75528 2012-01-04 01:04:07 -0800 Optimize the multiply-add in Biquad.cpp::process 2013-07-09 10:43:54 -0700 1 1 1 Unclassified WebKit Web Audio 528+ (Nightly build) Unspecified Unspecified UNCONFIRMED P2 Normal --- 81168 102524 102544 0 xingnan.wang webkit-unassigned abarth bfulgham crogers danceoffwithyourpantsoff dglazkov rtoy rtoy webkit.review.bot oldest_to_newest 529398 0 xingnan.wang 2012-01-04 01:04:07 -0800 Optimize the multiply-add by piplining with SSE2 instructions. 529405 1 121082 xingnan.wang 2012-01-04 01:17:37 -0800 Created attachment 121082 Patch 529923 2 crogers 2012-01-04 18:04:48 -0800 Thanks for the optimization! Any idea how much improvement this gives? I'm adding Raymond as reviewer here, and think we should get at least a very simple Biquad layout test written to validate both the existing code, and your new code. The layout test can do the following: 1. create a BiquadFilterNode, configure it as lowpass at a certain cutoff frequency 2. feed an impulse into the filter in an OfflineAudioContext 3. compare the result with the expected output using JavaScript to generate the expected response (we can borrow the math from Biquad.cpp and convert to JS. 529925 3 crogers 2012-01-04 18:06:30 -0800 Xingnan do you know how to write a layout test. If not, maybe Raymond can help you there, since he's just starting to write them himself. 529930 4 xingnan.wang 2012-01-04 18:12:11 -0800 (In reply to comment #2) > Thanks for the optimization! Any idea how much improvement this gives? > From my test, it boosted at least 20%(for 128 frames size), and reached 40% with larger size(like 1280). > I'm adding Raymond as reviewer here, and think we should get at least a very simple Biquad layout test written to validate both the existing code, and your new code. The layout test can do the following: > > 1. create a BiquadFilterNode, configure it as lowpass at a certain cutoff frequency > 2. feed an impulse into the filter in an OfflineAudioContext > 3. compare the result with the expected output using JavaScript to generate the expected response (we can borrow the math from Biquad.cpp and convert to JS. 529932 5 xingnan.wang 2012-01-04 18:16:41 -0800 (In reply to comment #3) > Xingnan do you know how to write a layout test. If not, maybe Raymond can help you there, since he's just starting to write them himself. I did not write layout test before but very interest in it, I would like to run the layout test, it`s very pleasure if Raymond can help. Thanks. 530326 6 rtoy 2012-01-05 08:53:24 -0800 (In reply to comment #5) > (In reply to comment #3) > > Xingnan do you know how to write a layout test. If not, maybe Raymond can help you there, since he's just starting to write them himself. > > I did not write layout test before but very interest in it, I would like to run the layout test, it`s very pleasure if Raymond can help. Thanks. I would be my pleasure to write a simple layout test that you can start with. 530422 7 121300 rtoy 2012-01-05 10:45:40 -0800 Created attachment 121300 Biquad layout test script Here is a start at a test of the biquad (lowpass) filter. Place this file in the third_party/WebKit/LayoutTests/webaudio directory. A biquad-test-expected.txt file needs to be created that contains the expected output. To run the layout test, see http://www.chromium.org/developers/testing/webkit-layout-tests. Perhaps it is enough to test just the lowpass filter for now. But eventually, I think we need a test for all the filters. In that case biquad-test.html should be broken up into a common piece that contains the test infrastructure and a probably several tests that create the appropriate filter tests. Should probably follow the style of convolution-mono-mono.html tests. 530446 8 121082 rtoy 2012-01-05 11:07:10 -0800 Comment on attachment 121082 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=121082&action=review > Source/WebCore/platform/audio/Biquad.cpp:90 > +#ifdef __SSE2__ Did you test this on windows? I think the asm syntax below only works for gcc, so this conditionalization should test for gcc too. (But I guess on windows, __SSE2__ is probably never defined.) > Source/WebCore/platform/audio/Biquad.cpp:93 > + __asm__( I am far from an expert in sse2, but this seems rather complex. Could we do something like this? Create array yy of length 2 (or maybe use the destination array directly?): yy[0] = y yy[1] = y1 load xmm0 with (b0 b1 b2 0) load xmm1 with (-a0 -a1 0 0) load xmm2 from *source to get (x[0] x[1] x[2] x[3]) xmm2 = xmm2 * xmm0 to get (b0*x0 b1*x1 b2*x2 0) load xmm3 from yy to get (y0 y1 junk junk) xmm3 = xmm3 * xmm1 to get (-a0*y0 a1*y1 0 0) xmm3 = xmm3 + xmm2 to get (b0*x0-a0*y0, b1*x1-a1*y1, b2*x2, 0) Extract each part of xmm3 and add them together. (We could gain something here if we had SSE3 to do the add, I think.) yy[0] = yy[1] yy[1] = result of sum. Don't know if this is faster or slower. This will change results slightly because we do everything in single precision. 530458 9 rtoy 2012-01-05 11:20:26 -0800 Oops. In the test script, you need to uncomment the first 5 commented lines in runTest() when running the layout test. You can use the script as is when running in a browser (provided the other scripts are available). 530475 10 crogers 2012-01-05 11:34:28 -0800 (In reply to comment #8) > Don't know if this is faster or slower. This will change results slightly because we do everything in single precision. I'd like to keep everything in double-precision because of filter stability problems with certain types of filters. 530831 11 xingnan.wang 2012-01-05 17:31:21 -0800 (In reply to comment #6) > (In reply to comment #5) > > (In reply to comment #3) > > > Xingnan do you know how to write a layout test. If not, maybe Raymond can help you there, since he's just starting to write them himself. > > > > I did not write layout test before but very interest in it, I would like to run the layout test, it`s very pleasure if Raymond can help. Thanks. > > I would be my pleasure to write a simple layout test that you can start with. That`s so nice if there is a sample, thanks Raymond. 530844 12 xingnan.wang 2012-01-05 17:59:50 -0800 (In reply to comment #8) > (From update of attachment 121082 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=121082&action=review > > > Source/WebCore/platform/audio/Biquad.cpp:90 > > +#ifdef __SSE2__ > > Did you test this on windows? I think the asm syntax below only works for gcc, so this conditionalization should test for gcc too. (But I guess on windows, __SSE2__ is probably never defined.) Yes, there is another macro in windows. I just started with linux support only, in case of any potential issues. I can add the windows support and do the test. > > > Source/WebCore/platform/audio/Biquad.cpp:93 > > + __asm__( > > I am far from an expert in sse2, but this seems rather complex. Could we do something like this? > > Create array yy of length 2 (or maybe use the destination array directly?): > yy[0] = y > yy[1] = y1 > > load xmm0 with (b0 b1 b2 0) > load xmm1 with (-a0 -a1 0 0) > load xmm2 from *source to get (x[0] x[1] x[2] x[3]) > xmm2 = xmm2 * xmm0 to get (b0*x0 b1*x1 b2*x2 0) > load xmm3 from yy to get (y0 y1 junk junk) > xmm3 = xmm3 * xmm1 to get (-a0*y0 a1*y1 0 0) > xmm3 = xmm3 + xmm2 to get (b0*x0-a0*y0, b1*x1-a1*y1, b2*x2, 0) > > Extract each part of xmm3 and add them together. (We could gain something here if we had SSE3 to do the add, I think.) > > yy[0] = yy[1] > yy[1] = result of sum. > > Don't know if this is faster or slower. This will change results slightly because we do everything in single precision. Your solution may work in single precision and I didn`t try. I have tried this way in double precision but no too much improvement. Then I changed to use inline asm so that I can pipeline the instructions, also get the benefit from SSE2. 533974 13 121082 rtoy 2012-01-11 09:54:53 -0800 Comment on attachment 121082 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=121082&action=review >>> Source/WebCore/platform/audio/Biquad.cpp:93 >>> + __asm__( >> >> I am far from an expert in sse2, but this seems rather complex. Could we do something like this? >> >> Create array yy of length 2 (or maybe use the destination array directly?): >> yy[0] = y >> yy[1] = y1 >> >> load xmm0 with (b0 b1 b2 0) >> load xmm1 with (-a0 -a1 0 0) >> load xmm2 from *source to get (x[0] x[1] x[2] x[3]) >> xmm2 = xmm2 * xmm0 to get (b0*x0 b1*x1 b2*x2 0) >> load xmm3 from yy to get (y0 y1 junk junk) >> xmm3 = xmm3 * xmm1 to get (-a0*y0 a1*y1 0 0) >> xmm3 = xmm3 + xmm2 to get (b0*x0-a0*y0, b1*x1-a1*y1, b2*x2, 0) >> >> Extract each part of xmm3 and add them together. (We could gain something here if we had SSE3 to do the add, I think.) >> >> yy[0] = yy[1] >> yy[1] = result of sum. >> >> Don't know if this is faster or slower. This will change results slightly because we do everything in single precision. > > Your solution may work in single precision and I didn`t try. > I have tried this way in double precision but no too much improvement. Then I changed to use inline asm so that I can pipeline the instructions, also get the benefit from SSE2. This looks fine then especially considering Chris's comment about keeping things in double for stability reasons. 534983 14 rtoy 2012-01-12 11:10:25 -0800 If you are not in a hurry, bug 71413 will include a bunch of tests for the biquad filters so you do not have to write any. 535504 15 xingnan.wang 2012-01-13 00:51:05 -0800 (In reply to comment #14) > If you are not in a hurry, bug 71413 will include a bunch of tests for the biquad filters so you do not have to write any. That will be nice. I was going to update the patch with the low pass filter layout test, but if bug 71413 provides all the tests, I will wait for it. 553885 16 126465 xingnan.wang 2012-02-09 23:44:37 -0800 Created attachment 126465 Patch Update the patch as layout tests of biquad filter landed, it`s OK to pass all these tests. 553941 17 126465 webkit.review.bot 2012-02-10 01:20:22 -0800 Comment on attachment 126465 Patch Attachment 126465 did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/11486599 New failing tests: webaudio/biquad-highpass.html webaudio/biquad-lowpass.html webaudio/biquad-lowshelf.html webaudio/biquad-notch.html webaudio/biquad-highshelf.html webaudio/biquad-peaking.html webaudio/biquad-bandpass.html webaudio/biquad-allpass.html 554366 18 126465 rtoy 2012-02-10 13:56:13 -0800 Comment on attachment 126465 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=126465&action=review There also appears to be test failures on linux. Can you investigate why that is happening? > Source/WebCore/platform/audio/Biquad.cpp:121 > + "addsd %%xmm4, %%xmm5\n\t" // x*b0 + (x2*b2+x*b0) Doesn't xmm4 contain a1*y1 here so the result is a1*y1 + (x2*b2+x*b0)? > Source/WebCore/platform/audio/Biquad.cpp:137 > + :"eax", "edx", "ecx" Is it necessary to add all of the xmm registers to this clobber list? 555045 19 126465 xingnan.wang 2012-02-12 23:26:19 -0800 Comment on attachment 126465 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=126465&action=review >> Source/WebCore/platform/audio/Biquad.cpp:121 >> + "addsd %%xmm4, %%xmm5\n\t" // x*b0 + (x2*b2+x*b0) > > Doesn't xmm4 contain a1*y1 here so the result is a1*y1 + (x2*b2+x*b0)? Sorry for that, it should be a1*y1 + (x2*b2+x*b0). >> Source/WebCore/platform/audio/Biquad.cpp:137 >> + :"eax", "edx", "ecx" > > Is it necessary to add all of the xmm registers to this clobber list? OK, it`s better to add all the xmm registers to the clobber-list, through there is little possibility of conflict of xmm registers here. 555053 20 xingnan.wang 2012-02-12 23:32:34 -0800 (In reply to comment #18) > (From update of attachment 126465 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=126465&action=review > > There also appears to be test failures on linux. Can you investigate why that is happening? > It`s strange that I can pass the layout tests locally, and little message can be found from the error log http://queues.webkit.org/results/11486599. I`ll try to check my patch in different machine. 555065 21 126729 xingnan.wang 2012-02-13 00:03:59 -0800 Created attachment 126729 Patch 555388 22 rtoy 2012-02-13 11:04:32 -0800 (In reply to comment #21) > Created an attachment (id=126729) [details] > Patch Looks good, except for the test failures that happened in the previous patch. I didn't see any test failures from the bots for this patch, though. Perhaps they've been fixed? 555793 23 xingnan.wang 2012-02-13 17:23:21 -0800 (In reply to comment #22) > (In reply to comment #21) > > Created an attachment (id=126729) [details] [details] > > Patch > > Looks good, except for the test failures that happened in the previous patch. I didn't see any test failures from the bots for this patch, though. Perhaps they've been fixed? Actually I never reproduced the failures at all, just updated the patch as your comments. Perhaps the last patch would pass the tests if run it again. Whatever, does that mean the patch is OK now? BTW, how can I check the building details including these failures from the bots? Thanks. 556350 24 rtoy 2012-02-14 09:39:06 -0800 (In reply to comment #23) > (In reply to comment #22) > > (In reply to comment #21) > > > Created an attachment (id=126729) [details] [details] [details] > > > Patch > > > > Looks good, except for the test failures that happened in the previous patch. I didn't see any test failures from the bots for this patch, though. Perhaps they've been fixed? > > Actually I never reproduced the failures at all, just updated the patch as your comments. Perhaps the last patch would pass the tests if run it again. Whatever, does that mean the patch is OK now? Since there are no messages from the webkit review bot with your latest change, I think it's ok now. Not sure what changes made this work. 560824 25 xingnan.wang 2012-02-21 05:46:42 -0800 (In reply to comment #24) > (In reply to comment #23) > > (In reply to comment #22) > > > (In reply to comment #21) > > > > Created an attachment (id=126729) [details] [details] [details] [details] > > > > Patch > > > > > > Looks good, except for the test failures that happened in the previous patch. I didn't see any test failures from the bots for this patch, though. Perhaps they've been fixed? > > > > Actually I never reproduced the failures at all, just updated the patch as your comments. Perhaps the last patch would pass the tests if run it again. Whatever, does that mean the patch is OK now? > > Since there are no messages from the webkit review bot with your latest change, I think it's ok now. Not sure what changes made this work. Since the patch is OK, can it be submitted now? 562186 26 126729 rtoy 2012-02-22 11:35:37 -0800 Comment on attachment 126729 Patch LGTM 566351 27 xingnan.wang 2012-02-27 20:58:20 -0800 (In reply to comment #26) > (From update of attachment 126729 [details]) > LGTM Chris, does the patch need more review? 568286 28 xingnan.wang 2012-02-29 20:32:41 -0800 (In reply to comment #27) > (In reply to comment #26) > > (From update of attachment 126729 [details] [details]) > > LGTM > > Chris, does the patch need more review? Whether the patch is OK to commit? 578744 29 126729 webkit.review.bot 2012-03-14 14:01:51 -0700 Comment on attachment 126729 Patch Clearing flags on attachment: 126729 Committed r110744: <http://trac.webkit.org/changeset/110744> 578745 30 webkit.review.bot 2012-03-14 14:01:57 -0700 All reviewed patches have been landed. Closing bug. 578903 31 abarth 2012-03-14 16:25:32 -0700 This patch seems to crash in debug: [49220:49236:17280644243061:ERROR:process_util_posix.cc(142)] Received signal 11 base::debug::StackTrace::StackTrace() [0x851b56] base::(anonymous namespace)::StackDumpSignalHandler() [0x80b739] 0x7f352f9d1af0 WebCore::Biquad::process() [0x256d73e] WebCore::HRTFKernel::HRTFKernel() [0x25ac2e4] WebCore::HRTFKernel::create() [0x25ab549] WebCore::HRTFElevation::calculateKernelsForAzimuthElevation() [0x25aa703] WebCore::HRTFElevation::createForSubject() [0x25aa9db] WebCore::HRTFDatabase::HRTFDatabase() [0x25a987e] WebCore::HRTFDatabase::create() [0x25a9630] WebCore::HRTFDatabaseLoader::load() [0x257382d] WebCore::databaseLoaderEntry() [0x25737ac] WTF::threadEntryPoint() [0x1f509e5] WTF::wtfThreadEntryPoint() [0x793cb8] start_thread [0x7f3531e369ca] 0x7f352fa8470d None 578919 32 abarth 2012-03-14 16:34:51 -0700 Rolled out in http://trac.webkit.org/changeset/110782 579181 33 xingnan.wang 2012-03-15 00:33:53 -0700 (In reply to comment #31) > This patch seems to crash in debug: > > [49220:49236:17280644243061:ERROR:process_util_posix.cc(142)] Received signal 11 > base::debug::StackTrace::StackTrace() [0x851b56] > base::(anonymous namespace)::StackDumpSignalHandler() [0x80b739] > 0x7f352f9d1af0 > WebCore::Biquad::process() [0x256d73e] > WebCore::HRTFKernel::HRTFKernel() [0x25ac2e4] > WebCore::HRTFKernel::create() [0x25ab549] > WebCore::HRTFElevation::calculateKernelsForAzimuthElevation() [0x25aa703] > WebCore::HRTFElevation::createForSubject() [0x25aa9db] > WebCore::HRTFDatabase::HRTFDatabase() [0x25a987e] > WebCore::HRTFDatabase::create() [0x25a9630] > WebCore::HRTFDatabaseLoader::load() [0x257382d] > WebCore::databaseLoaderEntry() [0x25737ac] > WTF::threadEntryPoint() [0x1f509e5] > WTF::wtfThreadEntryPoint() [0x793cb8] > start_thread [0x7f3531e369ca] > 0x7f352fa8470d > None Hi, That`s strange that I cannot reproduce the issue (by several different machines). Could you provide more details about your platform? 579188 34 abarth 2012-03-15 01:03:40 -0700 It was the chromium-mac-linux configuration, which is Ubuntu Lucid. 579411 35 xingnan.wang 2012-03-15 08:38:09 -0700 (In reply to comment #34) > It was the chromium-mac-linux configuration, which is Ubuntu Lucid. I am confused by chromium-mac-linux, should not it be either chromium-linux or chromium-mac? The patch only acts in linux, I`ve tested in Ubuntu oneiric and it works very well. 579439 36 rtoy 2012-03-15 09:26:08 -0700 (In reply to comment #33) > (In reply to comment #31) > > This patch seems to crash in debug: > > > > [49220:49236:17280644243061:ERROR:process_util_posix.cc(142)] Received signal 11 > > base::debug::StackTrace::StackTrace() [0x851b56] > > base::(anonymous namespace)::StackDumpSignalHandler() [0x80b739] > > 0x7f352f9d1af0 > > WebCore::Biquad::process() [0x256d73e] > > WebCore::HRTFKernel::HRTFKernel() [0x25ac2e4] > > WebCore::HRTFKernel::create() [0x25ab549] > > WebCore::HRTFElevation::calculateKernelsForAzimuthElevation() [0x25aa703] > > WebCore::HRTFElevation::createForSubject() [0x25aa9db] > > WebCore::HRTFDatabase::HRTFDatabase() [0x25a987e] > > WebCore::HRTFDatabase::create() [0x25a9630] > > WebCore::HRTFDatabaseLoader::load() [0x257382d] > > WebCore::databaseLoaderEntry() [0x25737ac] > > WTF::threadEntryPoint() [0x1f509e5] > > WTF::wtfThreadEntryPoint() [0x793cb8] > > start_thread [0x7f3531e369ca] > > 0x7f352fa8470d > > None > > Hi, > That`s strange that I cannot reproduce the issue (by several different machines). Could you provide more details about your platform? The stack trace looks totally bogus. (I'm pretty sure the HRTFKernel doesn't call Biquad). But I can reproduce the crash on my system. The stack trace is totally trash too. More work needed.... 579499 37 rtoy 2012-03-15 10:39:35 -0700 (In reply to comment #36) > (In reply to comment #33) > > (In reply to comment #31) > > > This patch seems to crash in debug: > > > > > > [49220:49236:17280644243061:ERROR:process_util_posix.cc(142)] Received signal 11 > > > base::debug::StackTrace::StackTrace() [0x851b56] > > > base::(anonymous namespace)::StackDumpSignalHandler() [0x80b739] > > > 0x7f352f9d1af0 > > > WebCore::Biquad::process() [0x256d73e] > > > WebCore::HRTFKernel::HRTFKernel() [0x25ac2e4] > > > WebCore::HRTFKernel::create() [0x25ab549] > > > WebCore::HRTFElevation::calculateKernelsForAzimuthElevation() [0x25aa703] > > > WebCore::HRTFElevation::createForSubject() [0x25aa9db] > > > WebCore::HRTFDatabase::HRTFDatabase() [0x25a987e] > > > WebCore::HRTFDatabase::create() [0x25a9630] > > > WebCore::HRTFDatabaseLoader::load() [0x257382d] > > > WebCore::databaseLoaderEntry() [0x25737ac] > > > WTF::threadEntryPoint() [0x1f509e5] > > > WTF::wtfThreadEntryPoint() [0x793cb8] > > > start_thread [0x7f3531e369ca] > > > 0x7f352fa8470d > > > None > > > > Hi, > > That`s strange that I cannot reproduce the issue (by several different machines). Could you provide more details about your platform? > > The stack trace looks totally bogus. (I'm pretty sure the HRTFKernel doesn't call Biquad). > > But I can reproduce the crash on my system. The stack trace is totally trash too. More work needed.... Ok. The issue is that on my debug build, I get a 64-bit build so loading the source address into edx loses the top 32 bits, which are non-zero. We then try to load from that non-existent address and crash. Don't understand why this doesn't happen on a release build, but this is definitely a bug. 579601 38 132095 rtoy 2012-03-15 11:51:01 -0700 Created attachment 132095 Updated patch to handle 32 and 64-bit builds Here is an updated patch. This passes all of the webaudio tests in debug mode. (Release mode test in progress.) This basically replaces all uses of the edx and ecx registers with their 64-bit counterparts, rdx and rcx. However, the complexity of maintaining this assembly code is getting rather large, so I'm inclined not to do this. It would be much better if intrinsics could be used, or, even better, do what the FIXME comment says and unroll the while loop with carefully scheduled arithmetic (in C) if that can achieve the desired performance gains. 579682 39 abarth 2012-03-15 13:33:56 -0700 > I am confused by chromium-mac-linux, should not it be either chromium-linux or chromium-mac? The patch only acts in linux, I`ve tested in Ubuntu oneiric and it works very well. Sorry. I meant chromium-linux. :) 580037 40 xingnan.wang 2012-03-15 19:17:59 -0700 (In reply to comment #38) > Created an attachment (id=132095) [details] > Updated patch to handle 32 and 64-bit builds > > Here is an updated patch. This passes all of the webaudio tests in debug mode. (Release mode test in progress.) > > This basically replaces all uses of the edx and ecx registers with their 64-bit counterparts, rdx and rcx. > > However, the complexity of maintaining this assembly code is getting rather large, so I'm inclined not to do this. It would be much better if intrinsics could be used, or, even better, do what the FIXME comment says and unroll the while loop with carefully scheduled arithmetic (in C) if that can achieve the desired performance gains. Hi Ray, thank you very much for fixing this bug. I agree that we should not make the code here to get too large. I have tried some ways to use intrinsics and C code to unroll the loop, but they could not achieve the performance increased by assembly code, because it was not easy to pipeline and reschedule the executing order of the code as assembly did. But I can make a try to find whether there is a better way to optimize the code with intrinsics. 581588 41 132532 xingnan.wang 2012-03-18 20:18:12 -0700 Created attachment 132532 Patch 581592 42 xingnan.wang 2012-03-18 20:23:38 -0700 (In reply to comment #41) > Created an attachment (id=132532) [details] > Patch Re-implemented the pipeline by intrinsics. What surprised me is the performance is much better than assembly code, ~45% with original ~20%(for 128 frames size). 582029 43 132532 rtoy 2012-03-19 12:13:27 -0700 Comment on attachment 132532 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=132532&action=review This looks much nicer than the assembly code, and it's faster too! Code looks good, but adding some comments would help to understand what's going on. > Source/WebCore/platform/audio/Biquad.cpp:113 > + __m128d mma1 = _mm_set_sd(-a1); mma1 = (0, -a1) > Source/WebCore/platform/audio/Biquad.cpp:117 > + mm6 = mm1; Add comment mm6 = (y2, x) > Source/WebCore/platform/audio/Biquad.cpp:118 > + mm1 = _mm_shuffle_pd(mm1, mm4, 0); // y2 = y1 Add comment that mm1 = (y1, x) > Source/WebCore/platform/audio/Biquad.cpp:119 > + mm5 = _mm_mul_pd(mm2, mm0); // (x1*b1 x2*b2) Insert comma: (x1 * b1, x2 * b2) Webkit style includes a space around arithmetic operators. Do the same for next line. > Source/WebCore/platform/audio/Biquad.cpp:121 > + mm0 = _mm_shuffle_pd(mm0, mm1, 1); // (x1 = x, x2 = x1) Comment is confusing. Maybe say mm0 = (x, x1) > Source/WebCore/platform/audio/Biquad.cpp:122 > + mm4 = _mm_mul_sd(mm4, mma1); // -y1*a1 Be more explicit about contents of mm4. mm4 = (0, -y1 * a1) > Source/WebCore/platform/audio/Biquad.cpp:123 > + mm5 = _mm_add_pd(mm5, mm6); // (x1*b1-y2*a2 x2*b2+x*b0) Insert comma, add space around operators: (x1 * b1 - y2 * a2, x2 * b2 + x * b0) > Source/WebCore/platform/audio/Biquad.cpp:124 > + n--; Why decrement n here? Is this a pipeline issue? If not, can we put the decrement back into while (n--)? > Source/WebCore/platform/audio/Biquad.cpp:125 > + mm6 = mm5; mm6 = (x1 * b1 - y2 * a2, x2 * b2 + x * b0) > Source/WebCore/platform/audio/Biquad.cpp:126 > + mm7 = _mm_load_ss(sourceP); mm7 = (0, 0, 0, *sourceP) And we're loading the next x value. > Source/WebCore/platform/audio/Biquad.cpp:127 > + mm1 = _mm_cvtss_sd(mm1, mm7); // x = *sourceP Maybe say exactly what is in mm1? mm1 = (y1, x) > Source/WebCore/platform/audio/Biquad.cpp:128 > + mm5 = _mm_add_pd(mm4, mm5); // (x1*b1-y2*a2 x2*b2+x*b0-y1*a1) Same comment as for line 123, 119. Any advantage in doing mm_add_pd? It looks like we don't care about the high part of mm5. Is that right? > Source/WebCore/platform/audio/Biquad.cpp:129 > + mm6 = _mm_shuffle_pd(mm6, mm6, 1); mm6 = (x2 * b2 + x * b0, x1 * b1 - y2 * a2) > Source/WebCore/platform/audio/Biquad.cpp:130 > + mm5 = _mm_add_pd(mm5, mm6); // y mm5 = (x1*b1-y2*a2+x2*b2+x*b0, y = x2*b2+x*b0-y1*a1+x1*b1-y2*a2), and we ignore high part of mm5, so maybe just mm_add_sd? > Source/WebCore/platform/audio/Biquad.cpp:131 > + mm7 = _mm_cvtsd_ss(mm7, mm5); mm7 = (0, 0, 0, y) > Source/WebCore/platform/audio/Biquad.cpp:133 > + mm4 = mm5; // y1 = y mm4 = (x1*b1-y2*a2+x2*b2+x*b0, y) 582767 44 132760 xingnan.wang 2012-03-19 22:34:45 -0700 Created attachment 132760 Patch 582774 45 132532 xingnan.wang 2012-03-19 22:44:34 -0700 Comment on attachment 132532 Patch Hi Ray, patch is updated as your comments. View in context: https://bugs.webkit.org/attachment.cgi?id=132532&action=review >> Source/WebCore/platform/audio/Biquad.cpp:113 >> + __m128d mma1 = _mm_set_sd(-a1); > > mma1 = (0, -a1) Done. >> Source/WebCore/platform/audio/Biquad.cpp:117 >> + mm6 = mm1; > > Add comment mm6 = (y2, x) Done. >> Source/WebCore/platform/audio/Biquad.cpp:118 >> + mm1 = _mm_shuffle_pd(mm1, mm4, 0); // y2 = y1 > > Add comment that mm1 = (y1, x) Done. >> Source/WebCore/platform/audio/Biquad.cpp:119 >> + mm5 = _mm_mul_pd(mm2, mm0); // (x1*b1 x2*b2) > > Insert comma: (x1 * b1, x2 * b2) > Webkit style includes a space around arithmetic operators. Do the same for next line. Done. >> Source/WebCore/platform/audio/Biquad.cpp:121 >> + mm0 = _mm_shuffle_pd(mm0, mm1, 1); // (x1 = x, x2 = x1) > > Comment is confusing. Maybe say mm0 = (x, x1) Done. >> Source/WebCore/platform/audio/Biquad.cpp:122 >> + mm4 = _mm_mul_sd(mm4, mma1); // -y1*a1 > > Be more explicit about contents of mm4. > > mm4 = (0, -y1 * a1) Done. >> Source/WebCore/platform/audio/Biquad.cpp:123 >> + mm5 = _mm_add_pd(mm5, mm6); // (x1*b1-y2*a2 x2*b2+x*b0) > > Insert comma, add space around operators: > > (x1 * b1 - y2 * a2, x2 * b2 + x * b0) Done. >> Source/WebCore/platform/audio/Biquad.cpp:124 >> + n--; > > Why decrement n here? Is this a pipeline issue? If not, can we put the decrement back into while (n--)? This will impact the pipeline, using "while(n--)" pulls down the performance a little, 45% to 43%. >> Source/WebCore/platform/audio/Biquad.cpp:125 >> + mm6 = mm5; > > mm6 = (x1 * b1 - y2 * a2, x2 * b2 + x * b0) Done. >> Source/WebCore/platform/audio/Biquad.cpp:126 >> + mm7 = _mm_load_ss(sourceP); > > mm7 = (0, 0, 0, *sourceP) > > And we're loading the next x value. Done. >> Source/WebCore/platform/audio/Biquad.cpp:128 >> + mm5 = _mm_add_pd(mm4, mm5); // (x1*b1-y2*a2 x2*b2+x*b0-y1*a1) > > Same comment as for line 123, 119. > > Any advantage in doing mm_add_pd? It looks like we don't care about the high part of mm5. Is that right? Yes, _mm_add_sd is better. >> Source/WebCore/platform/audio/Biquad.cpp:129 >> + mm6 = _mm_shuffle_pd(mm6, mm6, 1); > > mm6 = (x2 * b2 + x * b0, x1 * b1 - y2 * a2) Done. >> Source/WebCore/platform/audio/Biquad.cpp:130 >> + mm5 = _mm_add_pd(mm5, mm6); // y > > mm5 = (x1*b1-y2*a2+x2*b2+x*b0, y = x2*b2+x*b0-y1*a1+x1*b1-y2*a2), and we ignore high part of mm5, so maybe just mm_add_sd? Done. >> Source/WebCore/platform/audio/Biquad.cpp:133 >> + mm4 = mm5; // y1 = y > > mm4 = (x1*b1-y2*a2+x2*b2+x*b0, y) Done. 583169 46 132760 rtoy 2012-03-20 10:00:27 -0700 Comment on attachment 132760 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=132760&action=review Code looks good; just a few more minor items to look at. > Source/WebCore/platform/audio/Biquad.cpp:109 > + __m128d mm4 = _mm_set_pd(0, y1); // mm4 = (0, y1) It looks like we only care about the the low part of mm4. Maybe add comment that we only use the low part of mm4. Same goes for mma1 in line 113. > Source/WebCore/platform/audio/Biquad.cpp:122 > + mm4 = _mm_mul_sd(mm4, mma1); // mm4 = (0, -y1 * a1) I think the comment should be mm4=-y1*a1 since we're doing mul_sd. We don't care what's in the high part of mm4. > Source/WebCore/platform/audio/Biquad.cpp:125 > + mm6 = mm5; // mm6 = (x1 * b1 - y2 * a2, x2 * b2 + x * b0) See note on for line 130. Shuffle mm5 to the right parts of mm6 now instead of later? > Source/WebCore/platform/audio/Biquad.cpp:126 > + mm7 = _mm_load_ss(sourceP); // Load next x value, mm7 = (0, 0, 0, x) We load mm7 from memory here, but use it in the very next instruction. Could this be moved further up to lessen any pipeline issues? We're not using the value of mm7 until line 127. > Source/WebCore/platform/audio/Biquad.cpp:128 > + mm5 = _mm_add_sd(mm4, mm5); // mm5 = (x1 * b1 - y2 * a2, x2 * b2 + x * b0 - y1 * a1) Since we're doing mm_add_sd, we don't care what's in the high part of mm5. Maybe the comment should be mm5 = x2*b2 + x*b0-y1*a1. > Source/WebCore/platform/audio/Biquad.cpp:130 > + mm5 = _mm_add_sd(mm5, mm6); // mm5 = (x1 * b1 - y2 * a2, x * b0 + x1 * b1 + x2 * b2 - y1 * a1 - y2 * a2) = (*, y) Same comment here as for line 128. We don't care what's in the high part of mm5. Also, in line 129, we swap the high and low parts of mm6. Isn't it possible to do that in line 125 where we assign mm5 to mm6? Just shuffle mm5 to the right parts of mm6? That's one less instruction, but I don't know what the pipeline affects will be. > Source/WebCore/platform/audio/Biquad.cpp:131 > + mm7 = _mm_cvtsd_ss(mm7, mm5); // mm7 = (0, 0, 0, y) Actually, I think the comment is wrong. (Sorry about that.) cvtsd_ss doesn't modify the other parts of mm7. Plus we're only using the low part, maybe just say mm7 = static_cast<float>(y). 583729 47 132954 xingnan.wang 2012-03-20 20:12:57 -0700 Created attachment 132954 Patch 583744 48 132760 xingnan.wang 2012-03-20 20:36:18 -0700 Comment on attachment 132760 Patch Thanks, Ray. Update the patch. View in context: https://bugs.webkit.org/attachment.cgi?id=132760&action=review >> Source/WebCore/platform/audio/Biquad.cpp:126 >> + mm7 = _mm_load_ss(sourceP); // Load next x value, mm7 = (0, 0, 0, x) > > We load mm7 from memory here, but use it in the very next instruction. Could this be moved further up to lessen any pipeline issues? We're not using the value of mm7 until line 127. There is little difference of performance after moving this line up, I think it is not in the key path of the pipeline. >> Source/WebCore/platform/audio/Biquad.cpp:130 >> + mm5 = _mm_add_sd(mm5, mm6); // mm5 = (x1 * b1 - y2 * a2, x * b0 + x1 * b1 + x2 * b2 - y1 * a1 - y2 * a2) = (*, y) > > Same comment here as for line 128. We don't care what's in the high part of mm5. > > Also, in line 129, we swap the high and low parts of mm6. Isn't it possible to do that in line 125 where we assign mm5 to mm6? Just shuffle mm5 to the right parts of mm6? That's one less instruction, but I don't know what the pipeline affects will be. It seems we should do that and make us to save one instruction, but it pulls down the performance after that, strange~ I guess the compiler does much for us when we use intrinsic, the actual assembly code looks much different from what I expect for after "gcc -s " . 584205 49 rtoy 2012-03-21 09:17:07 -0700 (In reply to comment #48) > (From update of attachment 132760 [details]) > Thanks, Ray. Update the patch. > > View in context: https://bugs.webkit.org/attachment.cgi?id=132760&action=review > > >> Source/WebCore/platform/audio/Biquad.cpp:126 > >> + mm7 = _mm_load_ss(sourceP); // Load next x value, mm7 = (0, 0, 0, x) > > > > We load mm7 from memory here, but use it in the very next instruction. Could this be moved further up to lessen any pipeline issues? We're not using the value of mm7 until line 127. > > There is little difference of performance after moving this line up, I think it is not in the key path of the pipeline. > > >> Source/WebCore/platform/audio/Biquad.cpp:130 > >> + mm5 = _mm_add_sd(mm5, mm6); // mm5 = (x1 * b1 - y2 * a2, x * b0 + x1 * b1 + x2 * b2 - y1 * a1 - y2 * a2) = (*, y) > > > > Same comment here as for line 128. We don't care what's in the high part of mm5. > > > > Also, in line 129, we swap the high and low parts of mm6. Isn't it possible to do that in line 125 where we assign mm5 to mm6? Just shuffle mm5 to the right parts of mm6? That's one less instruction, but I don't know what the pipeline affects will be. > > It seems we should do that and make us to save one instruction, but it pulls down the performance after that, strange~ > I guess the compiler does much for us when we use intrinsic, the actual assembly code looks much different from what I expect for after "gcc -s " . Ok. I'm happy with this patch as it is now. Looks good. I do wonder, though, if you moved load_ss to be after line 124 and changed line 125 to do the shuffle (removing the shuffle at line 129) would improve performance. Or maybe move the (new) shuffle down a few lines since we don't need mm6 until line 130. Perhaps these changes will help the pipeline. You don't have to do this experiment if you don't want to. I'm just kind of curious. Otherwise, the patch looks good to me and is ready to go. 584506 50 crogers 2012-03-21 12:38:13 -0700 Does this latest patch pass the layout test (and not crash) on the Debug build? 585084 51 132760 xingnan.wang 2012-03-22 02:41:26 -0700 Comment on attachment 132760 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=132760&action=review >>>> Source/WebCore/platform/audio/Biquad.cpp:126 >>>> + mm7 = _mm_load_ss(sourceP); // Load next x value, mm7 = (0, 0, 0, x) >>> >>> We load mm7 from memory here, but use it in the very next instruction. Could this be moved further up to lessen any pipeline issues? We're not using the value of mm7 until line 127. >> >> There is little difference of performance after moving this line up, I think it is not in the key path of the pipeline. > > Ok. I'm happy with this patch as it is now. Looks good. > > I do wonder, though, if you moved load_ss to be after line 124 and changed line 125 to do the shuffle (removing the shuffle at line 129) would improve performance. Or maybe move the (new) shuffle down a few lines since we don't need mm6 until line 130. Perhaps these changes will help the pipeline. > > You don't have to do this experiment if you don't want to. I'm just kind of curious. Otherwise, the patch looks good to me and is ready to go. It does not matter, I did it. But it seems it does not work, a tiny slowing down. Hard to explain, we cannot exactly schedule the pipeline by intrinsics as the way by assembly, I guess. 585089 52 xingnan.wang 2012-03-22 02:52:28 -0700 (In reply to comment #50) > Does this latest patch pass the layout test (and not crash) on the Debug build? The layout test passed in both debug and release build from my environment (32-bit). Last crash issue`s root cause is I used 32-bit registers for the address and it will failed in 64-bit system. I use intrinsics this time and suppose it would not meet this issue. Certainly more tests are needed, but I don`t have 64-bit system right now. Ray, would you mind to help me to check whether the patch is OK in your 64-bit system? 585444 53 rtoy 2012-03-22 10:28:13 -0700 (In reply to comment #52) > (In reply to comment #50) > > Does this latest patch pass the layout test (and not crash) on the Debug build? > > The layout test passed in both debug and release build from my environment (32-bit). > Last crash issue`s root cause is I used 32-bit registers for the address and it will failed in 64-bit system. I use intrinsics this time and suppose it would not meet this issue. Certainly more tests are needed, but I don`t have 64-bit system right now. I don't expect any issues with 64-bit builds. The compiler should use the right sized registers for the intrinsic operations. > > Ray, would you mind to help me to check whether the patch is OK in your 64-bit system? No problem. The debug build passes all tests, except for the convolution-mono-mono test, but I think that's a bug in the convolver that needs to be fixed. The release build passes all tests too. (convolution-mono-mono passes.) 585448 54 rtoy 2012-03-22 10:30:29 -0700 (In reply to comment #51) > (From update of attachment 132760 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=132760&action=review > > >>>> Source/WebCore/platform/audio/Biquad.cpp:126 > >>>> + mm7 = _mm_load_ss(sourceP); // Load next x value, mm7 = (0, 0, 0, x) > >>> > >>> We load mm7 from memory here, but use it in the very next instruction. Could this be moved further up to lessen any pipeline issues? We're not using the value of mm7 until line 127. > >> > >> There is little difference of performance after moving this line up, I think it is not in the key path of the pipeline. > > > > Ok. I'm happy with this patch as it is now. Looks good. > > > > I do wonder, though, if you moved load_ss to be after line 124 and changed line 125 to do the shuffle (removing the shuffle at line 129) would improve performance. Or maybe move the (new) shuffle down a few lines since we don't need mm6 until line 130. Perhaps these changes will help the pipeline. > > > > You don't have to do this experiment if you don't want to. I'm just kind of curious. Otherwise, the patch looks good to me and is ready to go. > > It does not matter, I did it. But it seems it does not work, a tiny slowing down. Hard to explain, we cannot exactly schedule the pipeline by intrinsics as the way by assembly, I guess. Thanks for doing the test. I would have thought there might be some small gain, but the compiler can rearrange things as you say. So I think this new patch is ready to go as is. Thanks for doing the new version! 586029 55 xingnan.wang 2012-03-22 19:58:56 -0700 (In reply to comment #53) > (In reply to comment #52) > > (In reply to comment #50) > > > Does this latest patch pass the layout test (and not crash) on the Debug build? > > > > The layout test passed in both debug and release build from my environment (32-bit). > > Last crash issue`s root cause is I used 32-bit registers for the address and it will failed in 64-bit system. I use intrinsics this time and suppose it would not meet this issue. Certainly more tests are needed, but I don`t have 64-bit system right now. > > I don't expect any issues with 64-bit builds. The compiler should use the right sized registers for the intrinsic operations. > > > > > Ray, would you mind to help me to check whether the patch is OK in your 64-bit system? > > No problem. The debug build passes all tests, except for the convolution-mono-mono test, but I think that's a bug in the convolver that needs to be fixed. > > The release build passes all tests too. (convolution-mono-mono passes.) Ray, thank you for doing the tests. 767523 56 bfulgham 2012-11-14 17:42:33 -0800 If this patch looked good back in March, why hasn't it been approved or landed? 768716 57 132954 bfulgham 2012-11-15 17:10:38 -0800 Comment on attachment 132954 Patch I'm landing based on prior reviewers comments. It seems like it was simply lost in the shuffle at some point. 768725 58 132954 webkit.review.bot 2012-11-15 17:20:24 -0800 Comment on attachment 132954 Patch Clearing flags on attachment: 132954 Committed r134867: <http://trac.webkit.org/changeset/134867> 768726 59 webkit.review.bot 2012-11-15 17:20:31 -0800 All reviewed patches have been landed. Closing bug. 769522 60 webkit.review.bot 2012-11-16 11:43:33 -0800 Re-opened since this is blocked by bug 102544 769526 61 bfulgham 2012-11-16 11:45:41 -0800 Reopened because this change broke one of the security tests. See https://bugs.webkit.org/show_bug.cgi?id=102524 for details. 769530 62 bfulgham 2012-11-16 11:47:12 -0800 Rollout was done via https://bugs.webkit.org/show_bug.cgi?id=102544. 907175 63 danceoffwithyourpantsoff 2013-07-09 10:24:55 -0700 I might be completely wrong (I just happened to skim past this patch), but shouldn't the SSE2 code be a runtime check? The #ifdef SSE2 is only good for knowing if you can use the intrinsics or not (if the compiler supports SSE2). I don't know the current standing expectation of having SSE2 support is, but last I remember it wasn't considered a requirement: https://groups.google.com/a/chromium.org/forum/#!topic/chromium-dev/66p8sJtluJA 907185 64 rtoy 2013-07-09 10:43:54 -0700 (In reply to comment #63) > I might be completely wrong (I just happened to skim past this patch), but shouldn't the SSE2 code be a runtime check? The #ifdef SSE2 is only good for knowing if you can use the intrinsics or not (if the compiler supports SSE2). I don't know the current standing expectation of having SSE2 support is, but last I remember it wasn't considered a requirement: > > https://groups.google.com/a/chromium.org/forum/#!topic/chromium-dev/66p8sJtluJA I think you are correct. It's been a while since I looked, but I think on Linux and Windows the compiler flag that enables SSE2 support is not set, so the SSE2 code is never compiled in. On Mac, SSE2 has always been available on x86 machines and I think SSE2 code is compiled by default. There should be no issues with this code. In any case, this particular patch was rolled out. Ideally, we would like a runtime check and use the SSE2 code when available, falling back to the C reference when SSE2 is not available. 121082 2012-01-04 01:17:37 -0800 2012-02-09 23:44:37 -0800 Patch Patch text/plain 4400 xingnan.wang ZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZyBiL1NvdXJjZS9XZWJDb3JlL0No YW5nZUxvZwppbmRleCAyNzBiNTllLi4zODZiNjAxIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29y ZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTYg QEAKKzIwMTItMDEtMDQgIFhpbmduYW4gV2FuZyAgPHhpbmduYW4ud2FuZ0BpbnRlbC5jb20+CisK KyAgICAgICAgT3B0aW1pemUgdGhlIG11bHRpcGx5LWFkZCBpbiBCaXF1YWQuY3BwOjpwcm9jZXNz CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD03NTUyOAor CisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFBpcGVsaW5l IHRoZSBtdWx0aXBseS1hZGQgd2l0aCBTU0UyIGluc3RydWN0aW9ucy4KKyAgICAgICAgR2V0IGFi b3V0IDIwJSB+IDQwJSBpbXByb3ZlbWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICog cGxhdGZvcm0vYXVkaW8vQmlxdWFkLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJv Y2Vzcyk6CisKIDIwMTItMDEtMDMgIEFkYW0gQmFydGggIDxhYmFydGhAd2Via2l0Lm9yZz4KIAog ICAgICAgICBIVE1MQ29uc3RydWN0aW9uU2l0ZTo6YXR0YWNoIHNob3VsZG4ndCByZXR1cm4gYSB2 YWx1ZQpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlxdWFkLmNw cCBiL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAKaW5kZXggOGRjMWY0 Zi4uNWUyMTE3ZCAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlx dWFkLmNwcAorKysgYi9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCkBA IC00MSw2ICs0MSwxMCBAQAogI2luY2x1ZGUgPEFjY2VsZXJhdGUvQWNjZWxlcmF0ZS5oPgogI2Vu ZGlmCiAKKyNpZmRlZiBfX1NTRTJfXworI2luY2x1ZGUgPGVtbWludHJpbi5oPgorI2VuZGlmCisK IG5hbWVzcGFjZSBXZWJDb3JlIHsKIAogY29uc3QgaW50IGtCdWZmZXJTaXplID0gMTAyNDsKQEAg LTgyLDkgKzg2LDYwIEBAIHZvaWQgQmlxdWFkOjpwcm9jZXNzKGNvbnN0IGZsb2F0KiBzb3VyY2VQ LCBmbG9hdCogZGVzdFAsIHNpemVfdCBmcmFtZXNUb1Byb2Nlc3MpCiAgICAgZG91YmxlIGIyID0g bV9iMjsKICAgICBkb3VibGUgYTEgPSBtX2ExOwogICAgIGRvdWJsZSBhMiA9IG1fYTI7Ci0KKy8v IE9wdGltaXplIHRoZSBob3QgbXVsdGlwbHktYWRkIGJ5IHBpcGVsaW5pbmcgd2l0aCBTU0UyIGlu c3RydWN0aW9ucy4KKyNpZmRlZiBfX1NTRTJfXworICAgIGRvdWJsZSBuYTEgPSAtYTE7CisgICAg ZG91YmxlIG5hMiA9IC1hMjsKKyAgICBfX2FzbV9fKAorICAgICAgICAibW92bCAgICAgJTAsICAg ICAgJSVlZHhcblx0IiAvLyBtb3ZlIHNvdXJjZVAgdG8gZWR4CisgICAgICAgICJtb3ZsICAgICAl MSwgICAgICAlJWVjeFxuXHQiIC8vIG1vdmUgZGVzdFAgdG8gZWN4CisgICAgICAgICJtb3ZsICAg ICAlMiwgICAgICAlJWVheFxuXHQiIC8vIG1vdmUgbiB0byBlYXgKKyAgICAgICAgInRlc3RsICAg ICUlZWF4LCAgICUlZWF4XG5cdCIKKyAgICAgICAgImplIC5MYWJlbEVuZFxuXHQiCisgICAgICAg ICJtb3ZzcyAgICAoJSVlZHgpLCAlJXhtbTdcblx0IiAvLyBsb2FkIHggdG8geG1tN1s2MzowXQor ICAgICAgICAiY3Z0c3Myc2QgJSV4bW03LCAgJSV4bW0xXG5cdCIgLy8gY29udmVydCB4IGZyb20g ZmxvYXQgdG8gZG91YmxlCisgICAgICAgICJtb3ZscGQgICAlNCwgICAgICAlJXhtbTBcblx0IiAv LyBtb3ZlIHgyIHRvIHhtbTBbNjM6MF0KKyAgICAgICAgIm1vdmxwZCAgICU5LCAgICAgICUleG1t MlxuXHQiIC8vIG1vdmUgYjIgdG8geG1tMls2MzowXQorICAgICAgICAibW92bHBkICAgJTcsICAg ICAgJSV4bW0zXG5cdCIgLy8gbW92ZSBiMCB0byB4bW0zWzYzOjBdCisgICAgICAgICJtb3ZocGQg ICAlMywgICAgICAlJXhtbTBcblx0IiAvLyBtb3ZlIHgxIHRvIHhtbTBbMTI3OjY0XSAtLS0tPiAo eDEgeDIpCisgICAgICAgICJtb3ZocGQgICAlNiwgICAgICAlJXhtbTFcblx0IiAvLyBtb3ZlIHky IHRvIHhtbTFbMTI3OjY0XSAtLS0tPiAoeTIgeCApCisgICAgICAgICJtb3ZocGQgICAlOCwgICAg ICAlJXhtbTJcblx0IiAvLyBtb3ZlIGIxIHRvIHhtbTJbMTI3OjY0XSAtLS0tPiAoYjEgYjIpCisg ICAgICAgICJtb3ZocGQgICAlMTEsICAgICAlJXhtbTNcblx0IiAvLyBtb3ZlIGEyIHRvIHhtbTNb MTI3OjY0XSAtLS0tPiAoYTIgYjApCisgICAgICAgICJtb3ZscGQgICAlNSwgICAgICAlJXhtbTRc blx0IiAvLyBtb3ZlIHkxIHRvIHhtbTRbNjM6MF0KKyAgICAgICAgIi5MYWJlbExvb3A6XG5cdCIK KyAgICAgICAgImFkZGwgICAgICQ0LCAgICAgICUlZWR4XG5cdCIgLy8gc291cmNlUCsrCisgICAg ICAgICJtb3ZhcGQgICAlJXhtbTAsICAlJXhtbTVcblx0IiAvLyBjb3B5ICh4MSB4MikKKyAgICAg ICAgIm1vdmFwZCAgICUleG1tMSwgICUleG1tNlxuXHQiIC8vIGNvcHkgKHkyIHggKQorICAgICAg ICAic2h1ZnBkICAgJDAsICUleG1tNCwgJSV4bW0xXG5cdCIgLy8geTI9eTEKKyAgICAgICAgIm11 bHBkICAgICUleG1tMiwgICUleG1tNVxuXHQiIC8vICh4MSpiMSB4MipiMikKKyAgICAgICAgIm11 bHBkICAgICUleG1tMywgICUleG1tNlxuXHQiIC8vICh5MiphMiB4ICpiMCkKKyAgICAgICAgInNo dWZwZCAgICQxLCAlJXhtbTEsICUleG1tMFxuXHQiIC8vIHgyPXgxIHgxPXgKKyAgICAgICAgIm11 bHNkICAgICUxMCwgICAgICUleG1tNFxuXHQiIC8vIGExKnkxCisgICAgICAgICJhZGRwZCAgICAl JXhtbTYsICAlJXhtbTVcblx0IiAvLyAoeDEqYjEreTIqYTIgeDIqYjIreCpiMCkKKyAgICAgICAg InN1YmwgICAgICQxLCAgICAgICUlZWF4XG5cdCIgLy8gbi0tCisgICAgICAgICJtb3ZhcGQgICAl JXhtbTUsICAlJXhtbTZcblx0IgorICAgICAgICAibW92c3MgICAgKCUlZWR4KSwgJSV4bW03XG5c dCIgLy8gbG9hZCB4CisgICAgICAgICJjdnRzczJzZCAlJXhtbTcsICAlJXhtbTFcblx0IiAvLyBj dnQgeCBmcm9tIGZsb2F0IHRvIGRvdWJsZSAgeCA9IG5ldyB4CisgICAgICAgICJhZGRzZCAgICAl JXhtbTQsICAlJXhtbTVcblx0IiAvLyB4KmIwICsgKHgyKmIyK3gqYjApCisgICAgICAgICJzaHVm cGQgICAkMSwgJSV4bW02LCAlJXhtbTZcblx0IiAvLyAoeDEqYjEreTIqYTIgeDIqYjIreCpiMCkg LT4gKHgyKmIyK3gqYjAgeDEqYjEreTIqYTIpCisgICAgICAgICJhZGRzZCAgICAlJXhtbTYsICAl JXhtbTVcblx0IiAvLyB5CisgICAgICAgICJjdnRzZDJzcyAlJXhtbTUsICAlJXhtbTdcblx0Igor ICAgICAgICAibW92c3MgICAgJSV4bW03LCAgKCUlZWN4KVxuXHQiIC8vIHkgLT4gKmRlc3RQCisg ICAgICAgICJtb3ZhcGQgICAlJXhtbTUsICAlJXhtbTRcblx0IiAvLyB5MSA9IHkKKyAgICAgICAg ImFkZGwgICAgICQ0LCAgICAgICUlZWN4XG5cdCIgLy8gZGVzdFArKworICAgICAgICAidGVzdGwg ICAgJSVlYXgsICAgJSVlYXhcblx0IgorICAgICAgICAiam5lIC5MYWJlbExvb3Bcblx0IiAvLyB3 aGlsZSgpCisgICAgICAgICIuTGFiZWxFbmQ6XG5cdCIKKyAgICAgICAgIm1vdmhwZCAgICUleG1t MCwgICUzXG5cdCIKKyAgICAgICAgIm1vdmxwZCAgICUleG1tMCwgICU0XG5cdCIKKyAgICAgICAg Im1vdmxwZCAgICUleG1tNCwgICU1XG5cdCIKKyAgICAgICAgIm1vdmhwZCAgICUleG1tMSwgICU2 XG5cdCIKKyAgICAgICAgOgorICAgICAgICA6Im0iKHNvdXJjZVApLCAibSIoZGVzdFApLCAibSIo biksICJtIih4MSksICJtIih4MiksICJtIih5MSksICJtIih5MiksICJtIihiMCksICJtIihiMSks ICJtIihiMiksICJtIihuYTEpLCAibSIobmEyKQorICAgICAgICA6ImVheCIsICJlZHgiLCAiZWN4 IgorICAgICk7CisjZWxzZQogICAgIHdoaWxlIChuLS0pIHsKLSAgICAgICAgLy8gRklYTUU6IHRo aXMgY2FuIGJlIG9wdGltaXplZCBieSBwaXBlbGluaW5nIHRoZSBtdWx0aXBseSBhZGRzLi4uCiAg ICAgICAgIGZsb2F0IHggPSAqc291cmNlUCsrOwogICAgICAgICBmbG9hdCB5ID0gYjAqeCArIGIx KngxICsgYjIqeDIgLSBhMSp5MSAtIGEyKnkyOwogCkBAIC05Niw2ICsxNTEsNyBAQCB2b2lkIEJp cXVhZDo6cHJvY2Vzcyhjb25zdCBmbG9hdCogc291cmNlUCwgZmxvYXQqIGRlc3RQLCBzaXplX3Qg ZnJhbWVzVG9Qcm9jZXNzKQogICAgICAgICB5MiA9IHkxOwogICAgICAgICB5MSA9IHk7CiAgICAg fQorI2VuZGlmCiAKICAgICAvLyBMb2NhbCB2YXJpYWJsZXMgYmFjayB0byBtZW1iZXIuIEZsdXNo IGRlbm9ybWFscyBoZXJlIHNvIHdlCiAgICAgLy8gZG9uJ3Qgc2xvdyBkb3duIHRoZSBpbm5lciBs b29wIGFib3ZlLgo= 121300 2012-01-05 10:45:40 -0800 2012-01-05 10:45:40 -0800 Biquad layout test script biquad-test.html text/html 4176 rtoy PCFET0NUWVBFIEhUTUwgUFVCTElDICItLy9XM0MvL0RURCBIVE1MIDQuMDEgVHJhbnNpdGlvbmFs Ly9FTiI+CjxodG1sPgogIDxoZWFkPgogICAgPHRpdGxlPkJpcXVhZCBUZXN0PC90aXRsZT4KICAg IDxsaW5rIHJlbD0ic3R5bGVzaGVldCIgaHJlZj0iLi4vZmFzdC9qcy9yZXNvdXJjZXMvanMtdGVz dC1zdHlsZS5jc3MiLz4KICAgIDxzY3JpcHQgc3JjPSIuLi9mYXN0L2pzL3Jlc291cmNlcy9qcy10 ZXN0LXByZS5qcyI+PC9zY3JpcHQ+CiAgPC9oZWFkPgoKICA8Ym9keT4KICAgIDxkaXYgaWQ9ImRl c2NyaXB0aW9uIj48L2Rpdj4KICAgIDxkaXYgaWQ9ImNvbnNvbGUiPjwvZGl2PgogICAgPHNjcmlw dD4KCWRlc2NyaXB0aW9uKCJUZXN0IGJpcXVhZCBmaWx0ZXIgcmVzcG9uc2UiKTsKCgl2YXIgY29u dGV4dDsKCXZhciBmaWx0ZXI7Cgl2YXIgc2lnbmFsOwoJdmFyIGltcHVsc2U7Cgl2YXIgcmVuZGVy ZWRCdWZmZXI7Cgl2YXIgcmVuZGVyZWREYXRhOwoJdmFyIHRydWVSZXN1bHQ7Cgl2YXIgc2FtcGxl UmF0ZSA9IDQ0MTAwLjA7Cgl2YXIgcmVuZGVyTGVuZ3RoU2Vjb25kcyA9IDE7Cgl2YXIgcHVsc2VM ZW5ndGhTZWNvbmRzID0gLjE7Cgl2YXIgcHVsc2VMZW5ndGhGcmFtZXMgPSBwdWxzZUxlbmd0aFNl Y29uZHMgKiBzYW1wbGVSYXRlOwoJdmFyIGN1dG9mZiA9IDAuNTsKCXZhciBxID0gMTAuMDsKCXZh ciBnYWluID0gMC4wOwoKCS8vIEZpbHRlciBjb2VmZmljaWVudHMKCS8vIEgoeikgPSAoYjAgKyBi MS96ICsgYjIvel4yKS8oMSArIGExL3orYTIvel4yKQoJdmFyIGIwOwoJdmFyIGIxOwoJdmFyIGIy OwoJdmFyIGExOwoJdmFyIGEyOwoJCglmdW5jdGlvbiBjcmVhdGVMb3dwYXNzRmlsdGVyKGZyZXEs IHEsIGdhaW4pIHsKCSAgdmFyIGcgPSBNYXRoLnBvdygxMCwgcS8yMCk7CgkgIHZhciBkID0gTWF0 aC5zcXJ0KCg0LU1hdGguc3FydCgxNi0xNi9nL2cpKS8yKTsKCSAgdmFyIHRoZXRhID0gTWF0aC5Q SSAqIGZyZXE7CgkgIHZhciBzbiA9IGQgKiBNYXRoLnNpbih0aGV0YSkvMjsKCSAgdmFyIGJldGEg PSAxLzIqKDEgLSBzbikvKDEgKyBzbik7CgkgIHZhciBnYW1tYSA9ICgxLzIgKyBiZXRhKSpNYXRo LmNvcyh0aGV0YSk7CgkgIHZhciBhbHBoYSA9IDEvNCooMS8yICsgYmV0YSAtIGdhbW1hKTsKCSAg YjAgPSAyICogYWxwaGE7CgkgIGIxID0gNCAqIGFscGhhOwoJICBiMiA9IDIgKiBhbHBoYTsKCSAg YTEgPSAyICogKC1nYW1tYSk7CgkgIGEyID0gMiAqIGJldGE7Cgl9CgoJZnVuY3Rpb24gZmlsdGVy RGF0YShzaWduYWwsIGxlbikgewoJICB2YXIgeSA9IG5ldyBBcnJheShsZW4pOwoJICB5WzBdID0g YjAqc2lnbmFsWzBdOwoJICB5WzFdID0gYjAqc2lnbmFsWzFdICsgYjEqc2lnbmFsWzBdIC0gYTEq eVswXTsKCSAgZm9yICh2YXIgayA9IDI7IGsgPCBzaWduYWwubGVuZ3RoOyArK2spIHsKCSAgICB5 W2tdID0gYjAqc2lnbmFsW2tdICsgYjEqc2lnbmFsW2stMV0gKyBiMipzaWduYWxbay0yXSAtIGEx Knlbay0xXSAtIGEyKnlbay0yXTsKCSAgfQoJICBmb3IgKHZhciBrID0gc2lnbmFsLmxlbmd0aDsg ayA8IGxlbjsgKytrKSB7CgkgICAgeVtrXSA9IC0gYTEqeVtrLTFdIC0gYTIqeVtrLTJdOwoJICB9 CgoJICByZXR1cm4geTsKCX0KCQoJZnVuY3Rpb24gY2hlY2tGaWx0ZXJSZXNwb25zZShyZWZlcmVu Y2UpIHsKCSAgcmV0dXJuIGZ1bmN0aW9uKGV2ZW50KSB7CgkgICAgcmVuZGVyZWRCdWZmZXIgPSBl dmVudC5yZW5kZXJlZEJ1ZmZlcjsKCSAgICByZW5kZXJlZERhdGEgPSByZW5kZXJlZEJ1ZmZlci5n ZXRDaGFubmVsRGF0YSgwKTsKCSAgICB2YXIgbGVuID0gTWF0aC5taW4ocmVuZGVyZWREYXRhLmxl bmd0aCwgcmVmZXJlbmNlLmxlbmd0aCk7CgkgICAgdmFyIHN1Y2Nlc3MgPSB0cnVlOwoJICAgIHZh ciBtYXhFcnJvciA9IDA7CgkgICAgdmFyIG1heFBvc2l0aW9uID0gMDsKCgkKCSAgICAvLyBhbGxv d2VkRXJyb3IgZGV0ZXJtaW5lZCBleHBlcmltZW50YWxseS4gIFNldCB0byAwLCBhbmQKCSAgICAv LyBsb29rIGF0IHRoZSBtYXggZXJyb3IgcHJpbnRlZCBvdXQuICBBZGp1c3QKCSAgICAvLyBhY2Nv cmRpbmdseS4gIE5vdGUgdGhhdCBpbnRlcm5hbGx5LCBhcml0aG1ldGljIGlzIGRvbmUKCSAgICAv LyB1c2luZyBzaW5nbGUtcHJlY2lzaW9uLCBzbyB0aGUgZXJyb3IgY2Fubm90IGJlIG11Y2gKCSAg ICAvLyBiZXR0ZXIgdGhhbiBzaW5nbGUtcHJlY2lzaW9uIGVwc2lsb24gb2YgYWJvdXQgNS45NmUt OC4KCSAgICB2YXIgYWxsb3dlZEVycm9yID0gNWUtODsKCgkgICAgZm9yICh2YXIgayA9IDA7IGsg PCBsZW47ICsraykgewoJICAgICAgdmFyIGVyciA9IE1hdGguYWJzKHJlbmRlcmVkRGF0YVtrXSAt IHJlZmVyZW5jZVtrXSk7CgkgICAgICBpZiAoZXJyID4gbWF4RXJyb3IpIHsKCSAgICAgICAgbWF4 RXJyb3IgPSBlcnI7CgkgICAgICAgIG1heFBvc2l0aW9uID0gazsKCSAgICAgIH0KCSAgICB9Cgkg ICAgaWYgKG1heEVycm9yIDw9IGFsbG93ZWRFcnJvcikgewoJICAgICAgdGVzdFBhc3NlZCgiTG93 cGFzcyBmaWx0ZXIgcmVzcG9uc2UgaXMgY29ycmVjdC4iKTsKCSAgICB9IGVsc2UgewoJICAgICAg dGVzdEZhaWxlZCgiTG93cGFzcyBmaWx0ZXIgcmVzcG9uc2UgaXMgaW5jb3JyZWN0LiAgTWF4IGVy ciA9ICIgKyBtYXhFcnJvciArICIgYXQgIiArIG1heFBvc2l0aW9uKTsKICAgICAgICAgICAgICBz dWNjZXNzID0gZmFsc2U7CgkgICAgfQoJICAgIGlmIChzdWNjZXNzKSB7CgkgICAgICB0ZXN0UGFz c2VkKCJUZXN0IHNpZ25hbCB3YXMgY29ycmVjdGx5IGZpbHRlcmVkLiIpOwoJICAgIH0gZWxzZSB7 CgkgICAgICB0ZXN0RmFpbGVkKCJUZXN0IHNpZ25hbCB3YXMgbm90IGNvcnJlY3RseSBmaWx0ZXJl ZC4iKTsKCSAgICB9CgkgICAgZmluaXNoSlNUZXN0KCk7CgkgIH0KCX0KCQoJZnVuY3Rpb24gcnVu VGVzdCgpIHsKLy8JICBpZiAod2luZG93LmxheW91dFRlc3RDb250cm9sbGVyKSB7Ci8vCSAgICAg IGxheW91dFRlc3RDb250cm9sbGVyLmR1bXBBc1RleHQoKTsKLy8JICAgICAgbGF5b3V0VGVzdENv bnRyb2xsZXIud2FpdFVudGlsRG9uZSgpOwovLwkgIH0KCi8vCSAgd2luZG93LmpzVGVzdElzQXN5 bmMgPSB0cnVlOwoKCSAgLy8gQ3JlYXRlIG9mZmxpbmUgYXVkaW8gY29udGV4dAoJICBjb250ZXh0 ID0gbmV3IHdlYmtpdEF1ZGlvQ29udGV4dCgyLCBzYW1wbGVSYXRlICogcmVuZGVyTGVuZ3RoU2Vj b25kcywgc2FtcGxlUmF0ZSk7CgoJICAvLyBDcmVhdGUgYW4gaW1wdWxzZQoJICBpbXB1bHNlID0g Y29udGV4dC5jcmVhdGVCdWZmZXIoMSwgcHVsc2VMZW5ndGhGcmFtZXMsIGNvbnRleHQuc2FtcGxl UmF0ZSk7CgkgIHZhciBkYXRhID0gaW1wdWxzZS5nZXRDaGFubmVsRGF0YSgwKTsKCSAgZm9yICh2 YXIgayA9IDE7IGsgPCBkYXRhLmxlbmd0aDsgKytrKSB7CgkgICAgZGF0YVtrXSA9IDA7CgkgIH0K CSAgZGF0YVswXSA9IDE7CgoJICBzaWduYWwgPSBjb250ZXh0LmNyZWF0ZUJ1ZmZlclNvdXJjZSgp OwoJICBzaWduYWwuYnVmZmVyID0gaW1wdWxzZTsKCgkgIGZpbHRlciA9IGNvbnRleHQuY3JlYXRl QmlxdWFkRmlsdGVyKCk7CgkgIGZpbHRlci50eXBlID0gZmlsdGVyLkxPV1BBU1M7CiAgICAgICAg ICBmaWx0ZXIuZnJlcXVlbmN5LnZhbHVlID0gc2FtcGxlUmF0ZS8yICogY3V0b2ZmOwogICAgICAg ICAgZmlsdGVyLlEudmFsdWUgPSBxOwogICAgICAgICAgZmlsdGVyLmdhaW4udmFsdWUgPSBnYWlu OwoJICAKCSAgc2lnbmFsLmNvbm5lY3QoZmlsdGVyKTsKCSAgZmlsdGVyLmNvbm5lY3QoY29udGV4 dC5kZXN0aW5hdGlvbik7CgoJICBjcmVhdGVMb3dwYXNzRmlsdGVyKGN1dG9mZiwgcSwgZ2Fpbik7 CgkgIGNvbnNvbGUubG9nKCJiMCA9ICIgKyBiMCk7CgkgIHRydWVSZXN1bHQgPSBmaWx0ZXJEYXRh KGRhdGEsIHNhbXBsZVJhdGUgKiByZW5kZXJMZW5ndGhTZWNvbmRzKTsKCSAgCgkgIHNpZ25hbC5u b3RlT24oMCk7CgkgIGNvbnRleHQub25jb21wbGV0ZSA9IGNoZWNrRmlsdGVyUmVzcG9uc2UodHJ1 ZVJlc3VsdCk7CgkgIGNvbnRleHQuc3RhcnRSZW5kZXJpbmcoKTsKCX0KCglydW5UZXN0KCk7Cglz dWNjZXNzZnVsbHlQYXJzZWQgPSB0cnVlOwoKICAgICAgPC9zY3JpcHQ+CgogICAgICA8c2NyaXB0 IHNyYz0iLi4vZmFzdC9qcy9yZXNvdXJjZXMvanMtdGVzdC1wb3N0LmpzIj48L3NjcmlwdD4KICA8 L2JvZHk+CjwvaHRtbD4K 126465 2012-02-09 23:44:37 -0800 2012-02-13 00:03:59 -0800 Patch Patch text/plain 4196 xingnan.wang ZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZyBiL1NvdXJjZS9XZWJDb3JlL0No YW5nZUxvZwppbmRleCBhNTI5MDNlLi41MzU1N2MyIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29y ZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTUg QEAKKzIwMTItMDItMDkgIFhpbmduYW4gV2FuZyAgPHhpbmduYW4ud2FuZ0BpbnRlbC5jb20+CisK KyAgICAgICAgT3B0aW1pemUgdGhlIG11bHRpcGx5LWFkZCBpbiBCaXF1YWQuY3BwOjpwcm9jZXNz CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD03NTUyOAor CisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFBpcGVsaW5l IHRoZSBtdWx0aXBseS1hZGQgd2l0aCBTU0UyIGluc3RydWN0aW9ucyBhbmQgZ2V0IGFib3V0IDIw JSBpbXByb3ZlbWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICogcGxhdGZvcm0vYXVk aW8vQmlxdWFkLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJvY2Vzcyk6CisKIDIw MTItMDItMDkgIFhpYW56aHUgV2FuZyAgPHdhbmd4aWFuemh1QGNocm9taXVtLm9yZz4KIAogICAg ICAgICBbQ2hyb21pdW1dIFRpbGVkTGF5ZXJDaHJvbWl1bTo6cHJvdGVjdFZpc2libGVUaWxlVGV4 dHVyZXMoKSBzaG91bGQgb25seSBwcm90ZWN0IHRoZSB2aXNpYmxlIHRleHR1cmVzCmRpZmYgLS1n aXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwIGIvU291cmNlL1dl YkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlxdWFkLmNwcAppbmRleCBjZmVlOWNmLi5hMzJlNTdjIDEw MDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCisrKyBi L1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAKQEAgLTgzLDYgKzgzLDYw IEBAIHZvaWQgQmlxdWFkOjpwcm9jZXNzKGNvbnN0IGZsb2F0KiBzb3VyY2VQLCBmbG9hdCogZGVz dFAsIHNpemVfdCBmcmFtZXNUb1Byb2Nlc3MpCiAgICAgZG91YmxlIGExID0gbV9hMTsKICAgICBk b3VibGUgYTIgPSBtX2EyOwogCisvLyBPcHRpbWl6ZSB0aGUgaG90IG11bHRpcGx5LWFkZCBieSBw aXBlbGluaW5nIHdpdGggU1NFMiBpbnN0cnVjdGlvbnMuCisjaWZkZWYgX19TU0UyX18KKyAgICBk b3VibGUgbmExID0gLWExOworICAgIGRvdWJsZSBuYTIgPSAtYTI7CisKKyAgICBfX2FzbV9fKAor ICAgICAgICAibW92bCAgICAgJTQsICAgICAgJSVlZHhcblx0IiAvLyBtb3ZlIHNvdXJjZVAgdG8g ZWR4CisgICAgICAgICJtb3ZsICAgICAlNSwgICAgICAlJWVjeFxuXHQiIC8vIG1vdmUgZGVzdFAg dG8gZWN4CisgICAgICAgICJtb3ZsICAgICAlNiwgICAgICAlJWVheFxuXHQiIC8vIG1vdmUgbiB0 byBlYXgKKyAgICAgICAgInRlc3RsICAgICUlZWF4LCAgICUlZWF4XG5cdCIKKyAgICAgICAgImpl IC5MYWJlbEVuZFxuXHQiCisgICAgICAgICJtb3ZzcyAgICAoJSVlZHgpLCAlJXhtbTdcblx0IiAv LyBsb2FkIHggdG8geG1tN1s2MzowXQorICAgICAgICAiY3Z0c3Myc2QgJSV4bW03LCAgJSV4bW0x XG5cdCIgLy8gY29udmVydCB4IGZyb20gZmxvYXQgdG8gZG91YmxlCisgICAgICAgICJtb3ZscGQg ICAlMSwgICAgICAlJXhtbTBcblx0IiAvLyBtb3ZlIHgyIHRvIHhtbTBbNjM6MF0KKyAgICAgICAg Im1vdmxwZCAgICU5LCAgICAgICUleG1tMlxuXHQiIC8vIG1vdmUgYjIgdG8geG1tMls2MzowXQor ICAgICAgICAibW92bHBkICAgJTcsICAgICAgJSV4bW0zXG5cdCIgLy8gbW92ZSBiMCB0byB4bW0z WzYzOjBdCisgICAgICAgICJtb3ZocGQgICAlMCwgICAgICAlJXhtbTBcblx0IiAvLyBtb3ZlIHgx IHRvIHhtbTBbMTI3OjY0XSAtLS0tPiAoeDEgeDIpCisgICAgICAgICJtb3ZocGQgICAlMywgICAg ICAlJXhtbTFcblx0IiAvLyBtb3ZlIHkyIHRvIHhtbTFbMTI3OjY0XSAtLS0tPiAoeTIgeCApCisg ICAgICAgICJtb3ZocGQgICAlOCwgICAgICAlJXhtbTJcblx0IiAvLyBtb3ZlIGIxIHRvIHhtbTJb MTI3OjY0XSAtLS0tPiAoYjEgYjIpCisgICAgICAgICJtb3ZocGQgICAlMTEsICAgICAlJXhtbTNc blx0IiAvLyBtb3ZlIGEyIHRvIHhtbTNbMTI3OjY0XSAtLS0tPiAoYTIgYjApCisgICAgICAgICJt b3ZscGQgICAlMiwgICAgICAlJXhtbTRcblx0IiAvLyBtb3ZlIHkxIHRvIHhtbTRbNjM6MF0KKyAg ICAgICAgIi5MYWJlbExvb3A6XG5cdCIKKyAgICAgICAgImFkZGwgICAgICQ0LCAgICAgICUlZWR4 XG5cdCIgLy8gc291cmNlUCsrCisgICAgICAgICJtb3ZhcGQgICAlJXhtbTAsICAlJXhtbTVcblx0 IiAvLyBjb3B5ICh4MSB4MikKKyAgICAgICAgIm1vdmFwZCAgICUleG1tMSwgICUleG1tNlxuXHQi IC8vIGNvcHkgKHkyIHggKQorICAgICAgICAic2h1ZnBkICAgJDAsICUleG1tNCwgJSV4bW0xXG5c dCIgLy8geTI9eTEKKyAgICAgICAgIm11bHBkICAgICUleG1tMiwgICUleG1tNVxuXHQiIC8vICh4 MSpiMSB4MipiMikKKyAgICAgICAgIm11bHBkICAgICUleG1tMywgICUleG1tNlxuXHQiIC8vICh5 MiphMiB4ICpiMCkKKyAgICAgICAgInNodWZwZCAgICQxLCAlJXhtbTEsICUleG1tMFxuXHQiIC8v IHgyPXgxIHgxPXgKKyAgICAgICAgIm11bHNkICAgICUxMCwgICAgICUleG1tNFxuXHQiIC8vIGEx KnkxCisgICAgICAgICJhZGRwZCAgICAlJXhtbTYsICAlJXhtbTVcblx0IiAvLyAoeDEqYjEreTIq YTIgeDIqYjIreCpiMCkKKyAgICAgICAgInN1YmwgICAgICQxLCAgICAgICUlZWF4XG5cdCIgLy8g bi0tCisgICAgICAgICJtb3ZhcGQgICAlJXhtbTUsICAlJXhtbTZcblx0IgorICAgICAgICAibW92 c3MgICAgKCUlZWR4KSwgJSV4bW03XG5cdCIgLy8gbG9hZCB4CisgICAgICAgICJjdnRzczJzZCAl JXhtbTcsICAlJXhtbTFcblx0IiAvLyBjdnQgeCBmcm9tIGZsb2F0IHRvIGRvdWJsZSAgeCA9IG5l dyB4CisgICAgICAgICJhZGRzZCAgICAlJXhtbTQsICAlJXhtbTVcblx0IiAvLyB4KmIwICsgKHgy KmIyK3gqYjApCisgICAgICAgICJzaHVmcGQgICAkMSwgJSV4bW02LCAlJXhtbTZcblx0IiAvLyAo eDEqYjEreTIqYTIgeDIqYjIreCpiMCkgLT4gKHgyKmIyK3gqYjAgeDEqYjEreTIqYTIpCisgICAg ICAgICJhZGRzZCAgICAlJXhtbTYsICAlJXhtbTVcblx0IiAvLyB5CisgICAgICAgICJjdnRzZDJz cyAlJXhtbTUsICAlJXhtbTdcblx0IgorICAgICAgICAibW92c3MgICAgJSV4bW03LCAgKCUlZWN4 KVxuXHQiIC8vIHkgLT4gKmRlc3RQCisgICAgICAgICJtb3ZhcGQgICAlJXhtbTUsICAlJXhtbTRc blx0IiAvLyB5MSA9IHkKKyAgICAgICAgImFkZGwgICAgICQ0LCAgICAgICUlZWN4XG5cdCIgLy8g ZGVzdFArKworICAgICAgICAidGVzdGwgICAgJSVlYXgsICAgJSVlYXhcblx0IgorICAgICAgICAi am5lIC5MYWJlbExvb3Bcblx0IiAvLyB3aGlsZSgpCisgICAgICAgICJtb3ZocGQgICAlJXhtbTAs ICAlMFxuXHQiCisgICAgICAgICJtb3ZscGQgICAlJXhtbTAsICAlMVxuXHQiCisgICAgICAgICJt b3ZscGQgICAlJXhtbTQsICAlMlxuXHQiCisgICAgICAgICJtb3ZocGQgICAlJXhtbTEsICAlM1xu XHQiCisgICAgICAgICIuTGFiZWxFbmQ6XG5cdCIKKyAgICAgICAgOiIrbSIoeDEpLCAiK20iKHgy KSwgIittIih5MSksICIrbSIoeTIpCisgICAgICAgIDoibSIoc291cmNlUCksICJtIihkZXN0UCks ICJtIihuKSwgIm0iKGIwKSwgIm0iKGIxKSwgIm0iKGIyKSwgIm0iKG5hMSksICJtIihuYTIpCisg ICAgICAgIDoiZWF4IiwgImVkeCIsICJlY3giCisgICAgKTsKKyNlbHNlCiAgICAgd2hpbGUgKG4t LSkgewogICAgICAgICAvLyBGSVhNRTogdGhpcyBjYW4gYmUgb3B0aW1pemVkIGJ5IHBpcGVsaW5p bmcgdGhlIG11bHRpcGx5IGFkZHMuLi4KICAgICAgICAgZmxvYXQgeCA9ICpzb3VyY2VQKys7CkBA IC05Niw2ICsxNTAsNyBAQCB2b2lkIEJpcXVhZDo6cHJvY2Vzcyhjb25zdCBmbG9hdCogc291cmNl UCwgZmxvYXQqIGRlc3RQLCBzaXplX3QgZnJhbWVzVG9Qcm9jZXNzKQogICAgICAgICB5MiA9IHkx OwogICAgICAgICB5MSA9IHk7CiAgICAgfQorI2VuZGlmCiAKICAgICAvLyBMb2NhbCB2YXJpYWJs ZXMgYmFjayB0byBtZW1iZXIuIEZsdXNoIGRlbm9ybWFscyBoZXJlIHNvIHdlCiAgICAgLy8gZG9u J3Qgc2xvdyBkb3duIHRoZSBpbm5lciBsb29wIGFib3ZlLgo= 126729 2012-02-13 00:03:59 -0800 2012-03-18 20:17:54 -0700 Patch Patch text/plain 4206 xingnan.wang ZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZyBiL1NvdXJjZS9XZWJDb3JlL0No YW5nZUxvZwppbmRleCA1YTQ0MzQzLi45OTgxZmFjIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29y ZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTUg QEAKKzIwMTItMDItMTIgIFhpbmduYW4gV2FuZyAgPHhpbmduYW4ud2FuZ0BpbnRlbC5jb20+CisK KyAgICAgICAgT3B0aW1pemUgdGhlIG11bHRpcGx5LWFkZCBpbiBCaXF1YWQuY3BwOjpwcm9jZXNz CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD03NTUyOAor CisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFBpcGVsaW5l IHRoZSBtdWx0aXBseS1hZGQgd2l0aCBTU0UyIGluc3RydWN0aW9ucyBhbmQgZ2V0IGFib3V0IDIw JSBpbXByb3ZlbWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICogcGxhdGZvcm0vYXVk aW8vQmlxdWFkLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJvY2Vzcyk6CisKIDIw MTItMDItMTIgIEFkYW0gQmFydGggIDxhYmFydGhAd2Via2l0Lm9yZz4KIAogICAgICAgICBOYXZp Z2F0b3Iud2Via2l0R2V0VXNlck1lZGlhIGRvZXNuJ3QgbmVlZCB0byBiZSBjdXN0b20KZGlmZiAt LWdpdCBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAgYi9Tb3VyY2Uv V2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCmluZGV4IGNmZWU5Y2YuLjI4N2Q2Nzkg MTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAKKysr IGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlxdWFkLmNwcApAQCAtODMsNiArODMs NjAgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3QgZmxvYXQqIHNvdXJjZVAsIGZsb2F0KiBk ZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAgICBkb3VibGUgYTEgPSBtX2ExOwogICAg IGRvdWJsZSBhMiA9IG1fYTI7CiAKKy8vIE9wdGltaXplIHRoZSBob3QgbXVsdGlwbHktYWRkIGJ5 IHBpcGVsaW5pbmcgd2l0aCBTU0UyIGluc3RydWN0aW9ucy4KKyNpZmRlZiBfX1NTRTJfXworICAg IGRvdWJsZSBuYTEgPSAtYTE7CisgICAgZG91YmxlIG5hMiA9IC1hMjsKKworICAgIF9fYXNtX18o CisgICAgICAgICJtb3ZsICAgICAlNCwgICAgICAlJWVkeFxuXHQiIC8vIG1vdmUgc291cmNlUCB0 byBlZHgKKyAgICAgICAgIm1vdmwgICAgICU1LCAgICAgICUlZWN4XG5cdCIgLy8gbW92ZSBkZXN0 UCB0byBlY3gKKyAgICAgICAgIm1vdmwgICAgICU2LCAgICAgICUlZWF4XG5cdCIgLy8gbW92ZSBu IHRvIGVheAorICAgICAgICAidGVzdGwgICAgJSVlYXgsICAgJSVlYXhcblx0IgorICAgICAgICAi amUgLkxhYmVsRW5kXG5cdCIKKyAgICAgICAgIm1vdnNzICAgICglJWVkeCksICUleG1tN1xuXHQi IC8vIGxvYWQgeCB0byB4bW03WzYzOjBdCisgICAgICAgICJjdnRzczJzZCAlJXhtbTcsICAlJXht bTFcblx0IiAvLyBjb252ZXJ0IHggZnJvbSBmbG9hdCB0byBkb3VibGUKKyAgICAgICAgIm1vdmxw ZCAgICUxLCAgICAgICUleG1tMFxuXHQiIC8vIG1vdmUgeDIgdG8geG1tMFs2MzowXQorICAgICAg ICAibW92bHBkICAgJTksICAgICAgJSV4bW0yXG5cdCIgLy8gbW92ZSBiMiB0byB4bW0yWzYzOjBd CisgICAgICAgICJtb3ZscGQgICAlNywgICAgICAlJXhtbTNcblx0IiAvLyBtb3ZlIGIwIHRvIHht bTNbNjM6MF0KKyAgICAgICAgIm1vdmhwZCAgICUwLCAgICAgICUleG1tMFxuXHQiIC8vIG1vdmUg eDEgdG8geG1tMFsxMjc6NjRdIC0tLS0+ICh4MSB4MikKKyAgICAgICAgIm1vdmhwZCAgICUzLCAg ICAgICUleG1tMVxuXHQiIC8vIG1vdmUgeTIgdG8geG1tMVsxMjc6NjRdIC0tLS0+ICh5MiB4ICkK KyAgICAgICAgIm1vdmhwZCAgICU4LCAgICAgICUleG1tMlxuXHQiIC8vIG1vdmUgYjEgdG8geG1t MlsxMjc6NjRdIC0tLS0+IChiMSBiMikKKyAgICAgICAgIm1vdmhwZCAgICUxMSwgICAgICUleG1t M1xuXHQiIC8vIG1vdmUgYTIgdG8geG1tM1sxMjc6NjRdIC0tLS0+IChhMiBiMCkKKyAgICAgICAg Im1vdmxwZCAgICUyLCAgICAgICUleG1tNFxuXHQiIC8vIG1vdmUgeTEgdG8geG1tNFs2MzowXQor ICAgICAgICAiLkxhYmVsTG9vcDpcblx0IgorICAgICAgICAiYWRkbCAgICAgJDQsICAgICAgJSVl ZHhcblx0IiAvLyBzb3VyY2VQKysKKyAgICAgICAgIm1vdmFwZCAgICUleG1tMCwgICUleG1tNVxu XHQiIC8vIGNvcHkgKHgxIHgyKQorICAgICAgICAibW92YXBkICAgJSV4bW0xLCAgJSV4bW02XG5c dCIgLy8gY29weSAoeTIgeCApCisgICAgICAgICJzaHVmcGQgICAkMCwgJSV4bW00LCAlJXhtbTFc blx0IiAvLyB5Mj15MQorICAgICAgICAibXVscGQgICAgJSV4bW0yLCAgJSV4bW01XG5cdCIgLy8g KHgxKmIxIHgyKmIyKQorICAgICAgICAibXVscGQgICAgJSV4bW0zLCAgJSV4bW02XG5cdCIgLy8g KHkyKmEyIHggKmIwKQorICAgICAgICAic2h1ZnBkICAgJDEsICUleG1tMSwgJSV4bW0wXG5cdCIg Ly8geDI9eDEgeDE9eAorICAgICAgICAibXVsc2QgICAgJTEwLCAgICAgJSV4bW00XG5cdCIgLy8g YTEqeTEKKyAgICAgICAgImFkZHBkICAgICUleG1tNiwgICUleG1tNVxuXHQiIC8vICh4MSpiMSt5 MiphMiB4MipiMit4KmIwKQorICAgICAgICAic3VibCAgICAgJDEsICAgICAgJSVlYXhcblx0IiAv LyBuLS0KKyAgICAgICAgIm1vdmFwZCAgICUleG1tNSwgICUleG1tNlxuXHQiCisgICAgICAgICJt b3ZzcyAgICAoJSVlZHgpLCAlJXhtbTdcblx0IiAvLyBsb2FkIHgKKyAgICAgICAgImN2dHNzMnNk ICUleG1tNywgICUleG1tMVxuXHQiIC8vIGN2dCB4IGZyb20gZmxvYXQgdG8gZG91YmxlICB4ID0g bmV3IHgKKyAgICAgICAgImFkZHNkICAgICUleG1tNCwgICUleG1tNVxuXHQiIC8vIGExKnkxICsg KHgyKmIyK3gqYjApCisgICAgICAgICJzaHVmcGQgICAkMSwgJSV4bW02LCAlJXhtbTZcblx0IiAv LyAoeDEqYjEreTIqYTIgeDIqYjIreCpiMCkgLT4gKHgyKmIyK3gqYjAgeDEqYjEreTIqYTIpCisg ICAgICAgICJhZGRzZCAgICAlJXhtbTYsICAlJXhtbTVcblx0IiAvLyB5CisgICAgICAgICJjdnRz ZDJzcyAlJXhtbTUsICAlJXhtbTdcblx0IgorICAgICAgICAibW92c3MgICAgJSV4bW03LCAgKCUl ZWN4KVxuXHQiIC8vIHkgLT4gKmRlc3RQCisgICAgICAgICJtb3ZhcGQgICAlJXhtbTUsICAlJXht bTRcblx0IiAvLyB5MSA9IHkKKyAgICAgICAgImFkZGwgICAgICQ0LCAgICAgICUlZWN4XG5cdCIg Ly8gZGVzdFArKworICAgICAgICAidGVzdGwgICAgJSVlYXgsICAgJSVlYXhcblx0IgorICAgICAg ICAiam5lIC5MYWJlbExvb3Bcblx0IiAvLyB3aGlsZSgpCisgICAgICAgICJtb3ZocGQgICAlJXht bTAsICAlMFxuXHQiCisgICAgICAgICJtb3ZscGQgICAlJXhtbTAsICAlMVxuXHQiCisgICAgICAg ICJtb3ZscGQgICAlJXhtbTQsICAlMlxuXHQiCisgICAgICAgICJtb3ZocGQgICAlJXhtbTEsICAl M1xuXHQiCisgICAgICAgICIuTGFiZWxFbmQ6XG5cdCIKKyAgICAgICAgOiIrbSIoeDEpLCAiK20i KHgyKSwgIittIih5MSksICIrbSIoeTIpCisgICAgICAgIDoibSIoc291cmNlUCksICJtIihkZXN0 UCksICJtIihuKSwgIm0iKGIwKSwgIm0iKGIxKSwgIm0iKGIyKSwgIm0iKG5hMSksICJtIihuYTIp CisgICAgICAgIDoiZWF4IiwgImVkeCIsICJlY3giLCAieG1tMCIsICJ4bW0xIiwgInhtbTIiLCAi eG1tMyIsICJ4bW00IiwgInhtbTUiLCAieG1tNiIsICJ4bW03IgorICAgICk7CisjZWxzZQogICAg IHdoaWxlIChuLS0pIHsKICAgICAgICAgLy8gRklYTUU6IHRoaXMgY2FuIGJlIG9wdGltaXplZCBi eSBwaXBlbGluaW5nIHRoZSBtdWx0aXBseSBhZGRzLi4uCiAgICAgICAgIGZsb2F0IHggPSAqc291 cmNlUCsrOwpAQCAtOTYsNiArMTUwLDcgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3QgZmxv YXQqIHNvdXJjZVAsIGZsb2F0KiBkZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAgICAg ICAgeTIgPSB5MTsKICAgICAgICAgeTEgPSB5OwogICAgIH0KKyNlbmRpZgogCiAgICAgLy8gTG9j YWwgdmFyaWFibGVzIGJhY2sgdG8gbWVtYmVyLiBGbHVzaCBkZW5vcm1hbHMgaGVyZSBzbyB3ZQog ICAgIC8vIGRvbid0IHNsb3cgZG93biB0aGUgaW5uZXIgbG9vcCBhYm92ZS4K 132095 2012-03-15 11:51:01 -0700 2012-03-15 11:51:01 -0700 Updated patch to handle 32 and 64-bit builds biquad.patch text/plain 4413 rtoy ZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAgYi9T b3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCmluZGV4IGQ2Nzk5YzYuLmUw N2M5YzAgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5j cHAKKysrIGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlxdWFkLmNwcApAQCAtOTYs NiArOTYsOTUgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3QgZmxvYXQqIHNvdXJjZVAsIGZs b2F0KiBkZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAgICBkb3VibGUgYTEgPSBtX2Ex OwogICAgIGRvdWJsZSBhMiA9IG1fYTI7CiAKKy8vIE9wdGltaXplIHRoZSBob3QgbXVsdGlwbHkt YWRkIGJ5IHBpcGVsaW5pbmcgd2l0aCBTU0UyIGluc3RydWN0aW9ucy4KKyNpZmRlZiBfX1NTRTJf XworICAgIGRvdWJsZSBuYTEgPSAtYTE7CisgICAgZG91YmxlIG5hMiA9IC1hMjsKKworICAgIGZw cmludGYoc3RkZXJyLCAicGlkID0gJWRcbiIsIGdldHBpZCgpKTsKKworICAgIF9fYXNtX18oCisj aWYgZGVmaW5lZChfX0xQNjRfXykKKyAgICAgICAgIm1vdnEgICAgJTQsICAgICAgJSVyZHhcblx0 IiAvLyBtb3ZlIHNvdXJjZVAgdG8gcmR4CisgICAgICAgICJtb3ZxICAgICU1LCAgICAgICUlcmN4 XG5cdCIgLy8gbW92ZSBkZXN0UCB0byByY3gKKyNlbHNlCisgICAgICAgICJtb3ZsICAgICAlNCwg ICAgICAlJWVkeFxuXHQiIC8vIG1vdmUgc291cmNlUCB0byBlZHgKKyAgICAgICAgIm1vdmwgICAg ICU1LCAgICAgICUlZWN4XG5cdCIgLy8gbW92ZSBkZXN0UCB0byBlY3gKKyNlbmRpZgorICAgICAg ICAibW92bCAgICAgJTYsICAgICAgJSVlYXhcblx0IiAvLyBtb3ZlIG4gdG8gZWF4CisgICAgICAg ICJ0ZXN0bCAgICAlJWVheCwgICAlJWVheFxuXHQiCisgICAgICAgICJqZSAuTGFiZWxFbmRcblx0 IgorI2lmIGRlZmluZWQoX19MUDY0X18pCisgICAgICAgICJtb3ZzcyAgICAoJSVyZHgpLCAlJXht bTdcblx0IiAvLyBsb2FkIHggdG8geG1tN1s2MzowXQorI2Vsc2UKKyAgICAgICAgIm1vdnNzICAg ICglJWVkeCksICUleG1tN1xuXHQiIC8vIGxvYWQgeCB0byB4bW03WzYzOjBdCisjZW5kaWYKKyAg ICAgICAgImN2dHNzMnNkICUleG1tNywgICUleG1tMVxuXHQiIC8vIGNvbnZlcnQgeCBmcm9tIGZs b2F0IHRvIGRvdWJsZQorICAgICAgICAibW92bHBkICAgJTEsICAgICAgJSV4bW0wXG5cdCIgLy8g bW92ZSB4MiB0byB4bW0wWzYzOjBdCisgICAgICAgICJtb3ZscGQgICAlOSwgICAgICAlJXhtbTJc blx0IiAvLyBtb3ZlIGIyIHRvIHhtbTJbNjM6MF0KKyAgICAgICAgIm1vdmxwZCAgICU3LCAgICAg ICUleG1tM1xuXHQiIC8vIG1vdmUgYjAgdG8geG1tM1s2MzowXQorICAgICAgICAibW92aHBkICAg JTAsICAgICAgJSV4bW0wXG5cdCIgLy8gbW92ZSB4MSB0byB4bW0wWzEyNzo2NF0gLS0tLT4gKHgx IHgyKQorICAgICAgICAibW92aHBkICAgJTMsICAgICAgJSV4bW0xXG5cdCIgLy8gbW92ZSB5MiB0 byB4bW0xWzEyNzo2NF0gLS0tLT4gKHkyIHggKQorICAgICAgICAibW92aHBkICAgJTgsICAgICAg JSV4bW0yXG5cdCIgLy8gbW92ZSBiMSB0byB4bW0yWzEyNzo2NF0gLS0tLT4gKGIxIGIyKQorICAg ICAgICAibW92aHBkICAgJTExLCAgICAgJSV4bW0zXG5cdCIgLy8gbW92ZSBhMiB0byB4bW0zWzEy Nzo2NF0gLS0tLT4gKGEyIGIwKQorICAgICAgICAibW92bHBkICAgJTIsICAgICAgJSV4bW00XG5c dCIgLy8gbW92ZSB5MSB0byB4bW00WzYzOjBdCisgICAgICAgICIuTGFiZWxMb29wOlxuXHQiCisj aWYgZGVmaW5lZChfX0xQNjRfXykKKyAgICAgICAgImFkZHEgICAgICQ0LCAgICAgICUlcmR4XG5c dCIgLy8gc291cmNlUCsrCisjZWxzZQorICAgICAgICAiYWRkbCAgICAgJDQsICAgICAgJSVlZHhc blx0IiAvLyBzb3VyY2VQKysKKyNlbmRpZgorICAgICAgICAibW92YXBkICAgJSV4bW0wLCAgJSV4 bW01XG5cdCIgLy8gY29weSAoeDEgeDIpCisgICAgICAgICJtb3ZhcGQgICAlJXhtbTEsICAlJXht bTZcblx0IiAvLyBjb3B5ICh5MiB4ICkKKyAgICAgICAgInNodWZwZCAgICQwLCAlJXhtbTQsICUl eG1tMVxuXHQiIC8vIHkyPXkxCisgICAgICAgICJtdWxwZCAgICAlJXhtbTIsICAlJXhtbTVcblx0 IiAvLyAoeDEqYjEgeDIqYjIpCisgICAgICAgICJtdWxwZCAgICAlJXhtbTMsICAlJXhtbTZcblx0 IiAvLyAoeTIqYTIgeCAqYjApCisgICAgICAgICJzaHVmcGQgICAkMSwgJSV4bW0xLCAlJXhtbTBc blx0IiAvLyB4Mj14MSB4MT14CisgICAgICAgICJtdWxzZCAgICAlMTAsICAgICAlJXhtbTRcblx0 IiAvLyBhMSp5MQorICAgICAgICAiYWRkcGQgICAgJSV4bW02LCAgJSV4bW01XG5cdCIgLy8gKHgx KmIxK3kyKmEyIHgyKmIyK3gqYjApCisjaWYgZGVmaW5lZChfX0xQNjRfXykKKyAgICAgICAgInN1 YnEgICAgICQxLCAgICAgICUlcmF4XG5cdCIgLy8gbi0tCisjZWxzZQorICAgICAgICAic3VibCAg ICAgJDEsICAgICAgJSVlYXhcblx0IiAvLyBuLS0KKyNlbmRpZgorICAgICAgICAibW92YXBkICAg JSV4bW01LCAgJSV4bW02XG5cdCIKKyNpZiBkZWZpbmVkKF9fTFA2NF9fKQorICAgICAgICAibW92 c3MgICAgKCUlcmR4KSwgJSV4bW03XG5cdCIgLy8gbG9hZCB4CisjZWxzZQorICAgICAgICAibW92 c3MgICAgKCUlZWR4KSwgJSV4bW03XG5cdCIgLy8gbG9hZCB4CisjZW5kaWYKKyAgICAgICAgImN2 dHNzMnNkICUleG1tNywgICUleG1tMVxuXHQiIC8vIGN2dCB4IGZyb20gZmxvYXQgdG8gZG91Ymxl ICB4ID0gbmV3IHgKKyAgICAgICAgImFkZHNkICAgICUleG1tNCwgICUleG1tNVxuXHQiIC8vIGEx KnkxICsgKHgyKmIyK3gqYjApCisgICAgICAgICJzaHVmcGQgICAkMSwgJSV4bW02LCAlJXhtbTZc blx0IiAvLyAoeDEqYjEreTIqYTIgeDIqYjIreCpiMCkgLT4gKHgyKmIyK3gqYjAgeDEqYjEreTIq YTIpCisgICAgICAgICJhZGRzZCAgICAlJXhtbTYsICAlJXhtbTVcblx0IiAvLyB5CisgICAgICAg ICJjdnRzZDJzcyAlJXhtbTUsICAlJXhtbTdcblx0IgorI2lmIGRlZmluZWQoX19MUDY0X18pCisg ICAgICAgICJtb3ZzcyAgICAlJXhtbTcsICAoJSVyY3gpXG5cdCIgLy8geSAtPiAqZGVzdFAKKyNl bHNlCisgICAgICAgICJtb3ZzcyAgICAlJXhtbTcsICAoJSVlY3gpXG5cdCIgLy8geSAtPiAqZGVz dFAKKyNlbmRpZgorICAgICAgICAibW92YXBkICAgJSV4bW01LCAgJSV4bW00XG5cdCIgLy8geTEg PSB5CisjaWYgZGVmaW5lZChfX0xQNjRfXykKKyAgICAgICAgImFkZHEgICAgICQ0LCAgICAgICUl cmN4XG5cdCIgLy8gZGVzdFArKworI2Vsc2UKKyAgICAgICAgImFkZGwgICAgICQ0LCAgICAgICUl ZWN4XG5cdCIgLy8gZGVzdFArKworI2VuZGlmCisgICAgICAgICJ0ZXN0bCAgICAlJWVheCwgICAl JWVheFxuXHQiCisgICAgICAgICJqbmUgLkxhYmVsTG9vcFxuXHQiIC8vIHdoaWxlKCkKKyAgICAg ICAgIm1vdmhwZCAgICUleG1tMCwgICUwXG5cdCIKKyAgICAgICAgIm1vdmxwZCAgICUleG1tMCwg ICUxXG5cdCIKKyAgICAgICAgIm1vdmxwZCAgICUleG1tNCwgICUyXG5cdCIKKyAgICAgICAgIm1v dmhwZCAgICUleG1tMSwgICUzXG5cdCIKKyAgICAgICAgIi5MYWJlbEVuZDpcblx0IgorICAgICAg ICA6IittIih4MSksICIrbSIoeDIpLCAiK20iKHkxKSwgIittIih5MikKKyAgICAgICAgOiJtIihz b3VyY2VQKSwgIm0iKGRlc3RQKSwgIm0iKG4pLCAibSIoYjApLCAibSIoYjEpLCAibSIoYjIpLCAi bSIobmExKSwgIm0iKG5hMikKKyNpZiBkZWZpbmVkKF9fTFA2NF9fKQorICAgICAgICA6ImVheCIs ICJyZHgiLCAicmN4IiwgInhtbTAiLCAieG1tMSIsICJ4bW0yIiwgInhtbTMiLCAieG1tNCIsICJ4 bW01IiwgInhtbTYiLCAieG1tNyIKKyNlbHNlCisgICAgICAgIDoiZWF4IiwgImVkeCIsICJlY3gi LCAieG1tMCIsICJ4bW0xIiwgInhtbTIiLCAieG1tMyIsICJ4bW00IiwgInhtbTUiLCAieG1tNiIs ICJ4bW03IgorI2VuZGlmCisgICAgKTsKKyNlbHNlCiAgICAgd2hpbGUgKG4tLSkgewogICAgICAg ICAvLyBGSVhNRTogdGhpcyBjYW4gYmUgb3B0aW1pemVkIGJ5IHBpcGVsaW5pbmcgdGhlIG11bHRp cGx5IGFkZHMuLi4KICAgICAgICAgZmxvYXQgeCA9ICpzb3VyY2VQKys7CkBAIC0xMDksNiArMTk4 LDcgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3QgZmxvYXQqIHNvdXJjZVAsIGZsb2F0KiBk ZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAgICAgICAgeTIgPSB5MTsKICAgICAgICAg eTEgPSB5OwogICAgIH0KKyNlbmRpZgogCiAgICAgLy8gTG9jYWwgdmFyaWFibGVzIGJhY2sgdG8g bWVtYmVyLiBGbHVzaCBkZW5vcm1hbHMgaGVyZSBzbyB3ZQogICAgIC8vIGRvbid0IHNsb3cgZG93 biB0aGUgaW5uZXIgbG9vcCBhYm92ZS4K 132532 2012-03-18 20:18:12 -0700 2012-03-19 22:44:34 -0700 Patch bug-75528-20120319112125.patch text/plain 3250 xingnan.wang U3VidmVyc2lvbiBSZXZpc2lvbjogMTExMTQ1CmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggNjhkZDhjOThhNDJkYjZi Zjc5M2Q0OTgxNWIzMDMxYjE3OWY5Yjk4Mi4uNmU2ZGFhYWVlZDI1NTJiMDkxOWE0M2JkMDcwNDkx NWE1NzY0OTNlMSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE2IEBACisyMDEyLTAzLTE4ICBYaW5n bmFuIFdhbmcgIDx4aW5nbmFuLndhbmdAaW50ZWwuY29tPgorCisgICAgICAgIE9wdGltaXplIHRo ZSBtdWx0aXBseS1hZGQgaW4gQmlxdWFkLmNwcDo6cHJvY2VzcworICAgICAgICBodHRwczovL2J1 Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9NzU1MjgKKworICAgICAgICBSZXZpZXdlZCBi eSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBQaXBlbGluZSB0aGUgbXVsdGlwbHktYWRkIHdp dGggU1NFMiBpbnRyaW5zaWNzLgorICAgICAgICBHZXQgfjQ1JSBwZXJmb3JtYW5jZSBpbXByb3Zl bWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICogcGxhdGZvcm0vYXVkaW8vQmlxdWFk LmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJvY2Vzcyk6CisKIDIwMTItMDMtMTgg IERhbmEgSmFuc2VucyAgPGRhbmFrakBjaHJvbWl1bS5vcmc+CiAKICAgICAgICAgW2Nocm9taXVt XSBEb24ndCBvY2NsdWRlIG9uIG1haW4tdGhyZWFkIGJlaGluZCBsYXllcnMvc3VyZmFjZXMgd2l0 aCBpbXBsLXRocmVhZCBhbmltYXRpb25zCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0 Zm9ybS9hdWRpby9CaXF1YWQuY3BwIGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlx dWFkLmNwcAppbmRleCBkNjc5OWM2OWQzZmQ1ZDIxMTU0ZjM3ODA5MjkwMTk3MDdiNzBlNjlhLi44 NzMzYmM2NGJkYTBjODAyOGQzNDkyY2ZkZmMxY2ViMDExY2EwNDc1IDEwMDY0NAotLS0gYS9Tb3Vy Y2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCisrKyBiL1NvdXJjZS9XZWJDb3Jl L3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAKQEAgLTQxLDYgKzQxLDEwIEBACiAjaW5jbHVkZSA8 QWNjZWxlcmF0ZS9BY2NlbGVyYXRlLmg+CiAjZW5kaWYKIAorI2lmZGVmIF9fU1NFMl9fCisjaW5j bHVkZSA8ZW1taW50cmluLmg+CisjZW5kaWYKKwogbmFtZXNwYWNlIFdlYkNvcmUgewogCiBjb25z dCBpbnQga0J1ZmZlclNpemUgPSAxMDI0OwpAQCAtOTYsNiArMTAwLDQ0IEBAIHZvaWQgQmlxdWFk Ojpwcm9jZXNzKGNvbnN0IGZsb2F0KiBzb3VyY2VQLCBmbG9hdCogZGVzdFAsIHNpemVfdCBmcmFt ZXNUb1Byb2Nlc3MpCiAgICAgZG91YmxlIGExID0gbV9hMTsKICAgICBkb3VibGUgYTIgPSBtX2Ey OwogCisgICAgLy8gT3B0aW1pemUgdGhlIGhvdCBtdWx0aXBseS1hZGQgYnkgcGlwZWxpbmluZyB3 aXRoIFNTRTIgaW50cmluc2ljcy4KKyNpZmRlZiBfX1NTRTJfXworICAgIF9fbTEyOGQgbW0wID0g X21tX3NldF9wZCh4MSwgeDIpOyAvLyAoeDEsIHgyKQorICAgIF9fbTEyOGQgbW0xID0gX21tX3Nl dF9wZCh5Miwgc3RhdGljX2Nhc3Q8ZG91YmxlPigqc291cmNlUCkpOyAvLyAoeTIsIHgpCisgICAg X19tMTI4ZCBtbTIgPSBfbW1fc2V0X3BkKGIxLCBiMik7IC8vIChiMSwgYjIpCisgICAgX19tMTI4 ZCBtbTMgPSBfbW1fc2V0X3BkKC1hMiwgYjApOyAvLyAoLWEyLCBiMCkKKyAgICBfX20xMjhkIG1t NCA9IF9tbV9zZXRfcGQoMCwgeTEpOyAvLyAoMCwgeTEpCisgICAgX19tMTI4ZCBtbTU7CisgICAg X19tMTI4ZCBtbTY7CisgICAgX19tMTI4IG1tNzsKKyAgICBfX20xMjhkIG1tYTEgPSBfbW1fc2V0 X3NkKC1hMSk7CisKKyAgICB3aGlsZSAobikgeworICAgICAgICBzb3VyY2VQKys7CisgICAgICAg IG1tNiA9IG1tMTsKKyAgICAgICAgbW0xID0gX21tX3NodWZmbGVfcGQobW0xLCBtbTQsIDApOyAv LyB5MiA9IHkxCisgICAgICAgIG1tNSA9IF9tbV9tdWxfcGQobW0yLCBtbTApOyAvLyAoeDEqYjEg eDIqYjIpCisgICAgICAgIG1tNiA9IF9tbV9tdWxfcGQobW0zLCBtbTYpOyAvLyAoLXkyKmEyIHgg KmIwKQorICAgICAgICBtbTAgPSBfbW1fc2h1ZmZsZV9wZChtbTAsIG1tMSwgMSk7IC8vICh4MSA9 IHgsIHgyID0geDEpCisgICAgICAgIG1tNCA9IF9tbV9tdWxfc2QobW00LCBtbWExKTsgLy8gLXkx KmExCisgICAgICAgIG1tNSA9IF9tbV9hZGRfcGQobW01LCBtbTYpOyAvLyAoeDEqYjEteTIqYTIg eDIqYjIreCpiMCkKKyAgICAgICAgbi0tOworICAgICAgICBtbTYgPSBtbTU7CisgICAgICAgIG1t NyA9IF9tbV9sb2FkX3NzKHNvdXJjZVApOworICAgICAgICBtbTEgPSBfbW1fY3Z0c3Nfc2QobW0x LCBtbTcpOyAvLyB4ID0gKnNvdXJjZVAKKyAgICAgICAgbW01ID0gX21tX2FkZF9wZChtbTQsIG1t NSk7IC8vICh4MSpiMS15MiphMiB4MipiMit4KmIwLXkxKmExKSAKKyAgICAgICAgbW02ID0gX21t X3NodWZmbGVfcGQobW02LCBtbTYsIDEpOworICAgICAgICBtbTUgPSBfbW1fYWRkX3BkKG1tNSwg bW02KTsgLy8geQorICAgICAgICBtbTcgPSBfbW1fY3Z0c2Rfc3MobW03LCBtbTUpOworICAgICAg ICBfbW1fc3RvcmVfc3MoZGVzdFAsIG1tNyk7IC8vICpkZXN0UCA9IHkKKyAgICAgICAgbW00ID0g bW01OyAvLyB5MSA9IHkKKyAgICAgICAgZGVzdFArKzsKKyAgICB9CisgICAgX21tX3N0b3JlaF9w ZCgmeDEsIG1tMCk7CisgICAgX21tX3N0b3JlbF9wZCgmeDIsIG1tMCk7CisgICAgX21tX3N0b3Jl bF9wZCgmeTEsIG1tNCk7CisgICAgX21tX3N0b3JlaF9wZCgmeTIsIG1tMSk7CisjZWxzZQogICAg IHdoaWxlIChuLS0pIHsKICAgICAgICAgLy8gRklYTUU6IHRoaXMgY2FuIGJlIG9wdGltaXplZCBi eSBwaXBlbGluaW5nIHRoZSBtdWx0aXBseSBhZGRzLi4uCiAgICAgICAgIGZsb2F0IHggPSAqc291 cmNlUCsrOwpAQCAtMTA5LDYgKzE1MSw3IEBAIHZvaWQgQmlxdWFkOjpwcm9jZXNzKGNvbnN0IGZs b2F0KiBzb3VyY2VQLCBmbG9hdCogZGVzdFAsIHNpemVfdCBmcmFtZXNUb1Byb2Nlc3MpCiAgICAg ICAgIHkyID0geTE7CiAgICAgICAgIHkxID0geTsKICAgICB9CisjZW5kaWYgLy8gX19TU0UyX18K IAogICAgIC8vIExvY2FsIHZhcmlhYmxlcyBiYWNrIHRvIG1lbWJlci4gRmx1c2ggZGVub3JtYWxz IGhlcmUgc28gd2UKICAgICAvLyBkb24ndCBzbG93IGRvd24gdGhlIGlubmVyIGxvb3AgYWJvdmUu Cg== 132760 2012-03-19 22:34:45 -0700 2012-03-22 02:41:26 -0700 Patch bug-75528-20120320133759.patch text/plain 3719 xingnan.wang U3VidmVyc2lvbiBSZXZpc2lvbjogMTExMTQ1CmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggNjhkZDhjOThhNDJkYjZi Zjc5M2Q0OTgxNWIzMDMxYjE3OWY5Yjk4Mi4uNmU2ZGFhYWVlZDI1NTJiMDkxOWE0M2JkMDcwNDkx NWE1NzY0OTNlMSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE2IEBACisyMDEyLTAzLTE4ICBYaW5n bmFuIFdhbmcgIDx4aW5nbmFuLndhbmdAaW50ZWwuY29tPgorCisgICAgICAgIE9wdGltaXplIHRo ZSBtdWx0aXBseS1hZGQgaW4gQmlxdWFkLmNwcDo6cHJvY2VzcworICAgICAgICBodHRwczovL2J1 Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9NzU1MjgKKworICAgICAgICBSZXZpZXdlZCBi eSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBQaXBlbGluZSB0aGUgbXVsdGlwbHktYWRkIHdp dGggU1NFMiBpbnRyaW5zaWNzLgorICAgICAgICBHZXQgfjQ1JSBwZXJmb3JtYW5jZSBpbXByb3Zl bWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICogcGxhdGZvcm0vYXVkaW8vQmlxdWFk LmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJvY2Vzcyk6CisKIDIwMTItMDMtMTgg IERhbmEgSmFuc2VucyAgPGRhbmFrakBjaHJvbWl1bS5vcmc+CiAKICAgICAgICAgW2Nocm9taXVt XSBEb24ndCBvY2NsdWRlIG9uIG1haW4tdGhyZWFkIGJlaGluZCBsYXllcnMvc3VyZmFjZXMgd2l0 aCBpbXBsLXRocmVhZCBhbmltYXRpb25zCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0 Zm9ybS9hdWRpby9CaXF1YWQuY3BwIGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlx dWFkLmNwcAppbmRleCBkNjc5OWM2OWQzZmQ1ZDIxMTU0ZjM3ODA5MjkwMTk3MDdiNzBlNjlhLi42 YzMzMGFkNzBjMzY0YTdmMjA0NjJlMjNlMzg2YTRiZWZlMzZhNGE3IDEwMDY0NAotLS0gYS9Tb3Vy Y2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCisrKyBiL1NvdXJjZS9XZWJDb3Jl L3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5jcHAKQEAgLTQxLDYgKzQxLDEwIEBACiAjaW5jbHVkZSA8 QWNjZWxlcmF0ZS9BY2NlbGVyYXRlLmg+CiAjZW5kaWYKIAorI2lmZGVmIF9fU1NFMl9fCisjaW5j bHVkZSA8ZW1taW50cmluLmg+CisjZW5kaWYKKwogbmFtZXNwYWNlIFdlYkNvcmUgewogCiBjb25z dCBpbnQga0J1ZmZlclNpemUgPSAxMDI0OwpAQCAtOTYsOCArMTAwLDQ1IEBAIHZvaWQgQmlxdWFk Ojpwcm9jZXNzKGNvbnN0IGZsb2F0KiBzb3VyY2VQLCBmbG9hdCogZGVzdFAsIHNpemVfdCBmcmFt ZXNUb1Byb2Nlc3MpCiAgICAgZG91YmxlIGExID0gbV9hMTsKICAgICBkb3VibGUgYTIgPSBtX2Ey OwogCisgICAgLy8gT3B0aW1pemUgdGhlIGhvdCBtdWx0aXBseS1hZGQgYnkgcGlwZWxpbmluZyB3 aXRoIFNTRTIgaW50cmluc2ljcy4KKyNpZmRlZiBfX1NTRTJfXworICAgIF9fbTEyOGQgbW0wID0g X21tX3NldF9wZCh4MSwgeDIpOyAvLyBtbTAgPSAoeDEsIHgyKQorICAgIF9fbTEyOGQgbW0xID0g X21tX3NldF9wZCh5Miwgc3RhdGljX2Nhc3Q8ZG91YmxlPigqc291cmNlUCkpOyAvLyBtbTEgPSAo eTIsIHgpCisgICAgX19tMTI4ZCBtbTIgPSBfbW1fc2V0X3BkKGIxLCBiMik7IC8vIG1tMiA9IChi MSwgYjIpCisgICAgX19tMTI4ZCBtbTMgPSBfbW1fc2V0X3BkKC1hMiwgYjApOyAvLyBtbTMgPSAo LWEyLCBiMCkKKyAgICBfX20xMjhkIG1tNCA9IF9tbV9zZXRfcGQoMCwgeTEpOyAvLyBtbTQgPSAo MCwgeTEpCisgICAgX19tMTI4ZCBtbTU7CisgICAgX19tMTI4ZCBtbTY7CisgICAgX19tMTI4IG1t NzsKKyAgICBfX20xMjhkIG1tYTEgPSBfbW1fc2V0X3NkKC1hMSk7IC8vIG1tYTEgPSAoMCwgLWEx KQorCisgICAgd2hpbGUgKG4pIHsKKyAgICAgICAgc291cmNlUCsrOworICAgICAgICBtbTYgPSBt bTE7IC8vIG1tNiA9ICh5MiwgeCkKKyAgICAgICAgbW0xID0gX21tX3NodWZmbGVfcGQobW0xLCBt bTQsIDApOyAvLyBtbTEgPSAoeTEsIHgpCisgICAgICAgIG1tNSA9IF9tbV9tdWxfcGQobW0yLCBt bTApOyAvLyBtbTUgPSAoeDEgKiBiMSwgeDIgKiBiMikKKyAgICAgICAgbW02ID0gX21tX211bF9w ZChtbTMsIG1tNik7IC8vIG1tNiA9ICgteTIgKiBhMiwgeCAqIGIwKQorICAgICAgICBtbTAgPSBf bW1fc2h1ZmZsZV9wZChtbTAsIG1tMSwgMSk7IC8vIG1tMCA9ICh4LCB4MSkKKyAgICAgICAgbW00 ID0gX21tX211bF9zZChtbTQsIG1tYTEpOyAvLyBtbTQgPSAoMCwgLXkxICogYTEpCisgICAgICAg IG1tNSA9IF9tbV9hZGRfcGQobW01LCBtbTYpOyAvLyBtbTUgPSAoeDEgKiBiMSAtIHkyICogYTIs IHgyICogYjIgKyB4ICogYjApCisgICAgICAgIG4tLTsKKyAgICAgICAgbW02ID0gbW01OyAvLyBt bTYgPSAoeDEgKiBiMSAtIHkyICogYTIsIHgyICogYjIgKyB4ICogYjApCisgICAgICAgIG1tNyA9 IF9tbV9sb2FkX3NzKHNvdXJjZVApOyAvLyBMb2FkIG5leHQgeCB2YWx1ZSwgbW03ID0gKDAsIDAs IDAsIHgpCisgICAgICAgIG1tMSA9IF9tbV9jdnRzc19zZChtbTEsIG1tNyk7IC8vIG1tMSA9ICh5 MSwgeCkKKyAgICAgICAgbW01ID0gX21tX2FkZF9zZChtbTQsIG1tNSk7IC8vIG1tNSA9ICh4MSAq IGIxIC0geTIgKiBhMiwgeDIgKiBiMiArIHggKiBiMCAtIHkxICogYTEpCisgICAgICAgIG1tNiA9 IF9tbV9zaHVmZmxlX3BkKG1tNiwgbW02LCAxKTsgLy8gbW02ID0gKHgyICogYjIgKyB4ICogYjAs IHgxICogYjEgLSB5MiAqIGEyKQorICAgICAgICBtbTUgPSBfbW1fYWRkX3NkKG1tNSwgbW02KTsg Ly8gbW01ID0gKHgxICogYjEgLSB5MiAqIGEyLCB4ICogYjAgKyB4MSAqIGIxICsgeDIgKiBiMiAt IHkxICogYTEgLSB5MiAqIGEyKSA9ICgqLCB5KQorICAgICAgICBtbTcgPSBfbW1fY3Z0c2Rfc3Mo bW03LCBtbTUpOyAvLyBtbTcgPSAoMCwgMCwgMCwgeSkKKyAgICAgICAgX21tX3N0b3JlX3NzKGRl c3RQLCBtbTcpOyAvLyBTdG9yZSB5IHRvIGRlc3RQCisgICAgICAgIG1tNCA9IG1tNTsgLy8gbW00 ID0gKHgxICogYjEgLSB5MiAqIGEyLCB5KQorICAgICAgICBkZXN0UCsrOworICAgIH0KKyAgICBf bW1fc3RvcmVoX3BkKCZ4MSwgbW0wKTsKKyAgICBfbW1fc3RvcmVsX3BkKCZ4MiwgbW0wKTsKKyAg ICBfbW1fc3RvcmVsX3BkKCZ5MSwgbW00KTsKKyAgICBfbW1fc3RvcmVoX3BkKCZ5MiwgbW0xKTsK KyNlbHNlCiAgICAgd2hpbGUgKG4tLSkgewotICAgICAgICAvLyBGSVhNRTogdGhpcyBjYW4gYmUg b3B0aW1pemVkIGJ5IHBpcGVsaW5pbmcgdGhlIG11bHRpcGx5IGFkZHMuLi4KICAgICAgICAgZmxv YXQgeCA9ICpzb3VyY2VQKys7CiAgICAgICAgIGZsb2F0IHkgPSBiMCp4ICsgYjEqeDEgKyBiMip4 MiAtIGExKnkxIC0gYTIqeTI7CiAKQEAgLTEwOSw2ICsxNTAsNyBAQCB2b2lkIEJpcXVhZDo6cHJv Y2Vzcyhjb25zdCBmbG9hdCogc291cmNlUCwgZmxvYXQqIGRlc3RQLCBzaXplX3QgZnJhbWVzVG9Q cm9jZXNzKQogICAgICAgICB5MiA9IHkxOwogICAgICAgICB5MSA9IHk7CiAgICAgfQorI2VuZGlm IC8vIF9fU1NFMl9fCiAKICAgICAvLyBMb2NhbCB2YXJpYWJsZXMgYmFjayB0byBtZW1iZXIuIEZs dXNoIGRlbm9ybWFscyBoZXJlIHNvIHdlCiAgICAgLy8gZG9uJ3Qgc2xvdyBkb3duIHRoZSBpbm5l ciBsb29wIGFib3ZlLgo= 132954 2012-03-20 20:12:57 -0700 2012-11-15 17:20:24 -0800 Patch bug-75528-20120321111612.patch text/plain 3765 xingnan.wang U3VidmVyc2lvbiBSZXZpc2lvbjogMTExNDgxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggMjk5YzAwMWEyOTFkMzFj ZDRmNjk4MTYyNDg2NWNkMTZjZjQyZWEyMy4uZmFiMTJhZWRhMWM3NDBkMTdjNjVlNzg2Y2MzOGI4 ZjA3NmE4M2RhZiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE2IEBACisyMDEyLTAzLTIwICBYaW5n bmFuIFdhbmcgIDx4aW5nbmFuLndhbmdAaW50ZWwuY29tPgorCisgICAgICAgIE9wdGltaXplIHRo ZSBtdWx0aXBseS1hZGQgaW4gQmlxdWFkLmNwcDo6cHJvY2VzcworICAgICAgICBodHRwczovL2J1 Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9NzU1MjgKKworICAgICAgICBSZXZpZXdlZCBi eSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBQaXBlbGluZSB0aGUgbXVsdGlwbHktYWRkIHdp dGggU1NFMiBpbnRyaW5zaWNzLgorICAgICAgICBHZXQgfjQ1JSBwZXJmb3JtYW5jZSBpbXByb3Zl bWVudCBmb3IgdGhlIGZ1bmN0aW9uLgorCisgICAgICAgICogcGxhdGZvcm0vYXVkaW8vQmlxdWFk LmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OkJpcXVhZDo6cHJvY2Vzcyk6CisKIDIwMTItMDMtMTgg IFRpbSBIb3J0b24gIDx0aW1vdGh5X2hvcnRvbkBhcHBsZS5jb20+CiAKICAgICAgICAgSW5maW5p dGUgcmVwYWludCBsb29wIHdpdGggU1ZHSW1hZ2VDYWNoZSBhbmQgZGVmZXJyZWQgcmVwYWludCB0 aW1lcnMKZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2F1ZGlvL0JpcXVhZC5j cHAgYi9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9hdWRpby9CaXF1YWQuY3BwCmluZGV4IGQ2Nzk5 YzY5ZDNmZDVkMjExNTRmMzc4MDkyOTAxOTcwN2I3MGU2OWEuLjFjZjVjNmM3ZWI0YzVkY2FiOTVk NmI0OTg1MDQ3ZDYxMDFjOGI0NGMgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3Jt L2F1ZGlvL0JpcXVhZC5jcHAKKysrIGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vYXVkaW8vQmlx dWFkLmNwcApAQCAtNDEsNiArNDEsMTAgQEAKICNpbmNsdWRlIDxBY2NlbGVyYXRlL0FjY2VsZXJh dGUuaD4KICNlbmRpZgogCisjaWZkZWYgX19TU0UyX18KKyNpbmNsdWRlIDxlbW1pbnRyaW4uaD4K KyNlbmRpZgorCiBuYW1lc3BhY2UgV2ViQ29yZSB7CiAKIGNvbnN0IGludCBrQnVmZmVyU2l6ZSA9 IDEwMjQ7CkBAIC05Niw4ICsxMDAsNDUgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3QgZmxv YXQqIHNvdXJjZVAsIGZsb2F0KiBkZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAgICBk b3VibGUgYTEgPSBtX2ExOwogICAgIGRvdWJsZSBhMiA9IG1fYTI7CiAKKyAgICAvLyBPcHRpbWl6 ZSB0aGUgaG90IG11bHRpcGx5LWFkZCBieSBwaXBlbGluaW5nIHdpdGggU1NFMiBpbnRyaW5zaWNz LgorI2lmZGVmIF9fU1NFMl9fCisgICAgX19tMTI4ZCBtbTAgPSBfbW1fc2V0X3BkKHgxLCB4Mik7 IC8vIG1tMCA9ICh4MSwgeDIpCisgICAgX19tMTI4ZCBtbTEgPSBfbW1fc2V0X3BkKHkyLCBzdGF0 aWNfY2FzdDxkb3VibGU+KCpzb3VyY2VQKSk7IC8vIG1tMSA9ICh5MiwgeCkKKyAgICBfX20xMjhk IG1tMiA9IF9tbV9zZXRfcGQoYjEsIGIyKTsgLy8gbW0yID0gKGIxLCBiMikKKyAgICBfX20xMjhk IG1tMyA9IF9tbV9zZXRfcGQoLWEyLCBiMCk7IC8vIG1tMyA9ICgtYTIsIGIwKQorICAgIF9fbTEy OGQgbW00ID0gX21tX3NldF9zZCh5MSk7IC8vIG1tNCA9IHkxLCBvbmx5IHVzZSBsb3cgcGFydCBv ZiBtbTQuCisgICAgX19tMTI4ZCBtbTU7CisgICAgX19tMTI4ZCBtbTY7CisgICAgX19tMTI4IG1t NzsgLy8gT25seSB1c2UgbG93IHBhcnQgb2YgbW03LgorICAgIF9fbTEyOGQgbW1hMSA9IF9tbV9z ZXRfc2QoLWExKTsgLy8gbW1hMSA9IC1hMSwgb25seSB1c2UgbG93IHBhcnQgb2YgbW1hMS4KKwor ICAgIHdoaWxlIChuKSB7CisgICAgICAgIHNvdXJjZVArKzsKKyAgICAgICAgbW02ID0gbW0xOyAv LyBtbTYgPSAoeTIsIHgpCisgICAgICAgIG1tMSA9IF9tbV9zaHVmZmxlX3BkKG1tMSwgbW00LCAw KTsgLy8gbW0xID0gKHkxLCB4KQorICAgICAgICBtbTUgPSBfbW1fbXVsX3BkKG1tMiwgbW0wKTsg Ly8gbW01ID0gKHgxICogYjEsIHgyICogYjIpCisgICAgICAgIG1tNiA9IF9tbV9tdWxfcGQobW0z LCBtbTYpOyAvLyBtbTYgPSAoLXkyICogYTIsIHggKiBiMCkKKyAgICAgICAgbW0wID0gX21tX3No dWZmbGVfcGQobW0wLCBtbTEsIDEpOyAvLyBtbTAgPSAoeCwgeDEpCisgICAgICAgIG1tNCA9IF9t bV9tdWxfc2QobW00LCBtbWExKTsgLy8gbW00ID0gLXkxICogYTEKKyAgICAgICAgbW01ID0gX21t X2FkZF9wZChtbTUsIG1tNik7IC8vIG1tNSA9ICh4MSAqIGIxIC0geTIgKiBhMiwgeDIgKiBiMiAr IHggKiBiMCkKKyAgICAgICAgbi0tOworICAgICAgICBtbTYgPSBtbTU7IC8vIG1tNiA9ICh4MSAq IGIxIC0geTIgKiBhMiwgeDIgKiBiMiArIHggKiBiMCkKKyAgICAgICAgbW03ID0gX21tX2xvYWRf c3Moc291cmNlUCk7IC8vIG1tNyA9ICpzb3VyY2VQLCBsb2FkIG5leHQgeCB2YWx1ZS4KKyAgICAg ICAgbW0xID0gX21tX2N2dHNzX3NkKG1tMSwgbW03KTsgLy8gbW0xID0gKHkxLCB4KQorICAgICAg ICBtbTUgPSBfbW1fYWRkX3NkKG1tNCwgbW01KTsgLy8gbW01ID0geDIgKiBiMiArIHggKiBiMCAt IHkxICogYTEsIG9ubHkgY2FyZSBsb3cgcGFydCBvZiBtbTUgaGVyZS4KKyAgICAgICAgbW02ID0g X21tX3NodWZmbGVfcGQobW02LCBtbTYsIDEpOyAvLyBtbTYgPSAoeDIgKiBiMiArIHggKiBiMCwg eDEgKiBiMSAtIHkyICogYTIpCisgICAgICAgIG1tNSA9IF9tbV9hZGRfc2QobW01LCBtbTYpOyAv LyBtbTUgPSB4ICogYjAgKyB4MSAqIGIxICsgeDIgKiBiMiAtIHkxICogYTEgLSB5MiAqIGEyID0g eSwgb25seSBjYXJlIGxvdyBwYXJ0IG9mIG1tNS4KKyAgICAgICAgbW03ID0gX21tX2N2dHNkX3Nz KG1tNywgbW01KTsgLy8gbW03ID0gc3RhdGljX2Nhc3Q8ZmxvYXQ+KHkpCisgICAgICAgIF9tbV9z dG9yZV9zcyhkZXN0UCwgbW03KTsgLy8gU3RvcmUgeSB0byBkZXN0UAorICAgICAgICBtbTQgPSBt bTU7IC8vIG1tNCA9IHkKKyAgICAgICAgZGVzdFArKzsKKyAgICB9CisgICAgX21tX3N0b3JlaF9w ZCgmeDEsIG1tMCk7CisgICAgX21tX3N0b3JlbF9wZCgmeDIsIG1tMCk7CisgICAgX21tX3N0b3Jl bF9wZCgmeTEsIG1tNCk7CisgICAgX21tX3N0b3JlaF9wZCgmeTIsIG1tMSk7CisjZWxzZQogICAg IHdoaWxlIChuLS0pIHsKLSAgICAgICAgLy8gRklYTUU6IHRoaXMgY2FuIGJlIG9wdGltaXplZCBi eSBwaXBlbGluaW5nIHRoZSBtdWx0aXBseSBhZGRzLi4uCiAgICAgICAgIGZsb2F0IHggPSAqc291 cmNlUCsrOwogICAgICAgICBmbG9hdCB5ID0gYjAqeCArIGIxKngxICsgYjIqeDIgLSBhMSp5MSAt IGEyKnkyOwogCkBAIC0xMDksNiArMTUwLDcgQEAgdm9pZCBCaXF1YWQ6OnByb2Nlc3MoY29uc3Qg ZmxvYXQqIHNvdXJjZVAsIGZsb2F0KiBkZXN0UCwgc2l6ZV90IGZyYW1lc1RvUHJvY2VzcykKICAg ICAgICAgeTIgPSB5MTsKICAgICAgICAgeTEgPSB5OwogICAgIH0KKyNlbmRpZiAvLyBfX1NTRTJf XwogCiAgICAgLy8gTG9jYWwgdmFyaWFibGVzIGJhY2sgdG8gbWVtYmVyLiBGbHVzaCBkZW5vcm1h bHMgaGVyZSBzbyB3ZQogICAgIC8vIGRvbid0IHNsb3cgZG93biB0aGUgaW5uZXIgbG9vcCBhYm92 ZS4K