Commit Graph

333 Commits

Author SHA1 Message Date
Elliott Hughes
f2a760dca7 am a85606e1: am c100a100: Merge "\'Avoid confusing "read prevented write" log messages\' 2."
* commit 'a85606e1563c2153bea3c73dfe4ca1588e778f22':
  'Avoid confusing "read prevented write" log messages' 2.
2013-10-15 17:38:17 -07:00
Elliott Hughes
68b67113a4 'Avoid confusing "read prevented write" log messages' 2.
This time it's assembler.

Change-Id: Iae6369833b8046b8eda70238bb4ed0cae64269ea
2013-10-15 17:17:05 -07:00
Christopher Ferris
289c460c55 am ac6bc319: Remove new aligned memcpy path for cortex-a15.
* commit 'ac6bc31942e58c8893c0695d9766d0f3e39335fe':
  Remove new aligned memcpy path for cortex-a15.
2013-10-15 16:17:14 -07:00
Christopher Ferris
ac6bc31942 Remove new aligned memcpy path for cortex-a15.
For some reason the new cortex-a15 memcpy code from ARM is really bad
for really large copies. This change forces us to go down the old path
for all copies.

All of my benchmarks show the new version is faster for large copies, but
something is going on that I don't understand.

Bug: 10838353
Change-Id: I01c16d4a2575e76f4c69862c6f78fd9024eb3fb8
2013-10-15 14:54:02 -07:00
Christopher Ferris
d77ff4f423 am 76d78dc0: am dc9d8d05: Merge "Modify prefetch for krait memcpy."
* commit '76d78dc0441ba982438ef613a8d6f2f65c6834e0':
  Modify prefetch for krait memcpy.
2013-10-15 14:19:27 -07:00
Elliott Hughes
4681f8244e am 5b3f1d8b: am 22b83da4: Merge "Clean up the sigprocmask/pthread_sigmask implementation."
* commit '5b3f1d8b3a6a1caf059dc376e4967a27ddd57dac':
  Clean up the sigprocmask/pthread_sigmask implementation.
2013-10-15 14:16:02 -07:00
Christopher Ferris
dc9d8d050a Merge "Modify prefetch for krait memcpy." 2013-10-15 21:12:31 +00:00
Christopher Ferris
c3c58fb560 Modify prefetch for krait memcpy.
I originally modified the krait mainloop prefetch from cacheline * 8 to * 2.
This causes a perf degradation for copies bigger than will fit in the cache.
Fixing this back to the original * 8. I tried other multiples, but * 8 is th
sweet spot on krait.

Bug: 11221806

Change-Id: I1f75fad6440f7417e664795a6e7b5616f6a29c45
2013-10-15 12:10:06 -07:00
Elliott Hughes
19e62325c2 Clean up the sigprocmask/pthread_sigmask implementation.
Let's have both use rt_sigprocmask, like in glibc. The 64-bit ABIs
can share the same code as the 32-bit ABIs.

Also, let's test the return side of these calls, not just the
setting.

Bug: 11069919
Change-Id: I11da99f85b5b481870943c520d05ec929b15eddb
2013-10-15 11:23:57 -07:00
Elliott Hughes
2d38f0c1c7 am 21d70d5f: am 39385aa7: Merge "Clean up the cpuacct cruft."
* commit '21d70d5fd8ff2823a6879d59c7d5a8d6c3e0ce8b':
  Clean up the cpuacct cruft.
2013-10-09 21:28:51 -07:00
Elliott Hughes
232163cf70 Clean up the cpuacct cruft.
Change-Id: I6ed63af8dfc2368e211420389fa8af4d5dc0908f
2013-10-09 17:35:36 -07:00
Elliott Hughes
1cdc9fe6d5 am 8826ba8a: am bc545e8a: Merge "Fix x86_64 build, clean up intermediate libraries."
* commit '8826ba8ab6922927bdac2466d2903916d5806a06':
  Fix x86_64 build, clean up intermediate libraries.
2013-10-09 16:35:46 -07:00
Elliott Hughes
eb847bc866 Fix x86_64 build, clean up intermediate libraries.
The x86_64 build was failing because clone.S had a call to __thread_entry which
was being added to a different intermediate .a on the way to making libc.so,
and the linker couldn't guarantee statically that such a relocation would be
possible.

  ld: error: out/target/product/generic_x86_64/obj/STATIC_LIBRARIES/libc_common_intermediates/libc_common.a(clone.o): requires dynamic R_X86_64_PC32 reloc against '__thread_entry' which may overflow at runtime; recompile with -fPIC

This patch addresses that by ensuring that the caller and callee end up in the
same intermediate .a. While I'm here, I've tried to clean up some of the mess
that led to this situation too. In particular, this removes libc/private/ from
the default include path (except for the DNS code), and splits out the DNS
code into its own library (since it's a weird special case of upstream NetBSD
code that's diverged so heavily it's unlikely ever to get back in sync).

There's more cleanup of the DNS situation possible, but this is definitely a
step in the right direction, and it's more than enough to get x86_64 building
cleanly.

Change-Id: I00425a7245b7a2573df16cc38798187d0729e7c4
2013-10-09 16:00:17 -07:00
Elliott Hughes
42b618cffe am 58b8f225: am 5cf1f229: Merge "pthread_exit should call __NR_exit with status 0."
* commit '58b8f2256637c5b1e24b568b699fb3aa6cf0ca96':
  pthread_exit should call __NR_exit with status 0.
2013-10-08 15:06:49 -07:00
Elliott Hughes
c4c6e192ac pthread_exit should call __NR_exit with status 0.
We shouldn't have been passing the bottom 32 bits of the address used
for pthread_join to the kernel.

Change-Id: I487e5002d60c27adba51173719213abbee0f183f
2013-10-08 14:48:05 -07:00
Elliott Hughes
9258c7b23e am b7d77a9e: am 56777421: Merge "Add arch-x86_64/include/machine."
* commit 'b7d77a9eb1effd5f0472dff14a8fe77d9da2c7a2':
  Add arch-x86_64/include/machine.
2013-10-04 16:21:25 -07:00
Elliott Hughes
72645164b6 Add arch-x86_64/include/machine.
This is basically the other half of I5de76f6c46ac87779f207d568a86bb453e2414de
from Pavel Chupin <pavel.v.chupin@intel.com>, but taking the exact upstream
_types.h instead of the modified version. (I was confused when I suggested
otherwise.)

I've also cleaned up the internal_types.h situation; we weren't gaining
anything from these empty files, and there is no upstream internal_types.h
for x86_64.

Change-Id: I802a9a6a8df1c979e820659212c75a47c2ef392e
2013-10-04 16:10:13 -07:00
Nick Kralevich
7909bd19ab am 9da8b682: am c417d0a5: Merge "Make error messages even better!"
* commit '9da8b6829cb8febb70af46761b7f2ca19e6498ce':
  Make error messages even better!
2013-10-04 12:02:57 -07:00
Nick Kralevich
6861c6f85e Make error messages even better!
Change-Id: I72bd1eb1d526dc59833e5bc3c636171f7f9545af
2013-10-04 11:43:30 -07:00
Elliott Hughes
be7b6408b3 am 6d923d49: am 14102932: Merge "Move common arch-* code to arch-common directory"
* commit '6d923d497373c160f6fdddd42f5ed56e7bf0923a':
  Move common arch-* code to arch-common directory
2013-10-04 09:58:23 -07:00
Elliott Hughes
141029327c Merge "Move common arch-* code to arch-common directory" 2013-10-03 23:17:58 +00:00
Christopher Ferris
1a086a52d3 am 270c52da: am 1fe477fc: Merge "Add dependencies on included makefiles."
* commit '270c52da0db68c6961eb576b67adc0c41c418763':
  Add dependencies on included makefiles.
2013-10-03 14:57:21 -07:00
Christopher Ferris
d7a632948d Add dependencies on included makefiles.
Bug: 11050594

Merge from internal master.

(cherry-picked from f389284e86)

Change-Id: I2b3e38329a09d26c16870906f9ed1257e2a9dbc8
2013-10-03 14:17:14 -07:00
Christopher Ferris
bdc610aba5 am dc7f8ddc: am 6088047a: Merge "Remove dead files."
* commit 'dc7f8ddc227e833b889eb1eedc16b9b5827fc127':
  Remove dead files.
2013-10-03 13:26:38 -07:00
Christopher Ferris
fc4d70fe54 Remove dead files.
memcpy.a15.S/strcmp.a15.S files were submitted by ARM for use as the basis
for the memcpy/strcmp implementations in cortex-a15.

memset.S was moved in to the generic directory.

NOTE: memcpy.a9.S was submitted by Linaro to be the basis for the memcpy
for cortex-a9/cortex-a15 but has not been incorporated yet.

Bug: 10971279

Merge from internal master.

(cherry-picked from 48fc3e8b9f)

Change-Id: I8f9297578990d517f004e4e8840e2b2cbd5a47d8
2013-10-03 12:35:56 -07:00
Pavel Chupin
b49c17c2bf Move common arch-* code to arch-common directory
Will be helpful on adding x86_64

Change-Id: I96cf6fc7912c02f289c75f07ae0079c32d69173f
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
2013-10-03 11:14:33 +04:00
Christopher Ferris
de7ed18df0 am 0292c1c8: am 3901b1bd: Merge "Remove the __ARM_FEATURE_DSP check."
* commit '0292c1c81dfed23a32d46fb0b089b7eccb87e21e':
  Remove the __ARM_FEATURE_DSP check.
2013-10-02 23:27:06 -07:00
Christopher Ferris
aec1b3540a Remove the __ARM_FEATURE_DSP check.
The check for __ARM_FEATURE_DSP being defined is pointless since it
is always defined.

Bug: 10971279

Merge from internal master.

(cherry-picked from d2642fa70c)

Change-Id: If23ab3271f4da0c38cd531ffdc9a7e5eed6ec5dc
2013-10-02 23:14:01 -07:00
Nick Kralevich
670f372470 am b6201932: am 2c701a1b: Merge "libc: don\'t export unnecessary symbols"
* commit 'b6201932838c46a567e3411047719fd5c2797e96':
  libc: don't export unnecessary symbols
2013-10-02 17:03:42 -07:00
Nick Kralevich
32bbf8a63b libc: don't export unnecessary symbols
Symbols associated with the internal implementation of memcpy
like routines should be private.

Change-Id: I2b1d1f59006395c29d518c153928437b08f93d16
2013-10-02 16:54:58 -07:00
Elliott Hughes
5825f979b7 am 88f29444: am 7b538021: Merge "Make it easier to add syscalls for another architecture."
* commit '88f2944421839b725e02e97c62d119c9e608ec58':
  Make it easier to add syscalls for another architecture.
2013-09-26 09:02:46 -07:00
Elliott Hughes
d612165c67 Make it easier to add syscalls for another architecture.
Much of the per-architecture duplication can be removed, so let's do so
before we add the 64-bit architectures.

Change-Id: Ieb796503c8e5353ea38c3bab768bb9a690c9a767
2013-09-26 08:57:17 -07:00
Christopher Ferris
16e185c908 __memcpy_chk: Fix signed cmp of unsigned values.
I accidentally did a signed comparison of the size_t values passed in
for three of the _chk functions. Changing them to unsigned compares.

Add three new tests to verify this failure is fixed.

Bug: 10691831

Merge from internal master.

(cherry-picked from 883ef2499c)

Change-Id: Id9a96b549435f5d9b61dc132cf1082e0e30889f5
2013-09-20 20:12:09 -07:00
Christopher Ferris
a57c9c084b Fix all debug directives.
The backtrace when a fortify check failed was not correct. This change
adds all of the necessary directives to get a correct backtrace.

Fix the strcmp directives and change all labels to local labels.

Testing:
- Verify that the runtime can decode the stack for __memcpy_chk, __memset_chk,
  __strcpy_chk, __strcat_chk fortify failures.
- Verify that gdb can decode the stack properly when hitting a fortify check.
- Verify that the runtime can decode the stack for a seg fault for all of the
  _chk functions and for memcpy/memset.
- Verify that gdb can decode the stack for a seg fault for all of the _chk
  functions and for memcpy/memset.
- Verify that the runtime can decode the stack for a seg fault for strcmp.
- Verify that gdb can decode the stack for a seg fault in strcmp.

Bug: 10342460
Bug: 10345269

Merge from internal master.

(cherry-picked from 05332f2ce7)

Change-Id: Ibc919b117cfe72b9ae97e35bd48185477177c5ca
2013-09-20 18:59:58 -07:00
Christopher Ferris
bd7fe1d3c4 Update all debug directives.
The libcorkscrew stack unwinder does not understand cfi directives,
so add .save directives so that it can function properly.

Also add the directives in to strcmp.S and fix a missing set of
directives in cortex-a9/memcpy_base.S.

Bug: 10345269

Merge from internal master.

(cherry-picked from 5f7ccea3ff)

Change-Id: If48a216203216a643807f5d61906015984987189
2013-09-20 13:49:38 -07:00
Elliott Hughes
e74f77f92d am 48a909c9: am aad3c52e: Merge "Ensure we have the off64_t variant of every function that takes an off_t."
* commit '48a909c9fd6dbe9be5655ad172d1083fa69c4107':
  Ensure we have the off64_t variant of every function that takes an off_t.
2013-09-19 17:43:01 -07:00
Elliott Hughes
b4f7616fd6 Ensure we have the off64_t variant of every function that takes an off_t.
Change-Id: Ib2eee0cf13162be3b62559b84e90c6dcf5aab1c3
2013-09-19 16:27:24 -07:00
Christopher Ferris
2e7a8b44c2 resolved conflicts for merge of 883ef249 to klp-dev-plus-aosp
Change-Id: I1e2bd03a0cb5a0ab191c525d1574377bc7fd90ab
2013-09-10 19:49:04 -07:00
Christopher Ferris
883ef2499c __memcpy_chk: Fix signed cmp of unsigned values.
I accidentally did a signed comparison of the size_t values passed in
for three of the _chk functions. Changing them to unsigned compares.

Add three new tests to verify this failure is fixed.

Bug: 10691831
Change-Id: Ia831071f7dffd5972a748d888dd506c7cc7ddba3
2013-09-10 17:34:03 -07:00
Christopher Ferris
5cb0436930 am 1a88ca08: am 98c726ec: Merge "Add the dl_iterate_phdr function to libdl for arm."
* commit '1a88ca08046ea510bfc8d3de6875537f124b3ce3':
  Add the dl_iterate_phdr function to libdl for arm.
2013-09-06 10:57:40 -07:00
Christopher Ferris
24053a461e Add the dl_iterate_phdr function to libdl for arm.
Bug: 8410085

Merge from internal master.

(cherry-picked from cb491bc66d)

Change-Id: I94ed51bc5d4c626df7552c0e85c31ccee2d6568f
2013-09-06 09:53:54 -07:00
Christopher Ferris
153ec2a76f am 49c0d471: Merge "Fix all debug directives." into klp-dev
* commit '49c0d471a8ba9ed32e7fdadb7e4ea6118b4b4af4':
  Fix all debug directives.
2013-08-29 14:23:06 -07:00
Christopher Ferris
05332f2ce7 Fix all debug directives.
The backtrace when a fortify check failed was not correct. This change
adds all of the necessary directives to get a correct backtrace.

Fix the strcmp directives and change all labels to local labels.

Testing:
- Verify that the runtime can decode the stack for __memcpy_chk, __memset_chk,
  __strcpy_chk, __strcat_chk fortify failures.
- Verify that gdb can decode the stack properly when hitting a fortify check.
- Verify that the runtime can decode the stack for a seg fault for all of the
  _chk functions and for memcpy/memset.
- Verify that gdb can decode the stack for a seg fault for all of the _chk
  functions and for memcpy/memset.
- Verify that the runtime can decode the stack for a seg fault for strcmp.
- Verify that gdb can decode the stack for a seg fault in strcmp.

Bug: 10342460
Bug: 10345269

Change-Id: I1dedadfee207dce4a285e17a21e8952bbc63786a
2013-08-28 15:42:05 -07:00
Nick Kralevich
f6186e9592 am 1278ae38: am 22f5ef6b: Merge "Delete CAVEATS / fix spelling."
* commit '1278ae38366102369b5a35152f2213426a148512':
  Delete CAVEATS / fix spelling.
2013-08-27 17:11:06 -07:00
Nick Kralevich
bdbdbb8319 Delete CAVEATS / fix spelling.
Change-Id: I0ed504271b7c2e4434d0d5f53bc10335c8cf7b5b
2013-08-27 17:05:19 -07:00
Christopher Ferris
5f7ccea3ff Update all debug directives.
The libcorkscrew stack unwinder does not understand cfi directives,
so add .save directives so that it can function properly.

Also add the directives in to strcmp.S and fix a missing set of
directives in cortex-a9/memcpy_base.S.

Bug: 10345269

Change-Id: I043f493e0bb6c45bd3f4906fbe1d9f628815b015
2013-08-20 11:22:34 -07:00
Christopher Ferris
5f45d583b0 Create optimized __strcpy_chk/__strcat_chk.
This change pulls the memcpy code out into a new file so that the
__strcpy_chk and __strcat_chk can use it with an include.

The new versions of the two chk functions uses assembly versions
of strlen and memcpy to implement this check. This allows near
parity with the assembly versions of strcpy/strcat. It also means that
as memcpy implementations get faster, so do the chk functions.

Other included changes:
- Change all of the assembly labels to local labels. The other labels
  confuse gdb and mess up backtracing.
- Add .cfi_startproc and .cfi_endproc directives so that gdb is not
  confused when falling through from one function to another.
- Change all functions to use cfi directives since they are more powerful.
- Move the memcpy_chk fail code outside of the memcpy function definition
  so that backtraces work properly.
- Preserve lr before the calls to __fortify_chk_fail so that the backtrace
  actually works.

Testing:

- Ran the bionic unit tests. Verified all error messages in logs are set
  correctly.
- Ran libc_test, replacing strcpy with __strcpy_chk and replacing
  strcat with __strcat_chk.
- Ran the debugger on nexus10, nexus4, and old nexus7. Verified that the
  backtrace is correct for all fortify check failures. Also verify that
  when falling through from __memcpy_chk to memcpy that the backtrace is
  still correct. Also verified the same for __memset_chk and bzero.
  Verified the two different paths in the cortex-a9 memset routine that
  save variables to the stack still show the backtrace properly.

Bug: 9293744

(cherry-picked from 2be91915dc)

Change-Id: Ia407b74d3287d0b6af0139a90b6eb3bfaebf2155
2013-08-15 11:13:39 -07:00
Christopher Ferris
59a13c122e Optimize __memset_chk, __memcpy_chk. DO NOT MERGE.
This change creates assembler versions of __memcpy_chk/__memset_chk
that is implemented in the memcpy/memset assembler code. This change
avoids an extra call to memcpy/memset, instead allowing a simple fall
through to occur from the chk code into the body of the real
implementation.

Testing:

- Ran the libc_test on __memcpy_chk/__memset_chk on all nexus devices.
- Wrote a small test executable that has three calls to __memcpy_chk and
  three calls to __memset_chk. First call dest_len is length + 1. Second
  call dest_len is length. Third call dest_len is length - 1.
  Verified that the first two calls pass, and the third fails. Examined
  the logcat output on all nexus devices to verify that the fortify
  error message was sent properly.
- I benchmarked the new __memcpy_chk and __memset_chk on all systems. For
  __memcpy_chk and large copies, the savings is relatively small (about 1%).
  For small copies, the savings is large on cortex-a15/krait devices
  (between 5% to 30%).
  For cortex-a9 and small copies, the speed up is present, but relatively
  small (about 3% to 5%).
  For __memset_chk and large copies, the savings is also small (about 1%).
  However, all processors show larger speed-ups on small copies (about 30% to
  100%).

Bug: 9293744

Merge from internal master.

(cherry-picked from 7c860db074)

Change-Id: I916ad305e4001269460ca6ebd38aaa0be8ac7f52
2013-08-14 18:14:43 -07:00
Christopher Ferris
f0c3d90913 Create optimized __strcpy_chk/__strcat_chk.
This change pulls the memcpy code out into a new file so that the
__strcpy_chk and __strcat_chk can use it with an include.

The new versions of the two chk functions uses assembly versions
of strlen and memcpy to implement this check. This allows near
parity with the assembly versions of strcpy/strcat. It also means that
as memcpy implementations get faster, so do the chk functions.

Other included changes:
- Change all of the assembly labels to local labels. The other labels
  confuse gdb and mess up backtracing.
- Add .cfi_startproc and .cfi_endproc directives so that gdb is not
  confused when falling through from one function to another.
- Change all functions to use cfi directives since they are more powerful.
- Move the memcpy_chk fail code outside of the memcpy function definition
  so that backtraces work properly.
- Preserve lr before the calls to __fortify_chk_fail so that the backtrace
  actually works.

Testing:

- Ran the bionic unit tests. Verified all error messages in logs are set
  correctly.
- Ran libc_test, replacing strcpy with __strcpy_chk and replacing
  strcat with __strcat_chk.
- Ran the debugger on nexus10, nexus4, and old nexus7. Verified that the
  backtrace is correct for all fortify check failures. Also verify that
  when falling through from __memcpy_chk to memcpy that the backtrace is
  still correct. Also verified the same for __memset_chk and bzero.
  Verified the two different paths in the cortex-a9 memset routine that
  save variables to the stack still show the backtrace properly.

Bug: 9293744
Change-Id: Id5aec8c3cb14101d91bd125eaf3770c9c8aa3f57
(cherry picked from commit 2be91915dc)
2013-08-14 07:46:00 +00:00
Christopher Ferris
b922ed3498 Fix strcpy.c that should have been strcpy.S. DO NOT MERGE
Merge from internal master.

(cherry-picked from 1ce6654163)

Change-Id: I376b831df42248baadde7202a30a68112f752ff7
2013-08-08 12:09:37 -07:00
Christopher Ferris
4e24dcc8d8 Optimize strcat/strcpy, small tweaks to strlen. DO NOT MERGE
Create one version of strcat/strcpy/strlen for cortex-a15/krait and another
version for cortex-a9.

Tested with the libc_test strcat/strcpy/strlen tests.
Including new tests that verify that the src for strcat/strcpy do not
overread across page boundaries.

NOTE: The handling of unaligned strcpy (same code in strcat) could probably
be optimized further such that the src is read 64 bits at a time instead of
the partial reads occurring now.

strlen improves slightly since it was recently optimized.

Performance improvements for strcpy and strcat (using an empty dest string):

cortex-a9
- Small copies vary from about 5% to 20% as the size gets above 10 bytes.
- Copies >= 1024, about a 60% improvement.
- Unaligned copies, from about 40% improvement.

cortex-a15
- Most small copies exhibit a 100% improvement, a few copies only
  improve by 20%.
- Copies >= 1024, about 150% improvement.
- Unaligned copies, about 100% improvement.

krait
- Most small copies vary widely, but on average 20% improvement, then
  the performance gets better, hitting about a 100% improvement when
  copies 64 bytes of data.
- Copies >= 1024, about 100% improvement.
- When coping MBs of data, about 50% improvement.
- Unaligned copies, about 90% improvement.

As strcat destination strings get larger in size:

cortex-a9
- about 40% improvement for small dst strings (>= 32).
- about 250% improvement for dst strings >= 1024.

cortex-a15
- about 200% improvement for small dst strings (>=32).
- about 250% improvement for dst strings >= 1024.

krait
- about 25% improvement for small dst strings (>=32).
- about 100% improvement for dst strings >=1024.

Merge from internal master.

(cherry-picked from d119b7b6f4)

Change-Id: I296463b251ef9fab004ee4dded2793feca5b547a
2013-08-08 11:13:46 -07:00
Christopher Ferris
7c860db074 Optimize __memset_chk, __memcpy_chk.
This change creates assembler versions of __memcpy_chk/__memset_chk
that is implemented in the memcpy/memset assembler code. This change
avoids an extra call to memcpy/memset, instead allowing a simple fall
through to occur from the chk code into the body of the real
implementation.

Testing:

- Ran the libc_test on __memcpy_chk/__memset_chk on all nexus devices.
- Wrote a small test executable that has three calls to __memcpy_chk and
  three calls to __memset_chk. First call dest_len is length + 1. Second
  call dest_len is length. Third call dest_len is length - 1.
  Verified that the first two calls pass, and the third fails. Examined
  the logcat output on all nexus devices to verify that the fortify
  error message was sent properly.
- I benchmarked the new __memcpy_chk and __memset_chk on all systems. For
  __memcpy_chk and large copies, the savings is relatively small (about 1%).
  For small copies, the savings is large on cortex-a15/krait devices
  (between 5% to 30%).
  For cortex-a9 and small copies, the speed up is present, but relatively
  small (about 3% to 5%).
  For __memset_chk and large copies, the savings is also small (about 1%).
  However, all processors show larger speed-ups on small copies (about 30% to
  100%).

Bug: 9293744

Change-Id: I8926d59fe2673e36e8a27629e02a7b7059ebbc98
2013-08-06 15:38:29 -07:00
Christopher Ferris
1ce6654163 Fix strcpy.c that should have been strcpy.S.
Change-Id: Ib4609baad3a14c8b0f37556269781fa2b06916dc
2013-08-05 17:08:06 -07:00
Christopher Ferris
e1857431e8 Merge "Optimize strcat/strcpy, small tweaks to strlen." 2013-08-05 23:32:06 +00:00
Ben Cheng
b78f43579f am aa2733d1: Merge "Update the comments to reflect the current status."
* commit 'aa2733d17b87c607fccbd6e6a0f44d2d411ffd77':
  Update the comments to reflect the current status.
2013-08-02 17:58:47 -07:00
Elliott Hughes
d0313e7a2a am a6ed05c1: Merge "libgcc_compat: Introduce __aeabi_lasr for cortex-a9 and higher"
* commit 'a6ed05c1c4c787241b56df132e77512c64cbc595':
  libgcc_compat: Introduce __aeabi_lasr for cortex-a9 and higher
2013-08-02 17:58:46 -07:00
Ben Cheng
772b797b7b Update the comments to reflect the current status.
Change-Id: I3a6348b568230fe8b21d121e5b8d30561a9703c2
2013-08-02 15:53:18 -07:00
Christopher Ferris
d119b7b6f4 Optimize strcat/strcpy, small tweaks to strlen.
Create one version of strcat/strcpy/strlen for cortex-a15/krait and another
version for cortex-a9.

Tested with the libc_test strcat/strcpy/strlen tests.
Including new tests that verify that the src for strcat/strcpy do not
overread across page boundaries.

NOTE: The handling of unaligned strcpy (same code in strcat) could probably
be optimized further such that the src is read 64 bits at a time instead of
the partial reads occurring now.

strlen improves slightly since it was recently optimized.

Performance improvements for strcpy and strcat (using an empty dest string):

cortex-a9
- Small copies vary from about 5% to 20% as the size gets above 10 bytes.
- Copies >= 1024, about a 60% improvement.
- Unaligned copies, from about 40% improvement.

cortex-a15
- Most small copies exhibit a 100% improvement, a few copies only
  improve by 20%.
- Copies >= 1024, about 150% improvement.
- Unaligned copies, about 100% improvement.

krait
- Most small copies vary widely, but on average 20% improvement, then
  the performance gets better, hitting about a 100% improvement when
  copies 64 bytes of data.
- Copies >= 1024, about 100% improvement.
- When coping MBs of data, about 50% improvement.
- Unaligned copies, about 90% improvement.

As strcat destination strings get larger in size:

cortex-a9
- about 40% improvement for small dst strings (>= 32).
- about 250% improvement for dst strings >= 1024.

cortex-a15
- about 200% improvement for small dst strings (>=32).
- about 250% improvement for dst strings >= 1024.

krait
- about 25% improvement for small dst strings (>=32).
- about 100% improvement for dst strings >=1024.

Change-Id: Ifd091ebdbce70fe35a7c5d8f71d5914255f3af35
2013-08-02 10:31:51 -07:00
synergydev
efddf44c8e libgcc_compat: Introduce __aeabi_lasr for cortex-a9 and higher
This is needed when passing -mcpu=cortex-a9 or higher on a modern
toolchain for prebuilt library compatibility

Change-Id: I73eb2393377914ae26216a8c2828ad973d1c1225
2013-07-29 16:55:08 -07:00
Christopher Ferris
7ff868a630 am f63c28f0: Merge "Fix assembler errors in generic arm strlen.c."
* commit 'f63c28f0338fd647e88f1f9300b2220093af1aae':
  Fix assembler errors in generic arm strlen.c.
2013-07-16 17:22:05 -07:00
Christopher Ferris
b1d7fd4969 am 6f4fed74: Merge "Add new optimized strlen for arm."
* commit '6f4fed74cb9405c0f5322307085d15afed6be764':
  Add new optimized strlen for arm.
2013-07-16 17:21:55 -07:00
Christopher Ferris
9ad2a73ed6 Fix assembler errors in generic arm strlen.c.
Tested using a static version of the strlen libc_test program
on a nexus7 that uses the generic code.

Merge from internal master.

(cherry-picked from d8d10a8994)

Change-Id: I88f7dc01dc5b5c3ac2d5580d92153bc1bc36c564
2013-07-16 16:47:54 -07:00
Christopher Ferris
0aa9b52efa Add new optimized strlen for arm.
This optimized version is primarily targeted at cortex-a15.

Tested on all nexus devices using the system/extras/libc_test strlen test.
Tested alignments from 1 to 32 that are powers of 2.
Tested that strlen does not cross page boundaries at all alignments.

Speed improvements listed below:

cortex-a15
- Sizes >= 32 bytes, ~75% improvement.
- Sizes >= 1024 bytes, ~250% improvement.

cortex-a9
- Sizes >= 32 bytes, ~75% improvement.
- Sizes >= 1024 bytes, ~85% improvement.

krait
- Sizes >= 32 bytes, ~95% improvement.
- Sizes >= 1024 bytes, ~160% improvement.

Merge from internal master.

(cherry-picked from 2fc0717977)

Change-Id: I1ceceb4e745fd68e9d946f96d1d42e0cdaff6ccf
2013-07-16 16:47:37 -07:00
Elliott Hughes
62d6b7526a am 2a18ea14: am f152e386: Merge "EABI syscall cleanup."
* commit '2a18ea1462cf65cc51bfcb1a1c46972ee5af1d01':
  EABI syscall cleanup.
2013-07-16 15:36:11 -07:00
Elliott Hughes
2a18ea1462 am f152e386: Merge "EABI syscall cleanup."
* commit 'f152e386fcf477f3f5de9dc020c3660d4f9c4b81':
  EABI syscall cleanup.
2013-07-16 15:31:39 -07:00
Elliott Hughes
da4a3e6515 EABI syscall cleanup.
We cleaned up the auto-generated ones a while back to not touch
the stack unnecessarily if they have <= 4 arguments. This patch
cleans up some hand-crafted ones.

Also improve comments in clone.S.

Change-Id: I8850bf98f2b26829385315304472a760e6880ed8
2013-07-16 11:52:24 -07:00
Christopher Ferris
d8d10a8994 Fix assembler errors in generic arm strlen.c.
Tested using a static version of the strlen libc_test program
on a nexus7 that uses the generic code.

Change-Id: If04d15dcb6c0b18f27f2fefadca5510ed49016c5
2013-07-15 13:56:45 -07:00
Christopher Ferris
2fc0717977 Add new optimized strlen for arm.
This optimized version is primarily targeted at cortex-a15.

Tested on all nexus devices using the system/extras/libc_test strlen test.
Tested alignments from 1 to 32 that are powers of 2.
Tested that strlen does not cross page boundaries at all alignments.

Speed improvements listed below:

cortex-a15
- Sizes >= 32 bytes, ~75% improvement.
- Sizes >= 1024 bytes, ~250% improvement.

cortex-a9
- Sizes >= 32 bytes, ~75% improvement.
- Sizes >= 1024 bytes, ~85% improvement.

krait
- Sizes >= 32 bytes, ~95% improvement.
- Sizes >= 1024 bytes, ~160% improvement.

Change-Id: I361b1a36ed89ab991f2a8f0abbf0d7416d39c8f5
2013-07-15 12:37:51 -07:00
Elliott Hughes
be438a4c40 am fac9199c: am ebc8ce1d: Merge "libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings."
* commit 'fac9199c7698481805dd9b1adaf89a2584719f4c':
  libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
2013-07-03 10:28:19 -07:00
Elliott Hughes
fac9199c76 am ebc8ce1d: Merge "libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings."
* commit 'ebc8ce1de68a83d772106af98c7cb98150bb5662':
  libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
2013-07-03 10:23:41 -07:00
Will Newton
2753e12af5 libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
This memcpy code uses NEON/VFP to achieve very good performance
on ARMv7-A processors. It is specifically tuned for A15 but should
provide good performance on A9 also. It is equivalent to the code
in cortex-strings rev 116.

This patch is a follow up the existing gerrit change:

I7f6f77995f3ca903ad9c66d14261441667a2a935

This version includes a tweak for performance on misaligned
buffers and splits the header comment into license and
documentation sections.

Change-Id: Ibd2e23c8d8e01357ba0247be1d05192de3ceba69
Signed-off-by: Will Newton <will.newton@linaro.org>
2013-07-03 10:20:43 -07:00
Christopher Ferris
c6ac3ae269 am 269daac2: am 7c14d67b: Merge "libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings."
* commit '269daac2f1d76a478b83ba4cbb57d28b47eef5ec':
  libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
2013-07-01 10:39:08 -07:00
Christopher Ferris
269daac2f1 am 7c14d67b: Merge "libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings."
* commit '7c14d67bc1cc2679365a784e68518bf602b81dc7':
  libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
2013-07-01 10:32:17 -07:00
Will Newton
b61103dff4 libc/arch-arm/bionic/memcpy.a9.S: memcpy from cortex-strings.
This memcpy code uses NEON/VFP to achieve very good performance
on ARMv7-A processors. It is specifically tuned for A15 but should
provide good performance on A9 also. It is equivalent to the code
in cortex-strings rev 116.

This patch is a follow up the existing gerrit change:

I7f6f77995f3ca903ad9c66d14261441667a2a935

But this version includes a tweak for performance on misaligned
buffers.

Change-Id: I285abac0068f8ae29a1cbf7862ea8590aadaf0a7
Signed-off-by: Will Newton <will.newton@linaro.org>
2013-07-01 11:15:27 +01:00
Rom Lemarchand
6937468d37 am baa61864: am 995f17e6: Merge "libc: add swapon and swapoff syscalls"
* commit 'baa61864c515a56d4dbeac46b149b4317b01797b':
  libc: add swapon and swapoff syscalls
2013-06-25 17:02:06 -07:00
Rom Lemarchand
baa61864c5 am 995f17e6: Merge "libc: add swapon and swapoff syscalls"
* commit '995f17e6a9a9903f03f542192da9a83b1cabc684':
  libc: add swapon and swapoff syscalls
2013-06-25 15:28:21 -07:00
Rom Lemarchand
d206b560e7 libc: add swapon and swapoff syscalls
Change-Id: Ie79dc8e3f2ff1cd427dd6d95e3850920c4b407b0
Signed-off-by: Rom Lemarchand <romlem@google.com>
2013-06-25 13:18:03 -07:00
Ben Cheng
d20a04c5cf am 77f90de7: am fc104f89: Merge "Fix abort(3) to raise SIGABRT rather than causing SIGSEGV."
* commit '77f90de728b9fa60b83b7f12a45c1113f3189cb2':
  Fix abort(3) to raise SIGABRT rather than causing SIGSEGV.
2013-06-10 17:28:46 -07:00
Ben Cheng
77f90de728 am fc104f89: Merge "Fix abort(3) to raise SIGABRT rather than causing SIGSEGV."
* commit 'fc104f899d47916f76c91127caf9aeaf7b69d4ef':
  Fix abort(3) to raise SIGABRT rather than causing SIGSEGV.
2013-06-10 17:25:31 -07:00
Ben Cheng
7e6ce1a3c5 Fix abort(3) to raise SIGABRT rather than causing SIGSEGV.
tgkill() needs the .save stack unwinding directive to get the complete
stack trace.

BUG: https://code.google.com/p/android/issues/detail?id=16672

Change-Id: Ifb447dca2147a592c48baf32769dfc175d8aea72
2013-06-10 17:17:46 -07:00
Ben Cheng
72ce296f28 am 404d491e: Merge "Use bl instead of blx to support interworking properly."
* commit '404d491eb655839bf4260cc168bb79864473e129':
  Use bl instead of blx to support interworking properly.
2013-06-01 08:19:07 -07:00
Ben Cheng
a123b5d319 Use bl instead of blx to support interworking properly.
(cherry picked from commit 9e1905794b in
master)

Change-Id: I9b8c35ea9e201e00f84315f9f105013c23c94d85
2013-05-31 14:39:23 -07:00
Ben Cheng
9e1905794b Use bl instead of blx to support interworking properly.
BUG: 9227177
Change-Id: I742c2f2ecbe332f9c9743e3f4bde8de791a1d289
2013-05-31 14:25:48 -07:00
Erik Gilling
d5234a3b08 am 4c8eba6f: am 2e317075: Merge "libc/arm: add cortex-a8 cpu variant"
* commit '4c8eba6f2aaf351e29881ca4dc2ec47fc0246446':
  libc/arm: add cortex-a8 cpu variant
2013-05-16 13:20:53 -07:00
Erik Gilling
4c8eba6f2a am 2e317075: Merge "libc/arm: add cortex-a8 cpu variant"
* commit '2e317075b044e94fc75e36d08bec8a7eb5fc31ae':
  libc/arm: add cortex-a8 cpu variant
2013-05-16 13:19:07 -07:00
Rom Lemarchand
22bda4bd67 libc/arm: add cortex-a8 cpu variant
Change-Id: I30e8dd6d4b2e7889aea8f5ed21182a5941bfb489
2013-05-15 20:13:28 -07:00
Elliott Hughes
562804ff87 am f0f4fa3f: Merge "libc: add timerfd calls"
* commit 'f0f4fa3fb1ea8623b1e1bc59f7967e0470c8e532':
  libc: add timerfd calls
2013-05-14 14:59:16 -07:00
Todd Poynor
4200e6203a libc: add timerfd calls
(cherry-pick of 04c0ac14a49e0969333008a9522b64046d58fbdc.)

Change-Id: I06d0b6c2a8781602362b81f48faf1cca76b9ec05
2013-05-14 14:45:02 -07:00
Todd Poynor
04c0ac14a4 libc: add timerfd calls
Change-Id: Id63b907266d5b87c7422a51d393a1430551ca33d
2013-05-13 12:06:15 -07:00
Christopher Ferris
4d8fe5177e Tune the memcpy for krait.
Streamline the memcpy a bit removing some unnecessary instructions.

The biggest speed improvement comes from changing the size of
the preload. On krait, the sweet spot for the preload in the main
loop is twice the L1 cache line size.

In most cases, these small tweaks yield > 1000MB/s speed ups. As
the size of the memcpy approaches about 1MB, the speed improvement
disappears.

Change-Id: Ief79694d65324e2db41bee4707dae19b8c24be62
2013-05-02 14:04:31 -07:00
Andrew Hsieh
83966db80b am f7153fd1: Merge "Remove redundant space within square brackets"
* commit 'f7153fd13f469e9ba5aecbfa00fde42530ca2124':
  Remove redundant space within square brackets
2013-04-25 21:22:26 -07:00
Andrew Hsieh
e8f46e8edd Remove redundant space within square brackets
The new "as" in binutils-2.23 (with gcc4.8) is more picky:
it expects register right after [

Change-Id: I876124841582070ab2083ffafe38bc333b5812d0
2013-04-25 15:05:03 +08:00
Christopher Ferris
39e4ed9699 am 516a8970: Merge "Rewrite memset for cortexa15 to use strd."
* commit '516a89705378f43646678e75924529404e52b613':
  Rewrite memset for cortexa15 to use strd.
2013-04-12 12:30:22 -07:00
Christopher Ferris
796cbe249b Rewrite memset for cortexa15 to use strd.
Merge from internal master.

(cherry-picked from commit 7ffad9c120)

Change-Id: Ia67f2a545399f4fa37b63d5634a3565e4f5482f9
2013-04-12 10:58:25 -07:00
Christopher Ferris
101dadf6a6 am fc76c7d3: Merge "Add missing branch in memcpy.S dst aligned case."
* commit 'fc76c7d394ebe0e585777955efadf7cc8ed86636':
  Add missing branch in memcpy.S dst aligned case.
2013-04-10 17:37:49 -07:00
Christopher Ferris
3fe5b10948 am 68fd78ef: Merge "Update to latest cortexa15 memcpy code."
* commit '68fd78efa05fc61adfbdeadeb757caa45663570c':
  Update to latest cortexa15 memcpy code.
2013-04-10 17:37:49 -07:00
Christopher Ferris
bf0d1ad72b Add missing branch in memcpy.S dst aligned case.
Merge from internal master.

(cherry-picked from commit 6ffaa931c3)

Change-Id: Ifdcf01fd122866cf0d4c5b5f7a997803561d7889
2013-04-10 17:21:29 -07:00
Christopher Ferris
185ce72d00 Update to latest cortexa15 memcpy code.
This uses the new code original submitted as memcpy.a15.S as
the base. However, the old code handled unaligned src/dst better
so that was spliced in. I optimized the original unaligned code by
removing a few unnecessary instructions. I optimized the a15 code by
rewriting the pre and post code. I also modified the main loop to add
a pld so that larger copies would not stall waiting for memory.

Test cases for the new memcpy:

- Copy all sized values from 0 to 1024 bytes, using whatever alignment
  is returned by malloc.
For each alignment case described below, the test copied from 0 to 128
bytes.
- Src and dst pointers are both aligned to the same value, starting
  at one going through every power of two up to and including 128.
- Src aligned to double word boundary, dst aligned to word boundary.
- Src aligned to word boundary, dst aligned to double word boundary.
- Src aligned to 16 bit boundary, dst aligned to word boundary.
- Src aligned to word boundary, dst aligned to 16 byte boundary.
- Src aligned to word boundary, dst aligned to 1 byte from a word
  boundary.
- Src aligned to word boundary, dst aligned to 2 bytes from a word
  boundary.
- Src aligned to word boundary, dst aligned to 3 bytes from a word
  boundary.
- Src aligned to 1 byte from a word boundary, dst aligned to a word
  boundary.
- Src aligned to 2 bytes from a word boundary, dst aligned to a word
  boundary.
- Src aligned to 3 bytes from a word boundary, dst aligned to a word
  boundary.

Cases to verify the unaligned source code properly aligns to a 16 bit
boundary.
- Src aligned to 1 byte from a 128 bit boundary, dst aligned to
  4 + 128 bit boundary.
- Src aligned to 1 byte from a 128 bit boundary, dst aligned to
  8 + 128 bit boundary.
- Src aligned to 1 byte from a 128 bit boundary, dst aligned to
  12 + 128 bit boundary.
- Src aligned to 1 byte from a 128 bit boundary, dst aligned to
  16 + 128 bit boundary.

In all cases, a two byte fencepost was placed at the end of the
destination to verify that only the requested number of bytes were copied.

Bug: 8005082

Merge from internal master.

(cherry-picked from commit 21ede92d79)

Change-Id: Ief70c9e6dc8c6473ae245b6570b2c266fed9618c
2013-04-08 18:13:35 -07:00
Dima Zavin
369f92349f Merge "libc/arm: add cortex-a7 cpu variant" into jb-mr2-dev 2013-03-25 19:42:28 +00:00
Dima Zavin
0c973d7049 libc/arm: add cortex-a7 cpu variant
Change-Id: I541d665805ea69ca96bb6a5f4d50e56287f8c08c
Signed-off-by: Dima Zavin <dima@android.com>
2013-03-23 01:38:22 -07:00