Commit FIXME ("ARM: shmobile: Consolidate SCU mapping code") removed the
last user of the static mapping on emev2-based systems. Remove the
mapping and the legacy machine_desc.map_io() callback.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Commit 37201ba5c99d0be8 ("ARM: shmobile: r8a7740: Migrate to generic l2c
OF initialization") removed the last user of the legacy "IOMEM()" macro
on r8a7740-based systems. Hence there's no longer a need to set up a
transparent mapping of system I/O registers. Remove the mapping and the
legacy machine_desc.map_io() callback.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Currently the SCU registers are mapped in SoC-specific code, using
different methods, all involving the static mapping set up from
machine_desc.map_io():
- On emev2, a static (non-identity) mapping is used, with ioremap().
As the static mapping uses the MT_DEVICE type, ioremap() reuses it,
and the returned virtual address is suitable for passing to
shmobile_smp_hook(),
- On sh73a0 and r8a7779, a static identity mapping is used, with the
legacy IOMEM() macro.
As the static mapping uses the MT_DEVICE_NONSHARED type, replacing
IOMEM() by ioremap() would create a new mapping, whose virtual
address cannot be passed to shmobile_smp_hook().
Move the mapping of the SCU registers from SoC-specific code to common
code, always using ioremap(). To work in the absence of a static
mapping, this requires passing the physical SCU base address to
shmobile_smp_hook().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
ARM: shmobile: Move shmobile_scu_base from .text to .data
shmobile_scu_base is being written to, so it doesn't belong in the .text
section. Fix this by moving it from asm .text to C .data, as it's no
longer used from asm code since commit 4f6da36f7edd5790 ("ARM: shmobile:
Remove old SCU boot code").
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
ARM: shmobile: r8a7740: Remove mapping of L2 cache controller registers
Now all r8a7740-based platforms have been migrated to the generic l2c OF
initialization, it's no longer needed to map the L2 cache controller
registers from .map_io().
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Commit 0ca2894b5a900709 ("ARM: shmobile: Use shared SCU SMP boot code on
r8a7779") obsoleted the r8a7779-specific SCU boot code, but forgot to
remove the setup of shmobile_boot_fn and shmobile_boot_arg, which is
overwritten by shmobile_smp_scu_prepare_cpus().
Note that shmobile_scu_base wasn't initialized at that point yet anyway.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
ARM: dts: koelsch: Enable SCIF_CLK frequency and pins
Add and enable the external crystal for the SCIF_CLK and its pinctrl, to
be used by the Baud Rate Generator for External Clock (BRG) on (H)SCIF.
This increases the range and accuracy of supported baud rates:
- SCIF:
- Supports now 50, 75, 110, 1152000, 1500000, 2000000, and 4000000 bps,
- Perfect match for standard 50-460800, and 9216000 bps.
- More accurate 576000 bps.
- HSCIF:
- Supports now 50, 75, 110, 134, 150, and 200 bps,
- Perfect match for standard 50-460800, and 9216000 bps.
- More accurate 576000, 1152000, 3000000, 3500000, and 4000000
bps.
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (ZS_CLK and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF and HSCIF device nodes.
This increases the range and accuracy of supported baud rates on
(H)SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (ZS_CLK and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF and HSCIF device nodes.
This increases the range and accuracy of supported baud rates on
(H)SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (ZS_CLK and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF and HSCIF device nodes.
This increases the range and accuracy of supported baud rates on
(H)SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (ZS_CLK and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF and HSCIF device nodes.
This increases the range and accuracy of supported baud rates on
(H)SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (S1 and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF device nodes.
This increases the range and accuracy of supported baud rates on SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Add the device node for the external SCIF_CLK.
The presence of the SCIF_CLK crystal and its clock frequency depends on
the actual board.
Add the two optional clock sources (S1 and SCIF_CLK for the internal
resp. external clock) for the Baud Rate Generator for External Clock
(BRG) to all SCIF device nodes.
This increases the range and accuracy of supported baud rates on SCIF.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Linus Torvalds [Tue, 9 Feb 2016 18:15:16 +0000 (10:15 -0800)]
Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto fixes from Herbert Xu:
"This fixes the following issues:
API:
- Fix async algif_skcipher, it was broken by recent fixes.
- Fix potential race condition in algif_skcipher with ctx.
- Fix potential memory corruption in algif_skcipher.
- Add missing lock to crypto_user when doing an alg dump.
Drivers:
- marvell/cesa was testing the wrong variable for NULL after
allocation.
- Fix potential double-free in atmel-sha.
- Fix illegal call to sleepin function from atomic context in
atmel-sha"
* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
crypto: marvell/cesa - fix test in mv_cesa_dev_dma_init()
crypto: atmel-sha - remove calls of clk_prepare() from atomic contexts
crypto: atmel-sha - fix atmel_sha_remove()
crypto: algif_skcipher - Do not set MAY_BACKLOG on the async path
crypto: algif_skcipher - Do not dereference ctx without socket lock
crypto: algif_skcipher - Do not assume that req is unchanged
crypto: user - lock crypto_alg_list on alg dump
J. Bruce Fields [Wed, 10 Jul 2013 20:54:34 +0000 (16:54 -0400)]
scripts: add "prune-kernel" script to clean up old kernel images
Long ago, Dave Jones complained about CONFIG_LOCALVERSION_AUTO:
"I don't use the auto config, because I end up filling up /boot unless
I go through and clean them out by hand every time I install a new one
(which I do probably a dozen or so times a day). Is there some easy
way to prune old builds I'm missing?"
To which Bruce replied:
"I run this by hand every now and then. I'm probably doing it all wrong"
And if he is running it wrong, then so am I - because I've been using
this script ever since. It is true that CONFIG_LOCALVERSION_AUTO easily
ends up filling your /boot partition if you don't clean up old versions
regularly, and this script helps make that easier.
Checked with Bruce to see that it's fine to add this to the kernel
scripts. Maybe people will come up with enhancements, but more
importantly, this way I won't misplace this script whenever I install a
new machine and start doing custom kernels for it.
Mark Rutland [Tue, 9 Feb 2016 17:08:26 +0000 (17:08 +0000)]
asm-generic: make __set_fixmap_offset a macro again
Turning __set_fixmap_offset into a static inline breaks the build for
several architectures. Fixing this properly requires updates to a number
of architectures to make them agree on the prototype of __set_fixmap.
For the timebeing, restore __set_fixmap_offset to its prior state as a
macro function, reverting commit ac4c0ac73485867c ("asm-generic: make
__set_fixmap_offset a static inline"). To avoid the original issue with
namespace clashes, 'addr' is prefixed with a liberal sprinking of
underscores.
Yang Shi [Mon, 8 Feb 2016 22:49:24 +0000 (14:49 -0800)]
arm64: replace read_lock to rcu lock in call_step_hook
BUG: sleeping function called from invalid context at kernel/locking/rtmutex.c:917
in_atomic(): 1, irqs_disabled(): 128, pid: 383, name: sh
Preemption disabled at:[<ffff800000124c18>] kgdb_cpu_enter+0x158/0x6b8
Laura Abbott [Sat, 6 Feb 2016 00:24:48 +0000 (16:24 -0800)]
arm64: ptdump: Indicate whether memory should be faulting
With CONFIG_DEBUG_PAGEALLOC, pages do not have the valid bit
set when free in the buddy allocator. Add an indiciation to
the page table dumping code that the valid bit is not set,
'F' for fault, to make this easier to understand.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Tested-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Laura Abbott <labbott@fedoraproject.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Laura Abbott [Sat, 6 Feb 2016 00:24:47 +0000 (16:24 -0800)]
arm64: Add support for ARCH_SUPPORTS_DEBUG_PAGEALLOC
ARCH_SUPPORTS_DEBUG_PAGEALLOC provides a hook to map and unmap
pages for debugging purposes. This requires memory be mapped
with PAGE_SIZE mappings since breaking down larger mappings
at runtime will lead to TLB conflicts. Check if debug_pagealloc
is enabled at runtime and if so, map everyting with PAGE_SIZE
pages. Implement the functions to actually map/unmap the
pages at runtime.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Tested-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Laura Abbott <labbott@fedoraproject.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Alexander Duyck [Tue, 9 Feb 2016 10:49:54 +0000 (02:49 -0800)]
flow_dissector: Fix unaligned access in __skb_flow_dissector when used by eth_get_headlen
This patch fixes an issue with unaligned accesses when using
eth_get_headlen on a page that was DMA aligned instead of being IP aligned.
The fact is when trying to check the length we don't need to be looking at
the flow label so we can reorder the checks to first check if we are
supposed to gather the flow label and then make the call to actually get
it.
v2: Updated path so that either STOP_AT_FLOW_LABEL or KEY_FLOW_LABEL can
cause us to check for the flow label.
Reported-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by: Alexander Duyck <aduyck@mirantis.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Laura Abbott [Sat, 6 Feb 2016 00:24:46 +0000 (16:24 -0800)]
arm64: Drop alloc function from create_mapping
create_mapping is only used in fixmap_remap_fdt. All the create_mapping
calls need to happen on existing translation table pages without
additional allocations. Rather than have an alloc function be called
and fail, just set it to NULL and catch its use. Also change
the name to create_mapping_noalloc to better capture what exactly is
going on.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Laura Abbott <labbott@fedoraproject.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Andrew Pinski [Tue, 2 Feb 2016 12:46:26 +0000 (12:46 +0000)]
arm64: lib: patch in prfm for copy_page if requested
On ThunderX T88 pass 1 and pass 2, there is no hardware prefetching so
we need to patch in explicit software prefetching instructions
Prefetching improves this code by 60% over the original code and 2x
over the code without prefetching for the affected hardware using the
benchmark code at https://github.com/apinski-cavium/copy_page_benchmark
Signed-off-by: Andrew Pinski <apinski@cavium.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Tested-by: Andrew Pinski <apinski@cavium.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Will Deacon [Tue, 2 Feb 2016 12:46:25 +0000 (12:46 +0000)]
arm64: lib: improve copy_page to deal with 128 bytes at a time
We want to avoid lots of different copy_page implementations, settling
for something that is "good enough" everywhere and hopefully easy to
understand and maintain whilst we're at it.
This patch reworks our copy_page implementation based on discussions
with Cavium on the list and benchmarking on Cortex-A processors so that:
- The loop is unrolled to copy 128 bytes per iteration
- The reads are offset so that we read from the next 128-byte block
in the same iteration that we store the previous block
- Explicit prefetch instructions are removed for now, since they hurt
performance on CPUs with hardware prefetching
- The loop exit condition is calculated at the start of the loop
Signed-off-by: Will Deacon <will.deacon@arm.com> Tested-by: Andrew Pinski <apinski@cavium.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Will Deacon [Tue, 2 Feb 2016 12:46:24 +0000 (12:46 +0000)]
arm64: prefetch: add alternative pattern for CPUs without a prefetcher
Most CPUs have a hardware prefetcher which generally performs better
without explicit prefetch instructions issued by software, however
some CPUs (e.g. Cavium ThunderX) rely solely on explicit prefetch
instructions.
This patch adds an alternative pattern (ARM64_HAS_NO_HW_PREFETCH) to
allow our library code to make use of explicit prefetch instructions
during things like copy routines only when the CPU does not have the
capability to perform the prefetching itself.
Signed-off-by: Will Deacon <will.deacon@arm.com> Tested-by: Andrew Pinski <apinski@cavium.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Will Deacon [Tue, 2 Feb 2016 12:46:23 +0000 (12:46 +0000)]
arm64: prefetch: don't provide spin_lock_prefetch with LSE
The LSE atomics rely on us not dirtying data at L1 if we can avoid it,
otherwise many of the potential scalability benefits are lost.
This patch replaces spin_lock_prefetch with a nop when the LSE atomics
are in use, so that users don't shoot themselves in the foot by causing
needless coherence traffic at L1.
Signed-off-by: Will Deacon <will.deacon@arm.com> Tested-by: Andrew Pinski <apinski@cavium.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
projects / karo-tx-linux.git / log