Users of qdio buffers employ different strategies to manage these
buffers. The qeth driver uses huge contiguous buffers which leads
to high order allocations with all their downsides.
This patch provides helpers to allocate, free, and reset arrays of
qdio buffers using non contiguous pages.
Reviewed-by: Martin Peschke <mpeschke@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
To ungroup and deregister the group device always use the
ccwgroup_ungroup wrapper.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Since commit 0b60f9ead5
"s390: use device_remove_file_self() instead of device_schedule_callback()"
the return code of ccwgroup_ungroup_store is uninitialized. Make
sure the rc is always initialized.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Commit 0b60f9ead5
"s390: use device_remove_file_self() instead of device_schedule_callback()"
changed ccwgroup to use an extra work queue instead of
device_schedule_callback. This function obtained an extra device
reference for its async work which is missing in the new implementation
and results in a "freeing memory with a lock still held" BUG. Fix
this by obtaining an extra reference for the async work.
Reported-by: Stefan Raspl <raspl@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Keep the per-device dbf entries until module is removed, with
proper error checking for debug feature setup.
Signed-off-by: Stefan Raspl <raspl@linux.vnet.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Currently we set the device name at the time we call device_add after
we receive the interrupt for the first I/O. When something is not working
as expected during that first I/O (e.g. we don't receive an interrupt) we
print a message including the device name which has not yet been
initialized.
Set the device name after calling device_initialize (prior to starting
the first I/O) so that we have the name present if some unexpected error
occurs during that first I/O.
Reported-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reported-by: Jason J. Herne <jjherne@linux.vnet.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
airq_iv_(alloc|free) is called by some users with interrupts enabled
and by some with interrupts disabled which leads to the following
lockdep warning:
[ INFO: possible irq lock inversion dependency detected ]
3.14.0-15249-gbf29b7b-dirty #25 Not tainted
---------------------------------------------------------
insmod/2108 just changed the state of lock:
(&(&iv->lock)->rlock){+.....}, at: [<000000000046ee3e>] airq_iv_alloc+0x62/0x228
but this lock was taken by another, HARDIRQ-READ-safe lock in the past:
(&info->lock){.-.-..}
and interrupts could create inverse lock ordering between them.
other info that might help us debug this:
Possible interrupt unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&(&iv->lock)->rlock);
local_irq_disable();
lock(&info->lock);
lock(&(&iv->lock)->rlock);
<Interrupt>
lock(&info->lock);
*** DEADLOCK ***
Although this is a false alarm (since each airq user consistently
calls these functions from the same context) fix this by ensuring
that interrupts are disabled when the airq lock is held.
Reported-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
account_sbals is called by get_inbound_buffer_frontier and
get_outbound_buffer_frontier with 'count' value > 0 so we can safely
convert shift loop to ilog2.
Cc: Joe Perches <joe@perches.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
On systems where a ccw based console device is used a lockdep false alarm
could be triggered when a device driver calls printk while holding a
subchannels lock (e.g. in it's irq handler). Since this is valid behavior
fix this by introducing a separate lock class for the console subchannels
lock.
The lockdep warning was revealed by "printk: enable interrupts before calling
console_trylock_for_printk()" which changed console_unlock() to be called with
lockdep enabled.
[ INFO: possible recursive locking detected ]
3.15.0-rc5-next-20140520 #1 Not tainted
---------------------------------------------
ccwgroup/2239 is trying to acquire lock:
(&(sch->lock)->rlock){-.-...}, at: [<0000000000642a52>] raw3215_write+0x52/0x200
but task is already holding lock:
(&(sch->lock)->rlock){-.-...}, at: [<00000000005fd160>] do_cio_interrupt+0x60/0x108
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&(sch->lock)->rlock);
lock(&(sch->lock)->rlock);
*** DEADLOCK ***
May be due to missing lock nesting notation
8 locks held by ccwgroup/2239:
stack backtrace:
CPU: 3 PID: 2239 Comm: ccwgroup Not tainted 3.15.0-rc5-next-20140520 #1
0000000036fab518 0000000036fab528 0000000000000002 0000000000000000
0000000036fab5b8 0000000036fab530 0000000036fab530 00000000001116e8
0000000000000000 0000000000986ec4 00000000009701b6 000000000000000b
0000000036fab578 0000000036fab518 0000000000000000 0000000000000000
0000000000000000 00000000001116e8 0000000036fab518 0000000036fab578
Call Trace:
([<0000000000111626>] show_trace+0x14e/0x158)
[<000000000011169a>] show_stack+0x6a/0xe8
[<00000000007c6e72>] dump_stack+0x82/0xb0
[<00000000001a95f2>] validate_chain.isra.37+0xa4a/0xbb0
[<00000000001acaca>] __lock_acquire+0x4da/0xcd0
[<00000000001ada1a>] lock_acquire+0xba/0x218
[<00000000007cd634>] _raw_spin_lock_irqsave+0x6c/0xb8
[<0000000000642a52>] raw3215_write+0x52/0x200
[<0000000000643d16>] con3215_write+0x76/0xf8
[<00000000001bd87a>] call_console_drivers.constprop.25+0xfa/0x210
[<00000000001be0b0>] console_unlock+0x3e0/0x4e8
[<00000000001be450>] vprintk_emit+0x298/0x6e0
[<00000000005aa210>] dev_vprintk_emit+0xe0/0x1a8
[<00000000005aa320>] dev_printk_emit+0x48/0x50
[<00000000005aa390>] __dev_printk+0x68/0xb0
[<00000000005aa7c2>] _dev_info+0x62/0x70
[<0000000000657bf0>] qeth_l2_send_setmac_cb+0xd0/0x190
[<0000000000651a1e>] qeth_send_control_data_cb+0x3a6/0x6a8
[<0000000000655546>] qeth_irq+0x1a6/0xac0
[<000000000060a0ac>] ccw_device_call_handler+0xa4/0xc0
[<0000000000608b62>] ccw_device_irq+0x5a/0x190
[<00000000005fd1ca>] do_cio_interrupt+0xca/0x108
[<00000000001c0a2e>] handle_irq_event_percpu+0x5e/0x378
[<00000000001c46fc>] handle_percpu_irq+0x6c/0x98
[<00000000001c0066>] generic_handle_irq+0x46/0x68
[<000000000010b5b6>] do_IRQ+0x5e/0x88
[<00000000007cf304>] io_call+0x6/0x20
[<000000000064c63a>] qeth_send_control_data+0x322/0x570
([<000000000064c50e>] qeth_send_control_data+0x1f6/0x570)
[<0000000000651db2>] qeth_send_ipa_cmd+0x92/0x120
[<000000000065b310>] __qeth_l2_set_online+0x170/0xaa8
[<000000000060ebb6>] ccwgroup_set_online+0x56/0x90
[<000000000060ef96>] ccwgroup_online_store+0xd6/0xe0
[<000000000033d11a>] kernfs_fop_write+0x10a/0x188
[<00000000002bbd00>] vfs_write+0x98/0x1c0
[<00000000002bc8a0>] SyS_write+0x60/0xd0
[<00000000007cee3a>] sysc_nr_ok+0x22/0x28
[<000003fffd0c3f28>] 0x3fffd0c3f28
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Remove the 96-byte irb array from the lowcore and create a per-cpu
variable instead. That way we will pick up any change in the definition
of the struct irb automatically.
Acked-By: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Get rid of a useless assignment during argument evaluation.
No functional change.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Fix multiple definitions of struct channel_path_desc by moving it
to asm/chpid.h . Also change ccw_device_get_chp_desc to use proper
types.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Using a notification type mask for the store event information chsc
is unsupported on some firmware levels. Retry SEI with that mask set
to zero (which is the old way of requesting only channel subsystem
related events).
Cc: <stable@vger.kernel.org>
Reported-and-tested-by: Stefan Haberland <stefan.haberland@de.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Here's the big driver core / sysfs update for 3.15-rc1.
Lots of kernfs updates to make it useful for other subsystems, and a few
other tiny driver core patches.
All have been in linux-next for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iEYEABECAAYFAlM7A0wACgkQMUfUDdst+ynJNACfZlY+KNKIhNFt1OOW8rQfSZzy
1PYAnjYuOoly01JlPrpJD5b4TdxaAq71
=GVUg
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core and sysfs updates from Greg KH:
"Here's the big driver core / sysfs update for 3.15-rc1.
Lots of kernfs updates to make it useful for other subsystems, and a
few other tiny driver core patches.
All have been in linux-next for a while"
* tag 'driver-core-3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (42 commits)
Revert "sysfs, driver-core: remove unused {sysfs|device}_schedule_callback_owner()"
kernfs: cache atomic_write_len in kernfs_open_file
numa: fix NULL pointer access and memory leak in unregister_one_node()
Revert "driver core: synchronize device shutdown"
kernfs: fix off by one error.
kernfs: remove duplicate dir.c at the top dir
x86: align x86 arch with generic CPU modalias handling
cpu: add generic support for CPU feature based module autoloading
sysfs: create bin_attributes under the requested group
driver core: unexport static function create_syslog_header
firmware: use power efficient workqueue for unloading and aborting fw load
firmware: give a protection when map page failed
firmware: google memconsole driver fixes
firmware: fix google/gsmi duplicate efivars_sysfs_init()
drivers/base: delete non-required instances of include <linux/init.h>
kernfs: fix kernfs_node_from_dentry()
ACPI / platform: drop redundant ACPI_HANDLE check
kernfs: fix hash calculation in kernfs_rename_ns()
kernfs: add CONFIG_KERNFS
sysfs, kobject: add sysfs wrapper for kernfs_enable_ns()
...
Pull irq code updates from Thomas Gleixner:
"The irq department proudly presents:
- Another tree wide sweep of irq infrastructure abuse. Clear winner
of the trainwreck engineering contest was:
#include "../../../kernel/irq/settings.h"
- Tree wide update of irq_set_affinity() callbacks which miss a cpu
online check when picking a single cpu out of the affinity mask.
- Tree wide consolidation of interrupt statistics.
- Updates to the threaded interrupt infrastructure to allow explicit
wakeup of the interrupt thread and a variant of synchronize_irq()
which synchronizes only the hard interrupt handler. Both are
needed to replace the homebrewn thread handling in the mmc/sdhci
code.
- New irq chip callbacks to allow proper support for GPIO based irqs.
The GPIO based interrupts need to request/release GPIO resources
from request/free_irq.
- A few new ARM interrupt chips. No revolutionary new hardware, just
differently wreckaged variations of the scheme.
- Small improvments, cleanups and updates all over the place"
I was hoping that that trainwreck engineering contest was a April Fools'
joke. But no.
* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (68 commits)
irqchip: sun7i/sun6i: Disable NMI before registering the handler
ARM: sun7i/sun6i: dts: Fix IRQ number for sun6i NMI controller
ARM: sun7i/sun6i: irqchip: Update the documentation
ARM: sun7i/sun6i: dts: Add NMI irqchip support
ARM: sun7i/sun6i: irqchip: Add irqchip driver for NMI controller
genirq: Export symbol no_action()
arm: omap: Fix typo in ams-delta-fiq.c
m68k: atari: Fix the last kernel_stat.h fallout
irqchip: sun4i: Simplify sun4i_irq_ack
irqchip: sun4i: Use handle_fasteoi_irq for all interrupts
genirq: procfs: Make smp_affinity values go+r
softirq: Add linux/irq.h to make it compile again
m68k: amiga: Add linux/irq.h to make it compile again
irqchip: sun4i: Don't ack IRQs > 0, fix acking of IRQ 0
irqchip: sun4i: Fix a comment about mask register initialization
irqchip: sun4i: Fix irq 0 not working
genirq: Add a new IRQCHIP_EOI_THREADED flag
genirq: Document IRQCHIP_ONESHOT_SAFE flag
ARM: sunxi: dt: Convert to the new irq controller compatibles
irqchip: sunxi: Change compatibles
...
Pull timer updates from Ingo Molnar:
"The main purpose is to fix a full dynticks bug related to
virtualization, where steal time accounting appears to be zero in
/proc/stat even after a few seconds of competing guests running busy
loops in a same host CPU. It's not a regression though as it was
there since the beginning.
The other commits are preparatory work to fix the bug and various
cleanups"
* 'timers-nohz-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
arch: Remove stub cputime.h headers
sched: Remove needless round trip nsecs <-> tick conversion of steal time
cputime: Fix jiffies based cputime assumption on steal accounting
cputime: Bring cputime -> nsecs conversion
cputime: Default implementation of nsecs -> cputime conversion
cputime: Fix nsecs_to_cputime() return type cast
Pull s390 updates from Martin Schwidefsky:
"There are two memory management related changes, the CMMA support for
KVM to avoid swap-in of freed pages and the split page table lock for
the PMD level. These two come with common code changes in mm/.
A fix for the long standing theoretical TLB flush problem, this one
comes with a common code change in kernel/sched/.
Another set of changes is Heikos uaccess work, included is the initial
set of patches with more to come.
And fixes and cleanups as usual"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (36 commits)
s390/con3270: optionally disable auto update
s390/mm: remove unecessary parameter from pgste_ipte_notify
s390/mm: remove unnecessary parameter from gmap_do_ipte_notify
s390/mm: fixing comment so that parameter name match
s390/smp: limit number of cpus in possible cpu mask
hypfs: Add clarification for "weight_min" attribute
s390: update defconfigs
s390/ptrace: add support for PTRACE_SINGLEBLOCK
s390/perf: make print_debug_cf() static
s390/topology: Remove call to update_cpu_masks()
s390/compat: remove compat exec domain
s390: select CONFIG_TTY for use of tty in unconditional keyboard driver
s390/appldata_os: fix cpu array size calculation
s390/checksum: remove memset() within csum_partial_copy_from_user()
s390/uaccess: remove copy_from_user_real()
s390/sclp_early: Return correct HSA block count also for zero
s390: add some drivers/subsystems to the MAINTAINERS file
s390: improve debug feature usage
s390/airq: add support for irq ranges
s390/mm: enable split page table lock for PMD level
...
The architectures that override cputime_t (s390, ppc) don't provide
any version of nsecs_to_cputime(). Indeed this cputime_t implementation
by backend only happens when CONFIG_VIRT_CPU_ACCOUNTING_NATIVE=y under
which the core code doesn't make any use of nsecs_to_cputime().
At least for now.
We are going to make a broader use of it so lets provide a default
version with a per usecs granularity. It should be good enough for most
usecases.
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Merge the request/release callbacks which are in a separate branch for
consumption by the gpio folks.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
commit: 8f945a33 (genirq: Move kstat_incr_irqs_this_cpu() to core)
unearthed the following:
arch/s390/kernel/irq.c: In function 'init_IRQ':
>> arch/s390/kernel/irq.c:93:2: error: implicit declaration of function 'irq_reserve_irqs'
[-Werror=implicit-function-declaration]
....
cc1: some warnings being treated as errors
--
drivers/s390/cio/cio.c: In function 'init_cio_interrupts':
>> drivers/s390/cio/cio.c:594:2: error: implicit declaration of function
'irq_set_chip_and_handler' [-Werror=implicit-function-declaration]
[-Werror=implicit-function-declaration]
....
cc1: some warnings being treated as errors
The reason is that those files require linux/irq.h and magically
pulled that in via linux/kernel_stat.h
The commit above got rid of the pointless include of linux/irq.h in
linux/kernel_stat.h and therefor broke the build.
Include linux/irq.h
Reported-by: fengguang.wu@intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: s390 <linux-s390@vger.kernel.org>
Let the core do the irq_desc resolution.
No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: s390 <linux-s390@vger.kernel.org>
Link: http://lkml.kernel.org/r/20140223212737.983433636@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The maximum usable buffer size of the s390 debug feature (when using
the sprintf_view) is 11 * sizeof(long) (1 pointer for the format
string + 10 arguments). When a larger buffer size is specified the
additional memory is unused and wasted per debug entry. So reducing
the buffer size to its maximum (or to the actual buffer size used)
will make more precious debug feature space usable.
For pci_msg, chsc_msg, and cio_crw we use the additional usable dbf
space to reduce the number of allocated pages.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Add airq_iv_alloc and airq_iv_free to allocate and free consecutive
ranges of irqs from the interrupt vector.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Interrupts which happen on ccw consoles prior to their registration
with the driver core are not accounted to the respective device
driver. Fix this by setting the proper interrupt class during
initialization of ccw consoles.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Drivers for ccw consoles use ccw_device_probe_console to receive
an initialized ccw device which is already enabled for interrupts.
After that the device driver does the initialization of its private
data. This can race with unsolicited interrupts which can happen
once the device is enabled for interrupts.
Split ccw_device_probe_console into ccw_device_create_console and
ccw_device_enable_console and reorder the initialization of the ccw
console drivers.
While at it mark these functions as __init.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
ccw consoles are in use before they can be properly registered with
the driver core. For devices which are in use by a device driver we
rely on the ccw_device's pointer to the driver callbacks to be valid.
For ccw consoles this pointer is NULL until they are registered later
during boot and we dereferenced this pointer. This worked by
chance on 64 bit builds (cdev->drv was NULL but the optional callback
cdev->drv->path_event was also NULL by coincidence) and was unnoticed
until we received reports about boot failures on 31 bit systems.
Fix it by initializing the driver pointer for ccw consoles.
Cc: <stable@vger.kernel.org> # 3.10+
Reported-by: Mike Frysinger <vapier@gentoo.org>
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Performing a Channel-Path configure on operation on a Channel-Path ID
(CHPID) does not trigger a scan for subchannels that might have become
available through that CHPID. As a result, some subchannels and
associated I/O devices might be missing. Fix this by adding the missing
scan.
This problem was introduced by commit c820de39, "[S390] cio: Rework
css driver.", but wasn't noticed earlier because the machine usually
also generates a Channel-Report-Word when the first CHPID of a
subchannel is configured on, resulting in a separate scan for that
subchannel. The problem only becomes apparent when this first CHPID is
not working properly and additional working CHPIDs are subsequently
configured on without any effect on the availability of the affected
subchannel.
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
driver-core now supports synchrnous self-deletion of attributes and
the asynchrnous removal mechanism is scheduled for removal. Use it
instead of device_schedule_callback().
* Conversions in arch/s390/pci/pci_sysfs.c and
drivers/s390/block/dcssblk.c are straightforward.
* drivers/s390/cio/ccwgroup.c is a bit more tricky because
ccwgroup_notifier() was (ab)using device_schedule_callback() to
purely obtain a process context to kick off ungroup operation which
may block from a notifier callback.
Rename ccwgroup_ungroup_callback() to ccwgroup_ungroup() and make it
take ccwgroup_device * instead. The new function is now called
directly from ccwgroup_ungroup_store().
ccwgroup_notifier() chain is updated to explicitly bounce through
ccwgroup_device->ungroup_work. This also removes possible failure
from memory pressure.
Only compile-tested.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux390@de.ibm.com
Cc: linux-s390@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The modify subchannel wrapper cio_commit_config can fail when
(unexpected) status is pending on the subchannel.
Callers of cio_commit_config (that operated on enabled subchannels)
needed to do error handling for that case (clear the unexpected
status with test subchannel and retry). This error handling is
missing in some code paths and caused online setting of devices to
fail.
Fix this for all callers by moving the error handling inside
cio_commit_config.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Get rid of compile warning in qdio_int_handler_pci() when checking
for program-controlled interruption on outbound queues.
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
I observed that there are for_each macros that do an extra memory access
beyond the defined area.
Normally this does not cause problems.
But, this can cause exceptions. For example: if the area is allocated at
the end of a page and the next page is not accessible.
For correctness, I suggest changing the arguments of the 'for loop' like
others 'for_each' do in the kernel.
Signed-off-by: Jose Alonso <joalonsof@gmail.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Introduce function for the "Perform network-subchannel operation"
CHSC command with operation code "bridgeport information",
and bit definitions for "characteristics" pertaning to this command.
Signed-off-by: Eugene Crosser <eugene.crosser@ru.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
When calling set_{on,off}line of a ccwgroup device driver we hold
the module reference of the owner. This is pretty useless - we don't
want to prevent module unloading but driver unbinding. Use the
driver core's device_lock instead.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
When calling set_{on,off}line of a ccw device driver we hold the
module reference of the owner. This is pretty useless - we don't
want to prevent module unloading but driver unbinding. Use the
driver core's device_lock instead.
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Make sure that access to the online member of a ccw device is
guarded by the ccwlock.
Reported-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Move scheduling of a subchannel scan to those instances where new
devices may actually have become available. This reduces unnecessary
scan work in case devices were added to the blacklist.
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Subchannel looping function for_each_subchannel_staged() allocates a
subchannel-ID-bitmap to efficiently iterate over the list of known
and unknown subchannels. Since this function is also used to iterate
over known-subchannels only, optimize that case by not requiring the
ID-bitmap allocation and falling back to simple bus_for_each_dev()
looping.
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The CIO layer scans for newly available I/O devices by performing a scan
of available subchannels using the Store Subchannel (STSCH) instruction.
Performing too many STSCH instructions in a tight loop can cause high
Hypervisor overhead which can negatively impact the performance of the
virtual machine as a whole.
A subchannel scan is triggered for example during a hardware event that
indicates that a channel path has become available. It is also triggered
by the DASD device driver for each device that is set online.
This patch reduces the number of STSCH instructions being performed by
delaying the start of the actual subchannel scan by 1 second. Multiple
scan requests that are scheduled during this time will be merged into a
single scan loop.
The trade-off consists of a short delay that is introduced between
the time that the event is processed and a newly available device
becoming usable. This delay should be acceptable since it only
affects devices that have not been in use before.
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The CIO layer scans for newly available I/O devices by performing a scan
of available subchannels using the Store Subchannel (STSCH) instruction.
This processing can take a significant amount of time during which no
other task can run on the same CPU (unless CONFIG_PREEMPT has been
enabled). As a result, scheduling latencies for other tasks are
increased noticeably, especially on a single-CPU system.
Fix this problem by explicitly allowing other tasks to be scheduled
each time a subchannel has been processed.
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The CIO layer processes hardware events that indicate that a channel
path has become available by performing a scan of available subchannels
using the Store Subchannel (STSCH) instruction. Performing too many
STSCH instructions in a tight loop can cause high Hypervisor overhead
which can negatively impact the performance of the virtual machine as
a whole.
This patch reduces the number of STSCH instructions performed while
processing a resource accessibility event and while varying a CHPID
online.
In both cases, Linux first performs a STSCH instruction on each unused
subchannel to see if the subchannel has become available. If the STSCH
instruction indicates that the subchannel is available, a full
evaluation of this subchannel is scheduled. Since the full evaluation
includes performing a STSCH instruction, the initial STSCH is
unnecessary and can be removed.
Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Stop hiding scm_block's dependency to the eadm subchannel driver
(by using functions provided by the eadm subchannel instead of
wrappers provided by the scm bus).
This will help userspace recognizing module dependencies (e.g. for
building a ramdisk). As a side effect we can get rid of some code
reimplementing refcounting between those modules.
Reported-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Reviewed-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
When quiescing an eadm subchannel make sure that outstanding IO is
cleared and potential timeout handlers are canceled.
Reviewed-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
find_first_bit_left() and friends have nothing to do with the normal
LSB0 bit numbering for big endian machines used in Linux (least
significant bit has bit number 0).
Instead they use MSB0 bit numbering, where the most signficant bit has
bit number 0. So rename find_first_bit_left() and friends to
find_first_bit_inv(), to avoid any confusion.
Also provide inv versions of set_bit, clear_bit and test_bit.
This also removes the confusing use of e.g. set_bit() in airq.c which
uses a "be_to_le" bit number conversion, which could imply that instead
set_bit_le() could be used. But that is entirely wrong since the _le
bitops variant uses yet another bit numbering scheme.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Refactor direct debug level comparisons with the (internal) s390db->level
member. Use the debug_level_enabled() function instead.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
get_inbound_buffer_frontier() makes use of the return value of atomic_sub()
which shouldn't work, since atomic_sub() is supposed to return void.
This only works on s390 because atomic_sub() gets mapped to atomic_sub_return()
with a define without changing it's return value to void.
So use atomic_sub_return() instead of atomic_sub() in qeth code before fixing
atomic ops.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The result of the store-clock-fast (STCKF) instruction is a bit fuzzy.
It can happen that the value stored on one CPU is smaller than the value
stored on another CPU, although the order of the stores is the other
way around. This can cause deltas of get_tod_clock() values to become
negative when they should not be.
We need to be more careful with store-clock-fast, this patch partially
reverts git commit e4b7b4238e666682555461fa52eecd74652f36bb "time:
always use stckf instead of stck if available". The get_tod_clock()
function now uses the store-clock-extended (STCKE) instruction.
get_tod_clock_fast() can be used if the fuzziness of store-clock-fast
is acceptable e.g. for wait loops local to a CPU.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Access to the slow_subchannel_set has to be secured via the
slow_subchannel_lock.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Some functions that do arch specific resume actions are called
directly from swsusp_asm64.S . Before we add another function call
provide a generic s390_early_resume function which can be used
for this purpose.
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
