Re: [stable:linux-5.15.y 36/9999] arch/s390/kernel/vtime.c:132:2: error: expected absolute expression

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, May 11, 2023 at 03:21:01PM -0700, Kees Cook wrote:
> On Thu, May 11, 2023 at 05:14:16PM +0800, kernel test robot wrote:
> > FYI, the error/warning still remains.
> > 
> > tree:   https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git linux-5.15.y
> > head:   8a7f2a5c5aa1648edb4f2029c6ec33870afb7a95
> > commit: 33db9912ff7c491f839c89a08e98f755aa09598f [36/9999] ubsan: remove CONFIG_UBSAN_OBJECT_SIZE
> > [...]
> > >> arch/s390/kernel/vtime.c:132:2: error: cpu alternatives does not support instructions blocks > 254 bytes
> >            alternative_io("stpt %0\n .insn s,0xb2050000,%1\n",
> >            ^
> 
> Thanks for the report! I don't think the new errors are related to this
> removal, though. This patch was just disabling a potential build config.

Right, I am not really sure how the robot landed on this (perhaps there
was another error that was being exposed by CONFIG_UBSAN_OBJECT_SIZE
that disappeared with its removal?) but this is a problem with s390
specific code and clang's integrated assembler. I suspect it would be
fixed with one of the patches from this series that landed in 5.19 but I
have not looked too closely:

https://lore.kernel.org/20220511120532.2228616-1-hca@xxxxxxxxxxxxx/
https://lore.kernel.org/Yot+xzLZETNTxNo+@osiris/

We use LLVM_IAS=0 for s390 on 5.15 in CI, hence why we do not see this.

Cheers,
Nathan




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux