On Mon, Nov 9, 2020 at 5:53 PM kai <kai.k...@windriver.com> wrote:
>
> On 11/10/20 2:58 AM, Khem Raj wrote:
>
> On Mon, Nov 9, 2020 at 12:18 AM kai <kai.k...@windriver.com> wrote:
>
> On 11/9/20 4:07 PM, Alexander Kanavin wrote:
>
> Right, but then adding _remove to the upstream recipe doesn't seem right -
> rather you should use _pn-webkitgtk override in your custom config.
>
> I suppose '-fomit-frame-pointer' is common compile flag that other people may
> also meet this error.
>
> Regards,
> Kai
>
>
> Alex
>
> On Mon, 9 Nov 2020 at 08:52, kai <kai.k...@windriver.com> wrote:
>
> On 11/9/20 3:32 PM, Alexander Kanavin wrote:
>
> This is not seen on the autobuilder, which does world builds on qemux86.
> Something might be wrong on your side?
>
> Just the option '-fno-omit-frame-pointer' which is not set by default. We
> have a build with an extra config in local.conf:
>
> SELECTED_OPTIMIZATION = "${PROFILING_OPTIMIZATION}"
>
> which expands to
>
> SELECTED_OPTIMIZATION = "${FULL_OPTIMIZATION} -fno-omit-frame-pointer
> -fvisibility=default"
>
> then it fails to compile webkitgtk. It could be reproduced with upper config
> line.
>
> since we do not have global file to control this its fine if you alter
> PROFILING_OPTIMIZATION inside this recipe
> something like PROFILING_OPTIMIZATION_remove = "..."
>-fno-omit-frame-pointer
> OK. I'll follow the snippet such as:
>
> recipes-devtools/binutils/binutils_2.35.bb:SELECTED_OPTIMIZATION_remove_mipsarch
> = "-O2"
If you want to ensure that something is built without frame pointers
then appending "-fomit-frame-pointer" is clearer (and more robust)
than trying to remove "-fno-omit-frame-pointer".
> Thanks,
> Kai
>
>
>
> Regards,
> Kai
>
>
>
>
> Alex
>
> On Mon, 9 Nov 2020 at 01:32, kai <kai.k...@windriver.com> wrote:
>
> From: Kai Kang <kai.k...@windriver.com>
>
> It fails to compile webkitgtk for qemux86 if compile option
> '-fno-omit-frame-pointer' is applied:
>
> |
> TOPDIR/tmp-glibc/work/core2-32-wrsmllib32-linux/lib32-webkitgtk/2.30.1-r0/webkitgtk-2.30.1/
> Source/JavaScriptCore/runtime/JSFinalizationRegistry.cpp:47:1:
> | internal compiler error: in find_oldest_value_reg, at regcprop.c:438
> | 47 | }
>
> Signed-off-by: Kai Kang <kai.k...@windriver.com>
> ---
> meta/recipes-sato/webkit/webkitgtk_2.30.2.bb | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/meta/recipes-sato/webkit/webkitgtk_2.30.2.bb
> b/meta/recipes-sato/webkit/webkitgtk_2.30.2.bb
> index 1acba6be1d..11fdb59034 100644
> --- a/meta/recipes-sato/webkit/webkitgtk_2.30.2.bb
> +++ b/meta/recipes-sato/webkit/webkitgtk_2.30.2.bb
> @@ -79,6 +79,7 @@ EXTRA_OECMAKE_append_arc = " -DENABLE_JIT=OFF "
> # which is not enough for binaries larger than 32 MiB
> CFLAGS_append_arc = " -mlong-calls"
> CXXFLAGS_append_arc = " -mlong-calls"
> +CXXFLAGS_remove_x86 = "-fno-omit-frame-pointer"
>
> # Javascript JIT is not supported on powerpc
> EXTRA_OECMAKE_append_powerpc = " -DENABLE_JIT=OFF "
> --
> 2.17.1
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#144455):
https://lists.openembedded.org/g/openembedded-core/message/144455
Mute This Topic: https://lists.openembedded.org/mt/78125973/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
