With "mm: vmscan: scale number of pages reclaimed by reclaim/compaction
based on failures" reverted, Zdenek Kabelac reported the following
Hmm, so it's just took longer to hit the problem and observe
kswapd0 spinning on my CPU again - it's not as endless like before -
but still it easily eats minutes - it helps to turn off Firefox
or TB (memory hungry apps) so kswapd0 stops soon - and restart
those apps again. (And I still have like >1GB of cached memory)
kswapd0 R running task 0 30 2 0x00000000
Call Trace:
preempt_schedule+0x42/0x60
_raw_spin_unlock+0x55/0x60
put_super+0x31/0x40
drop_super+0x22/0x30
prune_super+0x149/0x1b0
shrink_slab+0xba/0x510
The sysrq+m indicates the system has no swap so it'll never reclaim
anonymous pages as part of reclaim/compaction. That is one part of the
problem but not the root cause as file-backed pages could also be
reclaimed.
The likely underlying problem is that kswapd is woken up or kept awake
for each THP allocation request in the page allocator slow path.
If compaction fails for the requesting process then compaction will be
deferred for a time and direct reclaim is avoided. However, if there
are a storm of THP requests that are simply rejected, it will still be
the the case that kswapd is awake for a prolonged period of time as
pgdat->kswapd_max_order is updated each time. This is noticed by the
main kswapd() loop and it will not call kswapd_try_to_sleep(). Instead
it will loopp, shrinking a small number of pages and calling
shrink_slab() on each iteration.
The temptation is to supply a patch that checks if kswapd was woken for
THP and if so ignore pgdat->kswapd_max_order but it'll be a hack and not
backed up by proper testing. As 3.7 is very close to release and this
is not a bug we should release with, a safer path is to revert "mm:
remove __GFP_NO_KSWAPD" for now and revisit it with the view to ironing
out the balance_pgdat() logic in general.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Zdenek Kabelac <zkabelac@redhat.com>
Cc: Seth Jennings <sjenning@linux.vnet.ibm.com>
Cc: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Cc: Jiri Slaby <jirislaby@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch fixes errors seen in identifying old Samsung SLC, due to the
following commits:
commit e2d3a35ee4
mtd: nand: detect Samsung K9GBG08U0A, K9GAG08U0F ID
commit e3b88bd604
mtd: nand: add generic READ ID length calculation functions
Some Samsung NAND with "5-byte" ID really appear to have 6-byte IDs, with
wraparound like:
Samsung K9K8G08U0D
ec d3 51 95 58 ec ec d3
Samsung K9F1G08U0C
ec f1 00 95 40 ec ec f1
Samsung K9F2G08U0B
ec da 10 95 44 00 ec da
This bad wraparound makes it hard to reliably detect the difference
between Samsung SLC with 5-byte ID and Samsung SLC with 6-byte ID.
The fix is to, for now, only use the new Samsung table for MLC. We
cannot support the new SLC (K9FAG08U0M) until Samsung gives better ID
decode information.
Note that this applies in addition to the previous regression fix:
commit bc86cf7af2
mtd: nand: fix Samsung SLC NAND identification regression
Together, these patches completely restore the previous detection
behavior so that we cannot see any more regressions in Samsung SLC NAND
(finger crossed). With luck, I can get a hold of a Samsung
representative and stop having to cross my fingers eventually.
Reported-by: Sylwester Nawrocki <sylvester.nawrocki@gmail.com>
Tested-by: Sylwester Nawrocki <sylvester.nawrocki@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
A combination of the following two commits caused a regression in 3.7-rc1
when identifying some Samsung NAND, so that some previously working NAND
were no longer detected properly:
commit e3b88bd604
mtd: nand: add generic READ ID length calculation functions
commit e2d3a35ee4
mtd: nand: detect Samsung K9GBG08U0A, K9GAG08U0F ID
Particularly, a regression was seen on Samsung K9F2G08U0B, with the
following full 8-byte READ ID string:
ec da 10 95 44 00 ec da
The basic problem is that Samsung manufactures both SLC and MLC NAND
that use a non-standard decoding table for deriving information from
their IDs. I have heuristically determined that all the chips that use
the new table have ID strings which wrap around after the 6th byte.
Unfortunately, I overlooked the fact that some older Samsung SLC (which
use a different decoding table) have "5 byte ID strings" which also wrap
around after the 6th byte.
This patch re-introduces a distinction between these old and new Samsung
NAND by checking that the 6th byte is non-zero, allowing both old and
new Samsung NAND to be detected properly.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Tested-by: Brian Norris <computersforpeace@gmail.com>
Reported-by: Marek Vasut <marex@denx.de>
Tested-by: Marek Vasut <marex@denx.de>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Fixes the following sparse warning:
drivers/mtd/onenand/onenand_base.c:3697:5: warning:
symbol 'flexonenand_set_boundary' was not declared. Should it be static?
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
The pointer returned by kzalloc should be tested for NULL
to avoid potential NULL pointer dereference later. Incorrect
pointer was being tested for NULL. Bug introduced by commit fbcf62a3
(mtd: physmap_of: move parse_obsolete_partitions to become separate
parser).
This patch fixes this bug.
Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Cc: Artem Bityutskiy <artem.bityutskiy@intel.com>
Cc: stable@kernel.org
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
A combination of the following two commits caused a regression in 3.7-rc1
when identifying some Samsung NAND, so that some previously working NAND
were no longer detected properly:
commit e3b88bd604
mtd: nand: add generic READ ID length calculation functions
commit e2d3a35ee4
mtd: nand: detect Samsung K9GBG08U0A, K9GAG08U0F ID
Particularly, a regression was seen on Samsung K9F2G08U0B, with the
following full 8-byte READ ID string:
ec da 10 95 44 00 ec da
The basic problem is that Samsung manufactures both SLC and MLC NAND
that use a non-standard decoding table for deriving information from
their IDs. I have heuristically determined that all the chips that use
the new table have ID strings which wrap around after the 6th byte.
Unfortunately, I overlooked the fact that some older Samsung SLC (which
use a different decoding table) have "5 byte ID strings" which also wrap
around after the 6th byte.
This patch re-introduces a distinction between these old and new Samsung
NAND by checking that the 6th byte is non-zero, allowing both old and
new Samsung NAND to be detected properly.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Tested-by: Brian Norris <computersforpeace@gmail.com>
Reported-by: Marek Vasut <marex@denx.de>
Tested-by: Marek Vasut <marex@denx.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This code was broken because it assumed that all MTD devices were map-based.
Disable it for now, until it can be fixed properly for the next merge window.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
A long time ago, in v2.4, VM_RESERVED kept swapout process off VMA,
currently it lost original meaning but still has some effects:
| effect | alternative flags
-+------------------------+---------------------------------------------
1| account as reserved_vm | VM_IO
2| skip in core dump | VM_IO, VM_DONTDUMP
3| do not merge or expand | VM_IO, VM_DONTEXPAND, VM_HUGETLB, VM_PFNMAP
4| do not mlock | VM_IO, VM_DONTEXPAND, VM_HUGETLB, VM_PFNMAP
This patch removes reserved_vm counter from mm_struct. Seems like nobody
cares about it, it does not exported into userspace directly, it only
reduces total_vm showed in proc.
Thus VM_RESERVED can be replaced with VM_IO or pair VM_DONTEXPAND | VM_DONTDUMP.
remap_pfn_range() and io_remap_pfn_range() set VM_IO|VM_DONTEXPAND|VM_DONTDUMP.
remap_vmalloc_range() set VM_DONTEXPAND | VM_DONTDUMP.
[akpm@linux-foundation.org: drivers/vfio/pci/vfio_pci.c fixup]
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Carsten Otte <cotte@de.ibm.com>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Eric Paris <eparis@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Morris <james.l.morris@oracle.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Kentaro Takeda <takedakn@nttdata.co.jp>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Venkatesh Pallipadi <venki@google.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When transparent huge pages were introduced, memory compaction and swap
storms were an issue, and the kernel had to be careful to not make THP
allocations cause pageout or compaction.
Now that we have working compaction deferral, kswapd is smart enough to
invoke compaction and the quadratic behaviour around isolate_free_pages
has been fixed, it should be safe to remove __GFP_NO_KSWAPD.
[minchan@kernel.org: Comment fix]
[mgorman@suse.de: Avoid direct reclaim for deferred compaction]
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Weinberger from Linutronix. Fastmap is designed to address UBI's slow scanning
issues. Namely, it introduces a new on-flash data-structure called "fastmap",
which stores the information about logical<->physical eraseblocks mappings.
So now to get this information just read the fastmap, instead of doing full
scan. More information here can be found in Richard's announcement in LKML
(Subject: UBI: Fastmap request for inclusion (v19)):
http://thread.gmane.org/gmane.linux.kernel/1364922/focus=1369109
One thing I want to explicitly say is that fastmap did not have large
enough linux-next exposure. It is partially my fault - I did not respond
quickly enough. I _really_ apologize for this. But it had good testing and
disabled by default, so I do not expect that we'll break anything.
Fastmap is declared as experimental so far, and it is off by default. We
did declare that the on-flash format may be changed. The reason for this is
that no one used it in real production so far, so there is a high risk that
something is missing. Besides, we do not have user-space tools supporting
fastmap so far.
Nevertheless, I suggest we merge this feature. Many people want UBI's scanning
bottleneck to be fixed and merging fastmap now should accelerate its production
use. The plan is to make it bullet-prove, somewhat clean-up, and make it the
default for UBI. I do not know how many kernel releases will it take.
Basically, I what I want to do for fastmap is something like Linus did for
btrfs few years ago.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)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=/Sqc
-----END PGP SIGNATURE-----
Merge tag 'upstream-3.7-rc1-fastmap' of git://git.infradead.org/linux-ubi
Pull UBI fastmap changes from Artem Bityutskiy:
"This pull request contains the UBI fastmap support implemented by
Richard Weinberger from Linutronix. Fastmap is designed to address
UBI's slow scanning issues. Namely, it introduces a new on-flash
data-structure called "fastmap", which stores the information about
logical<->physical eraseblocks mappings. So now to get this
information just read the fastmap, instead of doing full scan. More
information here can be found in Richard's announcement in LKML
(Subject: UBI: Fastmap request for inclusion (v19)):
http://thread.gmane.org/gmane.linux.kernel/1364922/focus=1369109
One thing I want to explicitly say is that fastmap did not have large
enough linux-next exposure. It is partially my fault - I did not
respond quickly enough. I _really_ apologize for this. But it had
good testing and disabled by default, so I do not expect that we'll
break anything.
Fastmap is declared as experimental so far, and it is off by default.
We did declare that the on-flash format may be changed. The reason
for this is that no one used it in real production so far, so there is
a high risk that something is missing. Besides, we do not have
user-space tools supporting fastmap so far.
Nevertheless, I suggest we merge this feature. Many people want UBI's
scanning bottleneck to be fixed and merging fastmap now should
accelerate its production use. The plan is to make it bullet-prove,
somewhat clean-up, and make it the default for UBI. I do not know how
many kernel releases will it take.
Basically, I what I want to do for fastmap is something like Linus did
for btrfs few years ago."
* tag 'upstream-3.7-rc1-fastmap' of git://git.infradead.org/linux-ubi:
UBI: Wire-up fastmap
UBI: Add fastmap core
UBI: Add fastmap support to the WL sub-system
UBI: Add fastmap stuff to attach.c
UBI: Wire-up ->fm_sem
UBI: Add fastmap bits to build.c
UBI: Add self_check_eba()
UBI: Export next_sqnum()
UBI: Add fastmap stuff to ubi.h
UBI: Add fastmap on-flash data structures
Pull powerpc updates from Benjamin Herrenschmidt:
"Some highlights in addition to the usual batch of fixes:
- 64TB address space support for 64-bit processes by Aneesh Kumar
- Gavin Shan did a major cleanup & re-organization of our EEH support
code (IBM fancy PCI error handling & recovery infrastructure) which
paves the way for supporting different platform backends, along
with some rework of the PCIe code for the PowerNV platform in order
to remove home made resource allocations and instead use the
generic code (which is possible after some small improvements to it
done by Gavin).
- Uprobes support by Ananth N Mavinakayanahalli
- A pile of embedded updates from Freescale folks, including new SoC
and board supports, more KVM stuff including preparing for 64-bit
BookE KVM support, ePAPR 1.1 updates, etc..."
Fixup trivial conflicts in drivers/scsi/ipr.c
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (146 commits)
powerpc/iommu: Fix multiple issues with IOMMU pools code
powerpc: Fix VMX fix for memcpy case
driver/mtd:IFC NAND:Initialise internal SRAM before any write
powerpc/fsl-pci: use 'Header Type' to identify PCIE mode
powerpc/eeh: Don't release eeh_mutex in eeh_phb_pe_get
powerpc: Remove tlb batching hack for nighthawk
powerpc: Set paca->data_offset = 0 for boot cpu
powerpc/perf: Sample only if SIAR-Valid bit is set in P7+
powerpc/fsl-pci: fix warning when CONFIG_SWIOTLB is disabled
powerpc/mpc85xx: Update interrupt handling for IFC controller
powerpc/85xx: Enable USB support in p1023rds_defconfig
powerpc/smp: Do not disable IPI interrupts during suspend
powerpc/eeh: Fix crash on converting OF node to edev
powerpc/eeh: Lock module while handling EEH event
powerpc/kprobe: Don't emulate store when kprobe stwu r1
powerpc/kprobe: Complete kprobe and migrate exception frame
powerpc/kprobe: Introduce a new thread flag
powerpc: Remove unused __get_user64() and __put_user64()
powerpc/eeh: Global mutex to protect PE tree
powerpc/eeh: Remove EEH PE for normal PCI hotplug
...
Make fastmap known to Kconfig, UBI Makefile and MAINTAINERS.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
To make fastmap possible the WL sub-system needs some
changes.
Mostly to support fastmaps pools.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Fastmap uses ->fm_sem to stop EBA changes while writing
a new fastmap.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
self_check_eba() compares two ubi_attach_info objects.
Fastmap uses this function for self checks.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Fastmap needs next_sqnum(), rename it to ubi_next_sqnum()
and make it non-static.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
This patch adds fastmap specific data structures to ubi.h.
It moves also struct ubi_work to ubi.h as it is now needed
for more than one c file.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Add the on-flash data structures neeed by fastmap
to ubi-media.h
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
handling. We used to reserve 2% of the partition, but now we are
more aggressive and we reserve 2% of the entire chip, which is
what actually manufacturers specify in data sheets. We introduced
an option to users to override the default, though.
There are a couple of fixes as well, and a number of cleanups.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)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=95SE
-----END PGP SIGNATURE-----
Merge tag 'upstream-3.7-rc1' of git://git.infradead.org/linux-ubi
Pull UBI changes from Artem Bityutskiy:
"The main change is the way we reserve eraseblocks for bad blocks
handling. We used to reserve 2% of the partition, but now we are more
aggressive and we reserve 2% of the entire chip, which is what
actually manufacturers specify in data sheets. We introduced an
option to users to override the default, though.
There are a couple of fixes as well, and a number of cleanups."
* tag 'upstream-3.7-rc1' of git://git.infradead.org/linux-ubi: (24 commits)
UBI: fix trivial typo 'it' => 'is'
UBI: load after mtd device drivers
UBI: print less
UBI: use pr_ helper instead of printk
UBI: comply with coding style
UBI: erase free PEB with bitflip in EC header
UBI: fix autoresize handling in R/O mode
UBI: add max_beb_per1024 to attach ioctl
UBI: allow specifying bad PEBs limit using module parameter
UBI: check max_beb_per1024 value in ubi_attach_mtd_dev
UBI: prepare for max_beb_per1024 module parameter addition
UBI: introduce MTD_PARAM_MAX_COUNT
UBI: separate bad_peb_limit in a function
arm: sam9_l9260_defconfig: correct CONFIG_MTD_UBI_BEB_LIMIT
UBI: use the whole MTD device size to get bad_peb_limit
mtd: mtdparts: introduce mtd_get_device_size
mtd: mark mtd_is_partition argument as constant
arm: sam9_l9260_defconfig: remove non-existing config option
UBI: kill CONFIG_MTD_UBI_BEB_RESERVE
UBI: limit amount of reserved eraseblocks for bad PEB handling
...
Pull workqueue changes from Tejun Heo:
"This is workqueue updates for v3.7-rc1. A lot of activities this
round including considerable API and behavior cleanups.
* delayed_work combines a timer and a work item. The handling of the
timer part has always been a bit clunky leading to confusing
cancelation API with weird corner-case behaviors. delayed_work is
updated to use new IRQ safe timer and cancelation now works as
expected.
* Another deficiency of delayed_work was lack of the counterpart of
mod_timer() which led to cancel+queue combinations or open-coded
timer+work usages. mod_delayed_work[_on]() are added.
These two delayed_work changes make delayed_work provide interface
and behave like timer which is executed with process context.
* A work item could be executed concurrently on multiple CPUs, which
is rather unintuitive and made flush_work() behavior confusing and
half-broken under certain circumstances. This problem doesn't
exist for non-reentrant workqueues. While non-reentrancy check
isn't free, the overhead is incurred only when a work item bounces
across different CPUs and even in simulated pathological scenario
the overhead isn't too high.
All workqueues are made non-reentrant. This removes the
distinction between flush_[delayed_]work() and
flush_[delayed_]_work_sync(). The former is now as strong as the
latter and the specified work item is guaranteed to have finished
execution of any previous queueing on return.
* In addition to the various bug fixes, Lai redid and simplified CPU
hotplug handling significantly.
* Joonsoo introduced system_highpri_wq and used it during CPU
hotplug.
There are two merge commits - one to pull in IRQ safe timer from
tip/timers/core and the other to pull in CPU hotplug fixes from
wq/for-3.6-fixes as Lai's hotplug restructuring depended on them."
Fixed a number of trivial conflicts, but the more interesting conflicts
were silent ones where the deprecated interfaces had been used by new
code in the merge window, and thus didn't cause any real data conflicts.
Tejun pointed out a few of them, I fixed a couple more.
* 'for-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: (46 commits)
workqueue: remove spurious WARN_ON_ONCE(in_irq()) from try_to_grab_pending()
workqueue: use cwq_set_max_active() helper for workqueue_set_max_active()
workqueue: introduce cwq_set_max_active() helper for thaw_workqueues()
workqueue: remove @delayed from cwq_dec_nr_in_flight()
workqueue: fix possible stall on try_to_grab_pending() of a delayed work item
workqueue: use hotcpu_notifier() for workqueue_cpu_down_callback()
workqueue: use __cpuinit instead of __devinit for cpu callbacks
workqueue: rename manager_mutex to assoc_mutex
workqueue: WORKER_REBIND is no longer necessary for idle rebinding
workqueue: WORKER_REBIND is no longer necessary for busy rebinding
workqueue: reimplement idle worker rebinding
workqueue: deprecate __cancel_delayed_work()
workqueue: reimplement cancel_delayed_work() using try_to_grab_pending()
workqueue: use mod_delayed_work() instead of __cancel + queue
workqueue: use irqsafe timer for delayed_work
workqueue: clean up delayed_work initializers and add missing one
workqueue: make deferrable delayed_work initializer names consistent
workqueue: cosmetic whitespace updates for macro definitions
workqueue: deprecate system_nrt[_freezable]_wq
workqueue: deprecate flush[_delayed]_work_sync()
...
This is a pretty significant branch. It's the introduction of the
first multiplatform support on ARM, and with this (and the later
branch) merged, it is now possible to build one kernel that contains
support for highbank, vexpress, mvebu, socfpga, and picoxcell. More
platforms will be convered over in the next few releases.
Two critical last things had to be done for this to be practical and
possible:
* Today each platform has its own include directory under
mach-<mach>/include/mach/*, and traditionally that is where a lot of
driver/platform shared definitions have gone, such as platform data
structures. They now need to move out to a common location instead,
and this branch moves a large number of those out to
include/linux/platform_data.
* Each platform used to list the device trees to compile for its
boards in mach-<mach>/Makefile.boot.
Both of the above changes will mean that there are some merge
conflicts to come (and some to resolve here). It's a one-time move and
once it settles in, we should be good for quite a while. Sorry for the
overhead.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=Tc6y
-----END PGP SIGNATURE-----
Merge tag 'multiplatform' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM soc multiplatform enablement from Olof Johansson:
"This is a pretty significant branch. It's the introduction of the
first multiplatform support on ARM, and with this (and the later
branch) merged, it is now possible to build one kernel that contains
support for highbank, vexpress, mvebu, socfpga, and picoxcell. More
platforms will be convered over in the next few releases.
Two critical last things had to be done for this to be practical and
possible:
* Today each platform has its own include directory under
mach-<mach>/include/mach/*, and traditionally that is where a lot
of driver/platform shared definitions have gone, such as platform
data structures. They now need to move out to a common location
instead, and this branch moves a large number of those out to
include/linux/platform_data.
* Each platform used to list the device trees to compile for its
boards in mach-<mach>/Makefile.boot.
Both of the above changes will mean that there are some merge
conflicts to come (and some to resolve here). It's a one-time move
and once it settles in, we should be good for quite a while. Sorry
for the overhead."
Fix conflicts as per Olof.
* tag 'multiplatform' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (51 commits)
ARM: add v7 multi-platform defconfig
ARM: msm: Move core.h contents into common.h
ARM: highbank: call highbank_pm_init from .init_machine
ARM: dtb: move all dtb targets to common Makefile
ARM: spear: move platform_data definitions
ARM: samsung: move platform_data definitions
ARM: orion: move platform_data definitions
ARM: vexpress: convert to multi-platform
ARM: initial multiplatform support
ARM: mvebu: move armada-370-xp.h in mach dir
ARM: vexpress: remove dependency on mach/* headers
ARM: picoxcell: remove dependency on mach/* headers
ARM: move all dtb targets out of Makefile.boot
ARM: picoxcell: move debug macros to include/debug
ARM: socfpga: move debug macros to include/debug
ARM: mvebu: move debug macros to include/debug
ARM: vexpress: move debug macros to include/debug
ARM: highbank: move debug macros to include/debug
ARM: move debug macros to common location
ARM: make mach/gpio.h headers optional
...
- A long-coming conversion of various platforms to a common LED
infrastructure
- AT91 is moved over to use the newer MCI driver for MMC
- Pincontrol conversions for samsung platforms
- DT bindings for gscaler on samsung
- i2c driver fixes for tegra, acked by i2c maintainer
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
iQIcBAABAgAGBQJQaO48AAoJEIwa5zzehBx3excP/ieEkRhvfnWxdYST6ekvGIQr
nEyskOh2rVjgYKmSXUJyKSbvG+7bZ8VIxvPvojeAJ/R84pTFMzbR2F0CaPKzAuSW
inDt6c0Bnx1NZlfUTAoXcz7feyq9zHYNs9BCLoPU0bYNchCCqcWSKzqnpXk2ph/P
LFnmNa0j6a4E3QJYAjM2zFvc3Tgk+MWTq1fWwNFvsWTh2WbQtmB/iGnzT5Xs4XQh
u1SSx5tz0lcF5zQRGmJhXgL5+nnIP4sRwRUBAkpe3Gv5cM6WBVEBRDANa5QpbUL2
RXK5YyCTIln2Me4bPk32zEBLjiZ/WXbmiA2uwoqVgy6XToubemDXd0PtKmjj5tZ1
BkTD1DND7BKBEQnJj/GBECEdvx2FbrKfruoPcJHvXPZ7Svn5Dt/MWYJQIkRFkuhL
zlVNoDGWlU8nScGrgmTM56UvWmGWC3UFsWSgdVQNfW9yEva+G1FvRUwUH02Ip5Ad
4r28JFIn6zyjtM99ZHipU6C6Rze2ordC7fl5X5LBLkVOobioblxCAhIhcqkhfKsk
rFriNsdfYs7SrJA7mK7GzvaMEJgp/5o1noJKXI7ZBcLI8yYagzbQbPu/vGi6G6d3
0xC7NaTEJbtoXoDAtmtilLRxmw0YCXgVBBGua0K2YKpcRwnzCHNbV4gsLMnDuOXS
HP4M96LxLHJlLGCxhEme
=ck7M
-----END PGP SIGNATURE-----
Merge tag 'drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM soc driver specific changes from Olof Johansson:
- A long-coming conversion of various platforms to a common LED
infrastructure
- AT91 is moved over to use the newer MCI driver for MMC
- Pincontrol conversions for samsung platforms
- DT bindings for gscaler on samsung
- i2c driver fixes for tegra, acked by i2c maintainer
Fix up conflicts as per Olof.
* tag 'drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (48 commits)
drivers: bus: omap_l3: use resources instead of hardcoded irqs
pinctrl: exynos: Fix wakeup IRQ domain registration check
pinctrl: samsung: Uninline samsung_pinctrl_get_soc_data
pinctrl: exynos: Correct the detection of wakeup-eint node
pinctrl: exynos: Mark exynos_irq_demux_eint as inline
pinctrl: exynos: Handle only unmasked wakeup interrupts
pinctrl: exynos: Fix typos in gpio/wkup _irq_mask
pinctrl: exynos: Set pin function to EINT in irq_set_type of GPIO EINTa
drivers: bus: Move the OMAP interconnect driver to drivers/bus/
i2c: tegra: dynamically control fast clk
i2c: tegra: I2_M_NOSTART functionality not supported in Tegra20
ARM: tegra: clock: remove unused clock entry for i2c
ARM: tegra: clock: add connection name in i2c clock entry
i2c: tegra: pass proper name for getting clock
ARM: tegra: clock: add i2c fast clock entry in clock table
ARM: EXYNOS: Adds G-Scaler device from Device Tree
ARM: EXYNOS: Add clock support for G-Scaler
ARM: EXYNOS: Enable pinctrl driver support for EXYNOS4 device tree enabled platform
ARM: dts: Add pinctrl node entries for SAMSUNG EXYNOS4210 SoC
ARM: EXYNOS: skip wakeup interrupt setup if pinctrl driver is used
...
Device tree conversion and enablement branch. Mostly a bunch of new
bindings and setup for various platforms, but the Via/Winchip VT8500
platform is also converted over from being 100% legacy to now use
device tree for probing. More of that will come for 3.8.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=q7ZY
-----END PGP SIGNATURE-----
Merge tag 'dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM soc device tree updates from Olof Johansson:
"Device tree conversion and enablement branch. Mostly a bunch of new
bindings and setup for various platforms, but the Via/Winchip VT8500
platform is also converted over from being 100% legacy to now use
device tree for probing. More of that will come for 3.8."
Trivial conflicts due to removal of vt8500 files, and one documentation
file that was added with slightly different contents both here and in
the USb tree.
* tag 'dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (212 commits)
arm: vt8500: Fixup for missing gpio.h
ARM: LPC32xx: LED fix in PHY3250 DTS file
ARM: dt: mmp-dma: add binding file
arm: vt8500: Update arch-vt8500 to devicetree support.
arm: vt8500: gpio: Devicetree support for arch-vt8500
arm: vt8500: doc: Add device tree bindings for arch-vt8500 devices
arm: vt8500: clk: Add Common Clock Framework support
video: vt8500: Add devicetree support for vt8500-fb and wm8505-fb
serial: vt8500: Add devicetree support for vt8500-serial
rtc: vt8500: Add devicetree support for vt8500-rtc
arm: vt8500: Add device tree files for VIA/Wondermedia SoC's
ARM: tegra: Add Avionic Design Tamonten Evaluation Carrier support
ARM: tegra: Add Avionic Design Medcom-Wide support
ARM: tegra: Add Avionic Design Plutux support
ARM: tegra: Add Avionic Design Tamonten support
ARM: tegra: dts: Add pwm label
ARM: ux500: Fix SSP register address format
ARM: ux500: Apply tc3589x's GPIO/IRQ properties to HREF's DT
ARM: ux500: Remove redundant #gpio-cell properties from Snowball DT
ARM: ux500: Add all encompassing sound node to the HREF Device Tree
...
This is a large branch that contains a handful of different cleanups:
- Fixing up the I/O space remapping on PCI on ARM. This is a series
from Rob Herring that restructures how all pci devices allocate I/O
space, and it's part of the work to allow multiplatform kernels.
- A number of cleanup series for OMAP, moving and removing some
headers, sparse irq rework and in general preparation for
multiplatform.
- Final removal of all non-DT boards for Tegra, it is now
device-tree-only!
- Removal of a stale platform, nxp4008. It's an old mobile chipset
that is no longer in use, and was very likely never really used with
a mainline kernel. We have not been able to find anyone interested
in keeping it around in the kernel.
- Removal of the legacy dmaengine driver on tegra
+ A handful of other things that I haven't described above.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=ck1u
-----END PGP SIGNATURE-----
Merge tag 'cleanup' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM soc general cleanups from Olof Johansson:
"This is a large branch that contains a handful of different cleanups:
- Fixing up the I/O space remapping on PCI on ARM. This is a series
from Rob Herring that restructures how all pci devices allocate I/O
space, and it's part of the work to allow multiplatform kernels.
- A number of cleanup series for OMAP, moving and removing some
headers, sparse irq rework and in general preparation for
multiplatform.
- Final removal of all non-DT boards for Tegra, it is now
device-tree-only!
- Removal of a stale platform, nxp4008. It's an old mobile chipset
that is no longer in use, and was very likely never really used
with a mainline kernel. We have not been able to find anyone
interested in keeping it around in the kernel.
- Removal of the legacy dmaengine driver on tegra
+ A handful of other things that I haven't described above."
Fix up some conflicts with the staging tree (and because nxp4008 was
removed)
* tag 'cleanup' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (184 commits)
ARM: OMAP2+: serial: Change MAX_HSUART_PORTS to 6
ARM: OMAP4: twl-common: Support for additional devices on i2c1 bus
ARM: mmp: using for_each_set_bit to simplify the code
ARM: tegra: harmony: fix ldo7 regulator-name
ARM: OMAP2+: Make omap4-keypad.h local
ARM: OMAP2+: Make l4_3xxx.h local
ARM: OMAP2+: Make l4_2xxx.h local
ARM: OMAP2+: Make l3_3xxx.h local
ARM: OMAP2+: Make l3_2xxx.h local
ARM: OMAP1: Move irda.h from plat to mach
ARM: OMAP2+: Make hdq1w.h local
ARM: OMAP2+: Make gpmc-smsc911x.h local
ARM: OMAP2+: Make gpmc-smc91x.h local
ARM: OMAP1: Move flash.h from plat to mach
ARM: OMAP2+: Make debug-devices.h local
ARM: OMAP1: Move board-voiceblue.h from plat to mach
ARM: OMAP1: Move board-sx1.h from plat to mach
ARM: OMAP2+: Make omap-wakeupgen.h local
ARM: OMAP2+: Make omap-secure.h local
ARM: OMAP2+: Make ctrl_module_wkup_44xx.h local
...
Datasheets for the following Samsung NAND parts (both MLC and SLC) describe
extensions to the Samsung 6-byte extended ID decoding table:
K9GBG08U0A (MLC, 6-byte ID)
K9GAG08U0F (MLC, 6-byte ID)
K9FAG08U0M (SLC, 6-byte ID)
The table found in K9GAG08U0F, p.44, contains a superset of the information
found in other previous datasheets.
This patch adds support for all of these chips, with 512B and 640B OOB sizes.
It also changes the detection pattern such that this table applies to all
Samsung 6-byte ID NAND, not just MLC. This is safe, according to the NAND
parameter data I have collected:
Note that nand_base.c does not yet support the bad block marker scheme defined
for these chips (i.e., scan 1st and last page for BB markers).
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Hynix has introduced a new ID decoding scheme for their newer MLC, some of
which don't support ONFI. The following devices all follow the pattern given in
the datasheet for Hynix H27UBG8T2B, p.22:
Hynix H27UAG8T2A
Hynix H27UBG8T2A
Hynix H27UBG8T2B
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When decoding the extended ID bytes of a NAND chip, we have to calculate the ID
length according to some heuristic patterns (e.g., Does the ID wrap around?
Does it end in trailing zeros?). Currently, these heuristics are built into
complicated if/else blocks that can be hard to understand.
Now, these checks can be done generically in a function, making them more
robust and reusable. In fact, this sort of calculation is needed in future
additions to nand_base.c. And with this advancement, we get the added benefit
of a more readable "extended ID decode".
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When detecting NAND parameters, the code gets a little ugly so that the
logic is obscured. Try to remedy that by moving code to separate functions
that have well-defined purposes.
This patch splits out the simple ID decode functionality, where all the
information regarding NAND size/blocksize/pagesize/oobsize/busw is encoded in
the first two bytes of the ID string.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When detecting NAND parameters, the code gets a little ugly so that the
logic is obscured. Try to remedy that by moving code to separate functions
that have well-defined purposes.
This patch splits out the extended ID decode functionality, which handles
decoding the 3rd-8th ID bytes to determine NAND device parameters.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When detecting NAND parameters, the code gets a little ugly so that the
logic is obscured. Try to remedy that by moving code to separate functions
that have well-defined purposes.
This patch splits the bad block marker options detection into its own function,
away from the other parameters (e.g., chip size, page size, etc.).
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Instead of reading 2 bytes then later 8 bytes, we can simply read all 8
bytes from the start.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
We don't actually use the 'ret' variable; we set it, test it, and then it dies.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
While building an allyesconfig for UML I received this error message(s):
drivers/mtd/nand/docg4.c: In function 'probe_docg4':
drivers/mtd/nand/docg4.c:1272:2: error: implicit declaration of function
'ioremap' [-Werror=implicit-function-declaration]
drivers/mtd/nand/docg4.c:1272:10: warning: assignment makes pointer from
integer without a cast [enabled by default]
drivers/mtd/nand/docg4.c:1327:2: error: implicit declaration of function
'iounmap' [-Werror=implicit-function-declaration]
which is caused by the missing implementations on UML.
This patch adds this missing HAS_IOMEM dependency and prevents the driver from
being build on platforms with no HAS_IOMEM
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Acked-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The current code initializes the timing registers at very time
we call the gpmi_begin(). This really wastes the cpu cycles.
Add a new flag to let the gpmi driver initializes the timing registers
only one time.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When the frequency on the nand chip pins is above 33MHz,
the nand EDO(extended Data Out) timing could be applied.
The GPMI implements a Feedback read strobe to sample the read data in
the EDO timing mode.
This patch adds the EDO feature for the gpmi-nand driver.
For some onfi nand chips, the mode 4 is the fastest;
while for other onfi nand chips, the mode 5 is the fastest.
This patch only adds the support for the fastest asynchronous timing mode.
So this patch only supports the mode 4 and mode 5.
I tested several Micron's ONFI nand chips with EDO enabled,
take Micron MT29F32G08MAA for example (in mode 5, 100MHz):
1) The test result BEFORE we add the EDO feature:
=================================================
mtd_speedtest: MTD device: 2
mtd_speedtest: MTD device size 209715200, eraseblock size 524288,
page size 4096, count of eraseblocks 400,
pages per eraseblock 128, OOB size 218
.......................................
mtd_speedtest: testing eraseblock read speed
mtd_speedtest: eraseblock read speed is 3632 KiB/s
.......................................
mtd_speedtest: testing page read speed
mtd_speedtest: page read speed is 3554 KiB/s
.......................................
mtd_speedtest: testing 2 page read speed
mtd_speedtest: 2 page read speed is 3592 KiB/s
.......................................
=================================================
2) The test result AFTER we add the EDO feature:
=================================================
mtd_speedtest: MTD device: 2
mtd_speedtest: MTD device size 209715200, eraseblock size 524288,
page size 4096, count of eraseblocks 400,
pages per eraseblock 128, OOB size 218
.......................................
mtd_speedtest: testing eraseblock read speed
mtd_speedtest: eraseblock read speed is 19555 KiB/s
.......................................
mtd_speedtest: testing page read speed
mtd_speedtest: page read speed is 17319 KiB/s
.......................................
mtd_speedtest: testing 2 page read speed
mtd_speedtest: 2 page read speed is 18339 KiB/s
.......................................
=================================================
3) The read data performance is much improved by more then 5 times.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The default frequencies of the extra clocks are 200MHz.
The current code sets the extra clocks to 44.5MHz.
When i add the EDO feature to gpmi, i have to revert the extra clocks
to 200MHz.
So it is better that we do not set the default values for the extra
clocks. The driver runs well even when we do not set the default values for
extra clocks.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The setting DLL code is a little mess.
Just simplify the code and the comments.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
add the WRN_DLY_SEL field for HW_GPMI_CTRL1.
This field is used as delay for gpmi write strobe.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The current code will gets the clock frequency which is used by
gpmi_nfc_compute_hardware_timing(). It makes the code a little mess.
So move the `get clock frequency` code to the
gpmi_nfc_compute_hardware_timing() itself. This makes the code tidy
and clean.
This patch also uses the macro NSEC_PER_SEC to replace the `1000000000`.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The gpmi_nfc_compute_hardware_timing{} should contains all the
fields setting for gpmi timing registers. It already contains the fields
for HW_GPMI_TIMING0 and HW_GPMI_CTRL1.
So it is better to add a new field setting for HW_GPMI_TIMING1 in
this data structure. This makes the code more clear in logic.
This patch also changes some comments to make the code more readable.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add the set-features(0xef)/get-features(0xee) helpers for ONFI nand.
Also add the necessary macros.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
If override size is too big, the module was actually loaded instead of
failing, because retval was not set.
This lead to memory corruption with the use of the freed structs nandsim
and nand_chip.
Cc: stable@vger.kernel.org
Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Currently the docg4's ecc.read_page() method returns -EBADMSG when
uncorrectable bitflips occur. This is wrong; 0 should be returned in
this case. An error code should only be returned by this method in the
case of a hardware error (probably -EIO).
Signed-off-by: Mike Dunn <mikedunn@newsguy.com>
Acked-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
