus Torvalds [Sun, 19 Mar 2017 22:45:02 +0000 (15:45 -0700)]
Merge branch 'libnvdimm-fixes' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull device-dax fixes from Dan Williams:
 "The device-dax driver was not being careful to handle falling back to
  smaller fault-granularity sizes.

  The driver already fails fault attempts that are smaller than the
  device's alignment, but it also needs to handle the cases where a
  larger page mapping could be established. For simplicity of the
  immediate fix the implementation just signals VM_FAULT_FALLBACK until
  fault-size == device-alignment.

  One fix is for -stable to address pmd-to-pte fallback from the
  original implementation, another fix is for the new (introduced in
  4.11-rc1) pud-to-pmd regression, and a typo fix comes along for the
  ride.

  These have received a build success notification from the kbuild
  robot"

* 'libnvdimm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  device-dax: fix debug output typo
  device-dax: fix pud fault fallback handling
  device-dax: fix pmd/pte fault fallback handling

7 years agoqla2xxx: Update driver version to 9.00.00.00-k
Himanshu Madhani [Wed, 15 Mar 2017 16:48:56 +0000 (09:48 -0700)]
qla2xxx: Update driver version to 9.00.00.00-k

Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
signed-off-by: Giridhar Malavali <giridhar.malavali@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Fix delayed response to command for loop mode/direct connect.
Quinn Tran [Wed, 15 Mar 2017 16:48:55 +0000 (09:48 -0700)]
qla2xxx: Fix delayed response to command for loop mode/direct connect.

Current driver wait for FW to be in the ready state before
processing in-coming commands. For Arbitrated Loop or
Point-to- Point (not switch), FW Ready state can take a while.
FW will transition to ready state after all Nports have been
logged in. In the mean time, certain initiators have completed
the login and starts IO. Driver needs to start processing all
queues if FW is already started.

Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Change scsi host lookup method.
Quinn Tran [Wed, 15 Mar 2017 16:48:54 +0000 (09:48 -0700)]
qla2xxx: Change scsi host lookup method.

For target mode, when new scsi command arrive, driver first performs
a look up of the SCSI Host. The current look up method is based on
the ALPA portion of the NPort ID. For Cisco switch, the ALPA can
not be used as the index. Instead, the new search method is based
on the full value of the Nport_ID via btree lib.

Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Add DebugFS node to display Port Database
Himanshu Madhani [Wed, 15 Mar 2017 16:48:53 +0000 (09:48 -0700)]
qla2xxx: Add DebugFS node to display Port Database

Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Giridhar Malavali <giridhar.malavali@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Use IOCB interface to submit non-critical MBX.
Quinn Tran [Wed, 15 Mar 2017 16:48:52 +0000 (09:48 -0700)]
qla2xxx: Use IOCB interface to submit non-critical MBX.

The Mailbox interface is currently over subscribed. We like
to reserve the Mailbox interface for the chip managment and
link initialization. Any non essential Mailbox command will
be routed through the IOCB interface. The IOCB interface is
able to absorb more commands.

Following commands are being routed through IOCB interface

- Get ID List (007Ch)
- Get Port DB (0064h)
- Get Link Priv Stats (006Dh)

Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Add async new target notification
Quinn Tran [Wed, 15 Mar 2017 16:48:51 +0000 (09:48 -0700)]
qla2xxx: Add async new target notification

Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Export DIF stats via debugfs
Anil Gurumurthy [Wed, 15 Mar 2017 16:48:50 +0000 (09:48 -0700)]
qla2xxx: Export DIF stats via debugfs

Signed-off-by: Anil Gurumurthy <anil.gurumurthy@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Improve T10-DIF/PI handling in driver.
Quinn Tran [Wed, 15 Mar 2017 16:48:49 +0000 (09:48 -0700)]
qla2xxx: Improve T10-DIF/PI handling in driver.

Add routines to support T10 DIF tag.

Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Anil Gurumurthy <anil.gurumurthy@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Allow relogin to proceed if remote login did not finish
Quinn Tran [Wed, 15 Mar 2017 16:48:48 +0000 (09:48 -0700)]
qla2xxx: Allow relogin to proceed if remote login did not finish

If the remote port have started the login process, then the
PLOGI and PRLI should be back to back. Driver will allow
the remote port to complete the process. For the case where
the remote port decide to back off from sending PRLI, this
local port sets an expiration timer for the PRLI. Once the
expiration time passes, the relogin retry logic is allowed
to go through and perform login with the remote port.

Signed-off-by: Quinn Tran <quinn.tran@qlogic.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Fix sess_lock & hardware_lock lock order problem.
Quinn Tran [Wed, 15 Mar 2017 16:48:47 +0000 (09:48 -0700)]
qla2xxx: Fix sess_lock & hardware_lock lock order problem.

The main lock that needs to be held for CMD or TMR submission
to upper layer is the sess_lock. The sess_lock is used to
serialize cmd submission and session deletion. The addition
of hardware_lock being held is not necessary. This patch removes
hardware_lock dependency from CMD/TMR submission.

Use hardware_lock only for error response in this case.

Path1
       CPU0                    CPU1
       ----                    ----
  lock(&(&ha->tgt.sess_lock)->rlock);
                               lock(&(&ha->hardware_lock)->rlock);
                               lock(&(&ha->tgt.sess_lock)->rlock);
  lock(&(&ha->hardware_lock)->rlock);

Path2/deadlock
*** DEADLOCK ***
Call Trace:
dump_stack+0x85/0xc2
print_circular_bug+0x1e3/0x250
__lock_acquire+0x1425/0x1620
lock_acquire+0xbf/0x210
_raw_spin_lock_irqsave+0x53/0x70
qlt_sess_work_fn+0x21d/0x480 [qla2xxx]
process_one_work+0x1f4/0x6e0

Cc: <stable@vger.kernel.org>
Cc: Bart Van Assche <Bart.VanAssche@sandisk.com>
Reported-by: Bart Van Assche <Bart.VanAssche@sandisk.com>
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Fix inadequate lock protection for ABTS.
Quinn Tran [Wed, 15 Mar 2017 16:48:46 +0000 (09:48 -0700)]
qla2xxx: Fix inadequate lock protection for ABTS.

Normally, ABTS is sent to Target Core as Task MGMT command.
In the case of error, qla2xxx needs to send response, hardware_lock
is required to prevent request queue corruption.

Cc: <stable@vger.kernel.org>
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Fix request queue corruption.
Quinn Tran [Wed, 15 Mar 2017 16:48:45 +0000 (09:48 -0700)]
qla2xxx: Fix request queue corruption.

When FW notify driver or driver detects low FW resource,
driver tries to send out Busy SCSI Status to tell Initiator
side to back off. During the send process, the lock was not held.

Cc: <stable@vger.kernel.org>
Signed-off-by: Quinn Tran <quinn.tran@qlogic.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Fix memory leak for abts processing
Quinn Tran [Wed, 15 Mar 2017 16:48:44 +0000 (09:48 -0700)]
qla2xxx: Fix memory leak for abts processing

Cc: <stable@vger.kernel.org>
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
7 years agoqla2xxx: Allow vref count to timeout on vport delete.
Joe Carnuccio [Wed, 15 Mar 2017 16:48:43 +0000 (09:48 -0700)]
qla2xxx: Allow vref count to timeout on vport delete.

Cc: <stable@vger.kernel.org>
Signed-off-by: Joe Carnuccio <joe.carnuccio@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>