[PATCH] D13482: Revised Initial patch for PS4 toolchain
Filipe Cabecinhas via cfe-commits
cfe-commits at lists.llvm.org
Thu Oct 15 06:18:22 PDT 2015
Hi Katya,
On Thu, Oct 15, 2015 at 2:01 AM, Romanova, Katya <
Katya_Romanova at playstation.sony.com> wrote:
> Filipe,
>
> I have one question regarding the change in debug-option.c that you made
> to get the builds green again… (r250294).
>
>
>
> Basically you changed the generic run line that used to test the drivers
> for every single platform
>
> // RUN: %clang -### -g %s 2>&1 | FileCheck -check-prefix=CI %s
>
>
>
> into the run line that tests x86_64 drivers only.
>
> // RUN: %clang -### -g -target x86_64-unknown-unknown %s 2>&1 \
>
> // | FileCheck -check-prefix=CI %s
>
>
>
> By making this change, we reduced driver coverage for the other platforms,
> e.g. AMD. This run line will continue to pass for AMD irrespective of
> whether the AMD driver enabled or disabled emitting column information by
> default. I suspect that **non**-x86_64 developers won't be happy that we
> reduced their drivers’ test coverage without letting them know.
>
>
> Is there a better way to address this issue?
>
As far as I can tell, only the PS4 differs from the default in column info:
// PS4 defaults to no column info
if (Args.hasFlag(options::OPT_gcolumn_info, options::OPT_gno_column_info,
/*Default=*/ !IsPS4CPU))
CmdArgs.push_back("-dwarf-column-info");
There are practically no other mentions of "column.?info" (case-insensitive
regex) in the clang driver. And almost no other mentions in clang, except
for actually emitting if the setting is on.
I think the platform that deviates from the default should add their tests,
but also make the tests that assumed everyone had the same default still
pass. This will force us to hard-code a triple, since we can't test every
other triple we can use in clang. I ended up using x86_64, because it was
already used in the file anyway. And I tend to use "unknown-unknown" for
vendor and OS, since those should get us to the "default" handling,
whatever that is.
>From what I understand, it’s impossible now to execute a specific RUN line
> for all the platforms, except platform <foo>. However, it’s easy to do it
> for the whole file (XFAIL: <foo>). Does anyone know how hard it would be to
> expand lit functionality to do XFAIL: <foo> for a block of RUN lines in a
> file (e.g. XFAIL : begin: <foo> …. XFAIL : end : <foo>? Or, even a simpler
> approach: force XFAIL to take effect only for the RUN lines that follow
> XFAIL line. It will definitely add much more power/flexibility to FAIL
> functionality in lit. It might also reduce test code duplication. What do
> you think?
>
Most tests are very generic and have to do with what clang supports
"everywhere". The driver has some special cases, especially due to the
default triple, but I'm not sure it's a problem big enough to warrant doing
all that work. I would be ok with it, but don't see a huge benefit coming
out of there.
But for now (for our specific problem) maybe it’s better to remove the run
> lines that work differently for PS4 and for the rest of the world from
> debug_options.c file and place them into a ps4 specific file (e.g.
> ps4-Gcolumn-info.c) and into non-ps4 generic file (e.g gcolumn-info.c) and
> make gcolumn-info.c to XFAIL for PS4. We could do this extraction for
> gcolumn_info, debug_aranges, and anything else that is different for PS4
> and everyone else,
>
Splitting the file and adding complementary XFAILs would bring up
additional problems and maintenance burden.
We would have to copy over all of the tests from debug_options.c that we
still want to do on the PS4 (at the very least all the generic ones).
Whenever something changed, people would have to change both files in the
same way.
And since we're the only ones that diverge from the "default" and we're
testing the default and our special case, I don't think there is a problem
with the test.
Filipe
Katya.
>
>
>
>
>
> *From:* filcab at gmail.com [mailto:filcab at gmail.com] *On Behalf Of *Filipe
> Cabecinhas
> *Sent:* Wednesday, October 14, 2015 6:11 AM
> *To:* Romanova, Katya
> *Cc:* Sean Silva; Eric Christopher;
> reviews+D13482+public+a1a9627af638cb2e at reviews.llvm.org; Alex Rosenberg;
> Robinson, Paul; Jonathan Roelofs; Bedwell, Greg;
> pierregousseau14 at gmail.com; Anton Korobeynikov; Takumi NAKAMURA;
> cfe-commits
>
> *Subject:* Re: [PATCH] D13482: Revised Initial patch for PS4 toolchain
>
>
>
> Hi all,
>
>
>
> I have harnessed the powers of time zones and did the last tweak we needed
> to get it comitted.
>
> It's still going through some bots, but I don't foresee big problems. The
> fast bots liked it.
>
>
>
> Thanks to everyone that reviewed it and worked on it!
>
>
>
> Filipe
>
>
>
> On Wed, Oct 14, 2015 at 8:54 AM, Romanova, Katya <
> Katya_Romanova at playstation.sony.com> wrote:
>
> Hi Sean,
>
>
>
> I don’t think that the whole patch should have been reverted. There were
> much easier way to make the PS4 bot green. A single test could have been
> marked as “XFAIL: ps4” for a few hours until a better solution is
> implemented. Reverting this huge patch is more drastic measure that might
> cause more problems later.
>
>
>
> As you saw in my previous email, I kept an eye on the bots, noticed the
> failure on PS4 buildbot, explained why it happened in the email and ask
> community for an opinion if it could be temporarily marked as “XFAIL” for
> ps4 only. I had to leave work at certain point. When I got home, I saw that
> you reverted the whole fix. Now I will have to go through all the pain with
> maintaining a huge commit, instead of making one line change affecting one
> of the tests on PS4 platform only… But, I guess, if it’s a general
> practice, then it’s a different story. I was not aware of this practice
> and I thought that the other solution was much more reasonable.
>
>
>
> Thank you for letting me know about
> LLVM_DEFAULT_TARGET_TRIPLE=x86_64-scei-ps4. I will definitely run it before
> the next commit.
>
>
>
> Katya.
>
>
>
>
>
> *From:* Sean Silva [mailto:chisophugis at gmail.com]
> *Sent:* Tuesday, October 13, 2015 11:56 PM
> *To:* Eric Christopher
> *Cc:* reviews+D13482+public+a1a9627af638cb2e at reviews.llvm.org; Romanova,
> Katya; Alex Rosenberg; Robinson, Paul; filcab+llvm.phabricator at gmail.com;
> Jonathan Roelofs; Bedwell, Greg; pierregousseau14 at gmail.com; Anton
> Korobeynikov; Takumi NAKAMURA; cfe-commits
> *Subject:* Re: [PATCH] D13482: Revised Initial patch for PS4 toolchain
>
>
>
>
>
>
>
> On Tue, Oct 13, 2015 at 11:40 PM, Eric Christopher <echristo at gmail.com>
> wrote:
>
>
>
> On Tue, Oct 13, 2015 at 11:38 PM Sean Silva <chisophugis at gmail.com> wrote:
>
> On Tue, Oct 13, 2015 at 11:33 PM, Eric Christopher <echristo at gmail.com>
> wrote:
>
> It was already reverted, but I agree, let's get this fixed first.
>
>
>
> It was reintroduced in r250252. It is breaking
> http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/1362
>
>
>
>
>
> Ah, I missed that. Yeah, please go ahead and revert for now.
>
>
>
> reverted in r250273
>
> (bot is back to green:
> http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/1363
> )
>
>
>
> Katya - remember to run the tests
> with LLVM_DEFAULT_TARGET_TRIPLE=x86_64-scei-ps4 (and also remember to keep
> an eye on the bots after committing)
>
>
>
> -- Sean Silva
>
>
>
>
>
> Thanks!
>
>
>
> -eric
>
>
>
> -- Sean Silva
>
>
>
>
>
> Thanks!
>
>
>
> -eric
>
>
>
> On Tue, Oct 13, 2015 at 11:33 PM Sean Silva <chisophugis at gmail.com> wrote:
>
> On Tue, Oct 13, 2015 at 7:51 PM, Katya Romanova <
> Katya_Romanova at playstation.sony.com> wrote:
>
> kromanova added a comment.
>
> Hi,
>
> The initial PS4 patch caused a test failure (debug-options.c) on the PS4
> bot. I suspect that I know why the problem happens, but I'm not sure what
> will be the best way to handle it. If someone knows how to fix this test
> more "elegantly", I would appreciate their advice.
>
> FAIL: Clang :: Driver/debug-options.c (3509 of 24708)
>
> - TEST 'Clang :: Driver/debug-options.c' FAILED ********************
>
> Script:
> -------
>
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/./bin/clang
> -### -c -g
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.src/tools/clang/test/Driver/debug-options.c
> -target x86_64-linux-gnu 2>&1 |
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/./bin/FileCheck
> -check-prefix=G /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-
> ....
>
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/./bin/clang
> -### -g
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.src/tools/clang/test/Driver/debug-options.c
> 2>&1 |
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/./bin/FileCheck
> -check-prefix=CI
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.src/tools/clang/test/Driver/debug-options.c
>
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Exit Code: 1
>
> Command Output (stderr):
> ------------------------
>
> /home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.src/tools/clang/test/Driver/debug-options.c:139:8:
> error: expected string not found in input
> // CI: "-dwarf-column-info"
>
> ^
>
> <stdin>:1:1: note: scanning from here
> clang version 3.8.0 (trunk 250262)
> ^
> <stdin>:5:438: note: possible intended match here
> "/home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/bin/clang-3.8"
> "-cc1" "-triple" "x86_64-scei-ps4" "-emit-obj" "-mrelax-all"
> "-disable-free" "-main-file-name" "debug-options.c" "-mrelocation-model"
> "pic" "-pic-level" "2" "-mthread-model" "posix" "-mdisable-fp-elim"
> "-masm-verbose" "-mconstructor-aliases" "-munwind-tables" "-target-cpu"
> "btver2" "-momit-leaf-frame-pointer" "-debug-info-kind=limited"
> "-dwarf-version=4" "-backend-option" "-generate-arange-section"
> "-resource-dir"
> "/home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/bin/../lib/clang/3.8.0"
> "-fdebug-compilation-dir"
> "/home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.obj/tools/clang/test/Driver"
> "-ferror-limit" "19" "-fmessage-length" "0" "-stack-protector" "2"
> "-fdeclspec" "-fobjc-runtime=gnustep" "-fdiagnostics-show-option" "-o"
> "/tmp/debug-options-1505f5.o" "-x" "c"
> "/home/buildbot/Buildbot/Slave/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/llvm.src/tools/clang/test/Driver/debug-options.c"
>
>
>
>
>
> ^
>
> -
>
> The latest driver patch introduced a change, causing the PS4 driver *not*
> to have -gcolumn-info enabled by default.
>
> Consequently, this generic line started to fail on the PS4 bot.
> // RUN: %clang -### -g %s 2>&1 | FileCheck -check-prefix=CI %s
>
> Does someone know what will be the best way to run the test line for all
> the platforms, except PS4?
>
> In the patch, we have added a couple of PS4 specific lines to this test,
> to verify that the behavior on PS4 is correct:
>
> // RUN: %clang -### -c %s -g -target x86_64-scei-ps4 2>&1 \
> // RUN: | FileCheck -check-prefix=NOCI %s
>
> // RUN: %clang -### -c %s -g -gcolumn-info -target x86_64-scei-ps4 2>&1 \
> // RUN: | FileCheck -check-prefix=CI %s
>
> I do not want to make this test XFAIL for PS4 (though I might do it as an
> interim solution). I would also prefer to avoid duplicating most of the
> content of this file into a separate ps4-specific file.
> Any ideas how to handle this issue "more elegantly"?
>
> If nobody objects, I will mark this test as XFAIL for PS4 for a time being.
>
>
>
> Please revert until you can solve the issue.
>
>
>
> -- Sean Silva
>
>
>
>
> Katya.
>
>
>
> Repository:
> rL LLVM
>
> http://reviews.llvm.org/D13482
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20151015/ee099a05/attachment-0001.html>
More information about the cfe-commits
mailing list