Created attachment 106454 [details] aarch64 fix While building tests on x86/aarch64 it fails to build few tests. Attached patches solve the problem, hopefully in a portable way, while keeping the functionality intact.
Created attachment 106455 [details] x86 fix
Created attachment 106456 [details] syscall fix
What is MUSL?
musl is an alternative system C library implementation for linux, like glibc or uclibc and others. see https://www.musl-libc.org/
These might fix building but they will also break the tests themselves. For instance there is no update to the scalar expected even though the line numbers will have changed.
Could you update the patches with expecteds that won't break glibc Linux?
Dear Bug Submitter, This bug has been in NEEDSINFO status with no change for at least 15 days. Please provide the requested information as soon as possible and set the bug status as REPORTED. Due to regular bug tracker maintenance, if the bug is still in NEEDSINFO status with no change in 30 days the bug will be closed as RESOLVED > WORKSFORME due to lack of needed information. For more information about our bug triaging procedures please read the wiki located here: https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging If you have already provided the requested information, please mark the bug as REPORTED so that the KDE team knows that the bug is ready to be confirmed. Thank you for helping us make KDE software even better for everyone!
(In reply to Paul Floyd from comment #6) > Could you update the patches with expecteds that won't break glibc Linux? Paul, I do not understand what 'expecteds' you are refrerring to here. If I understand it better, I will be able to make the needed change
*** Bug 453929 has been marked as a duplicate of this bug. ***
In order for there to be no breakage you need all of the following to add no new problems: make [builds Valgrind] make check [builds the testcases] make regtest [runs the regtest] regtest uses one ore more expected like [testcase].stderr.exp to capture and compare the Valgrind output. Looking again at the diffs I don't think that the expected need updating. I'd still rather not see any change for glibc in the syscall patch. Could you update that patch to do something like this #include "./../../config..h" #if defined(MUSL_LIBC) #define MODIFY_LDT(func, ptr, bytecount) syscall(SYS_modify_ldt, (func), (ptr), (bytecount) ); #else #define MODIFY_LDT(func, ptr, bytecount) __modify_ldt((func), (ptr), (bytecount) ); #endif
If you don't beat me to it I'll probably go ahead and try this tonight.
(In reply to Paul Floyd from comment #11) > If you don't beat me to it I'll probably go ahead and try this tonight. OK. I wont get to this till weekend.
This bug has been in NEEDSINFO status with no change for at least 30 days. The bug is now closed as RESOLVED > WORKSFORME due to lack of needed information. For more information about our bug triaging procedures please read the wiki located here: https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging Thank you for helping us make KDE software even better for everyone!
Memory like a sieve. I did apply the __THROW patches and changed the __modify_ldt patch so that it doesn't affect glibc in any way. commit f7e4bb4af294575432b629de4869812fdfee6feb Author: Paul Floyd <pjfloyd@wanadoo.fr> Date: Fri Jan 27 08:42:06 2023 +0100 Bug 382034 - Testcases build fixes for musl
Created attachment 158792 [details] Add missing includes to avoid 'SYS_modify_ldt is undefined' errors
There's one more fix needed (add needed includes on musl), I've attached the patch for it. Hope you don't mind reusing this ticket, which I'll reopen.
commit 306232d40e037274160de417f21fdb957e30fc73 (HEAD -> master, origin/master, origin/HEAD) Author: Paul Floyd <pjfloyd@wanadoo.fr> Date: Tue May 9 22:13:28 2023 +0200 musl: another fix for building testcases I'll try to do some musl testing before closing this.
> I'll try to do some musl testing before closing this. Did you use an x86 install of Alpine? Or is there way to add an x86 subsystem to amd64?
(In reply to Paul Floyd from comment #19) > > I'll try to do some musl testing before closing this. > > Did you use an x86 install of Alpine? > > Or is there way to add an x86 subsystem to amd64? I build valgrind using yocto set to use musl and 32bit x86. If you want to try that, do it on the most powerful (cpu cores wise) computer you have.
(In reply to Alexander Kanavin from comment #20) > I build valgrind using yocto set to use musl and 32bit x86. If you want to > try that, do it on the most powerful (cpu cores wise) computer you have. That would be my macbook ;-) I'll try installing Alpine x86 in VirtualBox.
Created attachment 158982 [details] Some more changes for musl x86 make check succeeds with this patch. Need to check that there is no impact on other platforms.
Should build now. Still around 100 regtest failures.