This series adds support for building the powerpc 64-bit
LE kernel using the new ABI v2. We already supported
running ABI v2 userspace programs but this adds support
for building the kernel itself using the new ABI.
Unaligned stores take alignment exceptions on POWER7 running in little-endian.
This is a dumb little-endian base memcpy that prevents unaligned stores.
Once booted the feature fixup code switches over to the VMX copy loops
(which are already endian safe).
The question is what we do before that switch over. The base 64bit
memcpy takes alignment exceptions on POWER7 so we can't use it as is.
Fixing the causes of alignment exception would slow it down, because
we'd need to ensure all loads and stores are aligned either through
rotate tricks or bytewise loads and stores. Either would be bad for
all other 64bit platforms.
[ I simplified the loop a bit - Anton ]
Signed-off-by: Philippe Bergheaud <felix@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If an assembly function that calls back into c code is exported to
modules, we need to ensure r2 is setup correctly. There are only
two places crazy enough to do it (two of which are my fault).
Signed-off-by: Anton Blanchard <anton@samba.org>
Some of the assembler files in lib/ make use of the fact that in the
ELFv1 ABI, the caller guarantees to provide stack space to save the
parameter registers r3 ... r10. This guarantee is no longer present
in ELFv2 for functions that have no variable argument list and no
more than 8 arguments.
Change the affected routines to temporarily store registers in the
red zone and/or the top of their own stack frame (in the space
provided to save r31 .. r29, which is actually not used in these
routines).
In opal_query_takeover, simply always allocate a stack frame;
the routine is not performance critical.
Signed-off-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
binutils is smart enough to know that a branch to a function
descriptor is actually a branch to the functions text address.
Alan tells me that binutils has been doing this for 9 years.
Signed-off-by: Anton Blanchard <anton@samba.org>
Turn Anton's memcpy / copy_tofrom_user test into something that can
live in tools/testing/selftests.
It requires one turd in arch/powerpc/lib/memcpy_64.S, but it's pretty
harmless IMHO.
We are sailing very close to the wind with the feature macros. We define
them to nothing, which currently means we get a few extra nops and
include the unaligned calls.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
GCC 4.8 now generates out-of-line vr save/restore functions when
optimizing for size. They are needed for the raid6 altivec support.
Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Merge a pile of fixes that went into the "merge" branch (3.13-rc's) such
as Anton Little Endian fixes.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The powerpc 64-bit __copy_tofrom_user() function uses shifts to handle
unaligned invocations. However, these shifts were designed for
big-endian systems: On little-endian systems, they must shift in the
opposite direction.
This commit relies on the C preprocessor to insert the correct shifts
into the assembly code.
[ This is a rare but nasty LE issue. Most of the time we use the POWER7
optimised __copy_tofrom_user_power7 loop, but when it hits an exception
we fall back to the base __copy_tofrom_user loop. - Anton ]
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
So that it can be used by other codes. No function change.
Signed-off-by: Kevin Hao <haokexin@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Add a VMX optimised xor, used primarily for RAID5. On a POWER7 blade
this is a decent win:
32regs : 17932.800 MB/sec
altivec : 19724.800 MB/sec
The bigger gain is when the same test is run in SMT4 mode, as it
would if there was a lot of work going on:
8regs : 8377.600 MB/sec
altivec : 15801.600 MB/sec
I tested this against an array created without the patch, and also
verified it worked as expected on a little endian kernel.
[ Fix !CONFIG_ALTIVEC build -- BenH ]
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This patch addresses unaligned single precision floating point loads
and stores in the single-step code. The old implementation
improperly treated an 8 byte structure as an array of two 4 byte
words, which is a classic little endian bug.
Signed-off-by: Tom Musta <tmusta@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This patch modifies the unaligned access routines of the sstep.c
module so that it properly reverses the bytes of storage operands
in the little endian kernel kernel. This is implemented by
breaking an unaligned little endian access into a combination of
single byte accesses plus an overal byte reversal operation.
Signed-off-by: Tom Musta <tmusta@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We need to fix some endian issues in our memcpy code. For now
just enable the generic memcpy routine for little endian builds.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We need to fix some endian issues in our checksum code. For now
just enable the generic checksum routines for little endian builds.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Fix the permute loops for little endian.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The csum_partial_copy_generic() function saves the PowerPC non-volatile
r14, r15, and r16 registers for the main checksum-and-copy loop.
Unfortunately, it fails to restore them upon error exit from this loop,
which results in silent corruption of these registers in the presumably
rare event of an access exception within that loop.
This commit therefore restores these register on error exit from the loop.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: stable@vger.kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The csum_partial_copy_generic() uses register r7 to adjust the remaining
bytes to process. Unfortunately, r7 also holds a parameter, namely the
address of the flag to set in case of access exceptions while reading
the source buffer. Lacking a quantum implementation of PowerPC, this
commit instead uses register r9 to do the adjusting, leaving r7's
pointer uncorrupted.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: stable@vger.kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We've been keeping that field in thread_struct for a while, it contains
the "limit" of the current stack pointer and is meant to be used for
detecting stack overflows.
It has a few problems however:
- First, it was never actually *used* on 64-bit. Set and updated but
not actually exploited
- When switching stack to/from irq and softirq stacks, it's update
is racy unless we hard disable interrupts, which is costly. This
is fine on 32-bit as we don't soft-disable there but not on 64-bit.
Thus rather than fixing 2 in order to implement 1 in some hypothetical
future, let's remove the code completely from 64-bit. In order to avoid
a clutter of ifdef's, we remove the updates from C code completely
during interrupt stack switching, and instead maintain it from the
asm helper that is used to do the stack switching in the first place.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The stmw instruction was incorrectly decoded as an update form instruction
and thus the RA register was being clobbered.
Also, the utility routine to write memory to unaligned addresses breaks the
operation into smaller aligned accesses but was incorrectly incrementing
the address by only one; it needs to increment the address by the size of
the smaller aligned chunk.
Signed-off-by: Tom Musta <tmusta@us.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The lppaca, slb_shadow and dtl_entry hypervisor structures are
big endian, so we have to byte swap them in little endian builds.
LE KVM hosts will also need to be fixed but for now add an #error
to remind us.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Check truncate_if_32bit() on final write to nip.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
No code changes, just documenting what's happening a little better.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Uprobes uses emulate_step in sstep.c, but we haven't explicitly specified
the dependency. On pseries HAVE_HW_BREAKPOINT protects us, but 44x has no
such luxury.
Consolidate other users that depend on sstep and create a new config option.
Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Signed-off-by: Suzuki K. Poulose <suzuki@in.ibm.com>
Cc: linuxppc-dev@ozlabs.org
Cc: stable@vger.kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Finally remove the two level TOC and build with -mcmodel=medium.
Unfortunately we can't build modules with -mcmodel=medium due to
the tricks the kernel module loader plays with percpu data:
# -mcmodel=medium breaks modules because it uses 32bit offsets from
# the TOC pointer to create pointers where possible. Pointers into the
# percpu data area are created by this method.
#
# The kernel module loader relocates the percpu data section from the
# original location (starting with 0xd...) to somewhere in the base
# kernel percpu data space (starting with 0xc...). We need a full
# 64bit relocation for this to work, hence -mcmodel=large.
On older kernels we fall back to the two level TOC (-mminimal-toc)
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We don't do the real store operation for kprobing 'stwu Rx,(y)R1'
since this may corrupt the exception frame, now we will do this
operation safely in exception return code after migrate current
exception frame below the kprobed function stack.
So we only update gpr[1] here and trigger a thread flag to mask
this.
Note we should make sure if we trigger kernel stack over flow.
Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
patch_instruction() can be called very early on ppc32, when the kernel
isn't yet running at it's linked address. That can cause the !
is_kernel_addr() test in __put_user() to trip and call might_sleep()
which is very bad at that point during boot.
Use a lower level function instead for now, at least until we get to
rework ppc32 boot process to do the code patching later, like ppc64
does.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The enhanced prefetch hint patches corrupt the condition register
that was used to check if we are in interrupt. Fix this by using cr1.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
"powerpc: Use enhanced touch instructions in POWER7
copy_to_user/copy_from_user" was applied twice. Remove one.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This allows the linker to know that calls to them do not need to switch
TOC and stop errors like the following when linking large configurations:
powerpc64-linux-ld: drivers/built-in.o: In function `.gpiochip_is_requested':
(.text+0x4): sibling call optimization to `_savegpr0_29' does not allow automatic multiple TOCs; recompile with -mminimal-toc or -fno-optimize-sibling-calls, or make `_savegpr0_29' extern
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
These macros are using integers where they could be using logical
names since they take registers.
We are going to enforce this soon, so fix these up now.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
mtocrf define is just a wrapper around the real instructions so we can
just use real register names here (ie. lower case).
Also remove braces in macro so this is possible.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Merge the defines of STACKFRAMESIZE, STK_REG, STK_PARAM from different
places.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Anything that uses a constructed instruction (ie. from ppc-opcode.h),
need to use the new R0 macro, as %r0 is not going to work.
Also convert usages of macros where we are just determining an offset
(usually for a load/store), like:
std r14,STK_REG(r14)(r1)
Can't use STK_REG(r14) as %r14 doesn't work in the STK_REG macro since
it's just calculating an offset.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
I blame Mikey for this. He elevated my slightly dubious testcase:
to benchmark status. And naturally we need to be number 1 at creating
zeros. So lets improve __clear_user some more.
As Paul suggests we can use dcbz for large lengths. This patch gets
the destination cacheline aligned then uses dcbz on whole cachelines.
Before:
10485760000 bytes (10 GB) copied, 0.414744 s, 25.3 GB/s
After:
10485760000 bytes (10 GB) copied, 0.268597 s, 39.0 GB/s
39 GB/s, a new record.
Signed-off-by: Anton Blanchard <anton@samba.org>
Tested-by: Olof Johansson <olof@lixom.net>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Implement a POWER7 optimised memcpy using VMX and enhanced prefetch
instructions.
This is a copy of the POWER7 optimised copy_to_user/copy_from_user
loop. Detailed implementation and performance details can be found in
commit a66086b819 (powerpc: POWER7 optimised
copy_to_user/copy_from_user using VMX).
I noticed memcpy issues when profiling a RAID6 workload:
.memcpy
.async_memcpy
.async_copy_data
.__raid_run_ops
.handle_stripe
.raid5d
.md_thread
I created a simplified testcase by building a RAID6 array with 4 1GB
ramdisks (booting with brd.rd_size=1048576):
# mdadm -CR -e 1.2 /dev/md0 --level=6 -n4 /dev/ram[0-3]
I then timed how long it took to write to the entire array:
# dd if=/dev/zero of=/dev/md0 bs=1M
Before: 892 MB/s
After: 999 MB/s
A 12% improvement.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Version 2.06 of the POWER ISA introduced enhanced touch instructions,
allowing us to specify a number of attributes including the length of
a stream.
This patch adds a software stream for both loads and stores in the
POWER7 copy_tofrom_user loop. Since the setup is quite complicated
and we have to use an eieio to ensure correct ordering of the "GO"
command we only do this for copies above 4kB.
To quantify any performance improvements we need a working set
bigger than the caches so we operate on a 1GB file:
# dd if=/dev/zero of=/tmp/foo bs=1M count=1024
And we compare how fast we can read the file:
# dd if=/tmp/foo of=/dev/null bs=1M
before: 7.7 GB/s
after: 9.6 GB/s
A 25% improvement.
The worst case for this patch will be a completely L1 cache contained
copy of just over 4kB. We can test this with the copy_to_user
testcase we used to tune copy_tofrom_user originally:
http://ozlabs.org/~anton/junkcode/copy_to_user.c
# time ./copy_to_user2 -l 4224 -i 10000000
before: 6.807 s
after: 6.946 s
A 2% slowdown, which seems reasonable considering our data is unlikely
to be completely L1 contained.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Implement a POWER7 optimised copy_page using VMX and enhanced
prefetch instructions. We use enhanced prefetch hints to prefetch
both the load and store side. We copy a cacheline at a time and
fall back to regular loads and stores if we are unable to use VMX
(eg we are in an interrupt).
The following microbenchmark was used to assess the impact of
the patch:
http://ozlabs.org/~anton/junkcode/page_fault_file.c
We test MAP_PRIVATE page faults across a 1GB file, 100 times:
# time ./page_fault_file -p -l 1G -i 100
Before: 22.25s
After: 18.89s
17% faster
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Subsequent patches will add more VMX library functions and it makes
sense to keep all the c-code helper functions in the one file.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Version 2.06 of the POWER ISA introduced enhanced touch instructions,
allowing us to specify a number of attributes including the length of
a stream.
This patch adds a software stream for both loads and stores in the
POWER7 copy_tofrom_user loop. Since the setup is quite complicated
and we have to use an eieio to ensure correct ordering of the "GO"
command we only do this for copies above 4kB.
To quantify any performance improvements we need a working set
bigger than the caches so we operate on a 1GB file:
# dd if=/dev/zero of=/tmp/foo bs=1M count=1024
And we compare how fast we can read the file:
# dd if=/tmp/foo of=/dev/null bs=1M
before: 7.7 GB/s
after: 9.6 GB/s
A 25% improvement.
The worst case for this patch will be a completely L1 cache contained
copy of just over 4kB. We can test this with the copy_to_user
testcase we used to tune copy_tofrom_user originally:
http://ozlabs.org/~anton/junkcode/copy_to_user.c
# time ./copy_to_user2 -l 4224 -i 10000000
before: 6.807 s
after: 6.946 s
A 2% slowdown, which seems reasonable considering our data is unlikely
to be completely L1 contained.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
I noticed __clear_user high up in a profile of one of my RAID stress
tests. The testcase was doing a dd from /dev/zero which ends up
calling __clear_user.
__clear_user is basically a loop with a single 4 byte store which
is horribly slow. We can do much better by aligning the desination
and doing 32 bytes of 8 byte stores in a loop.
The following testcase was used to verify the patch:
http://ozlabs.org/~anton/junkcode/stress_clear_user.c
To show the improvement in performance I ran a dd from /dev/zero
to /dev/null on a POWER7 box:
Before:
# dd if=/dev/zero of=/dev/null bs=1M count=10000
10485760000 bytes (10 GB) copied, 3.72379 s, 2.8 GB/s
After:
# time dd if=/dev/zero of=/dev/null bs=1M count=10000
10485760000 bytes (10 GB) copied, 0.728318 s, 14.4 GB/s
Over 5x faster.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
For ftrace to use the patch_instruction code, it needs to check for
faults on write. Ftrace updates code all over the kernel, and we need to
know if code is updated or not due to protections that are placed on
some portions of the kernel. If ftrace does not detect a fault, it will
error later on, and it will be much more difficult to find the problem.
By changing patch_instruction() to detect faults, then ftrace will be
able to make use of it too.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This is much the same as for SPARC except that we can do the find_zero()
function more efficiently using the count-leading-zeroes instructions.
Tested on 32-bit and 64-bit PowerPC.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove CONFIG_POWER4_ONLY, the option is badly named and only does two
things:
- It wraps the MMU segment table code. With feature fixups there is
little downside to compiling this in.
- It uses the newer mtocrf instruction in various assembly functions.
Instead of making this a compile option just do it at runtime via
a feature fixup.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Disintegrate asm/system.h for PowerPC.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
cc: linuxppc-dev@lists.ozlabs.org
This is no longer selectable, so just remove all the dependent code.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Implement a POWER7 optimised copy_to_user/copy_from_user using VMX.
For large aligned copies this new loop is over 10% faster, and for
large unaligned copies it is over 200% faster.
If we take a fault we fall back to the old version, this keeps
things relatively simple and easy to verify.
On POWER7 unaligned stores rarely slow down - they only flush when
a store crosses a 4KB page boundary. Furthermore this flush is
handled completely in hardware and should be 20-30 cycles.
Unaligned loads on the other hand flush much more often - whenever
crossing a 128 byte cache line, or a 32 byte sector if either sector
is an L1 miss.
Considering this information we really want to get the loads aligned
and not worry about the alignment of the stores. Microbenchmarks
confirm that this approach is much faster than the current unaligned
copy loop that uses shifts and rotates to ensure both loads and
stores are aligned.
We also want to try and do the stores in cacheline aligned, cacheline
sized chunks. If the store queue is unable to merge an entire
cacheline of stores then the L2 cache will have to do a
read/modify/write. Even worse, we will serialise this with the stores
in the next iteration of the copy loop since both iterations hit
the same cacheline.
Based on this, the new loop does the following things:
1 - 127 bytes
Get the source 8 byte aligned and use 8 byte loads and stores. Pretty
boring and similar to how the current loop works.
128 - 4095 bytes
Get the source 8 byte aligned and use 8 byte loads and stores,
1 cacheline at a time. We aren't doing the stores in cacheline
aligned chunks so we will potentially serialise once per cacheline.
Even so it is much better than the loop we have today.
4096 - bytes
If both source and destination have the same alignment get them both
16 byte aligned, then get the destination cacheline aligned. Do
cacheline sized loads and stores using VMX.
If source and destination do not have the same alignment, we get the
destination cacheline aligned, and use permute to do aligned loads.
In both cases the VMX loop should be optimal - we always do aligned
loads and stores and are always doing stores in cacheline aligned,
cacheline sized chunks.
To be able to use VMX we must be careful about interrupts and
sleeping. We don't use the VMX loop when in an interrupt (which should
be rare anyway) and we wrap the VMX loop in disable/enable_pagefault
and fall back to the existing copy_tofrom_user loop if we do need to
sleep.
The VMX breakpoint of 4096 bytes was chosen using this microbenchmark:
http://ozlabs.org/~anton/junkcode/copy_to_user.c
Since we are using VMX and there is a cost to saving and restoring
the user VMX state there are two broad cases we need to benchmark:
- Best case - userspace never uses VMX
- Worst case - userspace always uses VMX
In reality a userspace process will sit somewhere between these two
extremes. Since we need to test both aligned and unaligned copies we
end up with 4 combinations. The point at which the VMX loop begins to
win is:
0% VMX
aligned 2048 bytes
unaligned 2048 bytes
100% VMX
aligned 16384 bytes
unaligned 8192 bytes
Considering this is a microbenchmark, the data is hot in cache and
the VMX loop has better store queue merging properties we set the
breakpoint to 4096 bytes, a little below the unaligned breakpoints.
Some future optimisations we can look at:
- Looking at the perf data, a significant part of the cost when a
task is always using VMX is the extra exception we take to restore
the VMX state. As such we should do something similar to the x86
optimisation that restores FPU state for heavy users. ie:
/*
* If the task has used fpu the last 5 timeslices, just do a full
* restore of the math state immediately to avoid the trap; the
* chances of needing FPU soon are obviously high now
*/
preload_fpu = tsk_used_math(next_p) && next_p->fpu_counter > 5;
and
/*
* fpu_counter contains the number of consecutive context switches
* that the FPU is used. If this is over a threshold, the lazy fpu
* saving becomes unlazy to save the trap. This is an unsigned char
* so that after 256 times the counter wraps and the behavior turns
* lazy again; this to deal with bursty apps that only use FPU for
* a short time
*/
- We could create a paca bit to mirror the VMX enabled MSR bit and check
that first, avoiding multiple calls to calling enable_kernel_altivec.
That should help with iovec based system calls like readv.
- We could have two VMX breakpoints, one for when we know the user VMX
state is loaded into the registers and one when it isn't. This could
be a second bit in the paca so we can calculate the break points quickly.
- One suggestion from Ben was to save and restore the VSX registers
we use inline instead of using enable_kernel_altivec.
[BenH: Fixed a problem with preempt and fixed build without CONFIG_ALTIVEC]
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
kdump fails because we try to execute an HV only instruction. Feature
fixups are being applied after we copy the exception vectors down to 0
so they miss out on any updates.
We have always had this issue but it only became critical in v3.0
when we added CFAR support (breaks POWER5) and v3.1 when we added
POWERNV (breaks everyone).
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: <stable@kernel.org> [v3.0+]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
