Re: [PATCH v5 3/7] Change the symbols order when --ffunction-sections is enabled

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

 



The fix is now included in v6 of the patch series for review.

On Fri, Oct 25, 2024 at 5:27 PM Rong Xu <xur@xxxxxxxxxx> wrote:
>
> OK. Here is a simple fix to the issue.
>
> diff --git a/arch/sparc/kernel/vmlinux.lds.S b/arch/sparc/kernel/vmlinux.lds.S
> index d317a843f7ea9..f1b86eb303404 100644
> --- a/arch/sparc/kernel/vmlinux.lds.S
> +++ b/arch/sparc/kernel/vmlinux.lds.S
> @@ -48,6 +48,11 @@ SECTIONS
>         {
>                 _text = .;
>                 HEAD_TEXT
> +               ALIGN_FUNCTION();
> +#ifdef CONFIG_SPARC64
> +               /* Match text section symbols in head_64.S first */
> +               *head_64.o(.text)
> +#endif
>                 TEXT_TEXT
>                 SCHED_TEXT
>                 LOCK_TEXT
>
> If this is OK, I'll send the updated patch.
>
> -Rong
>
> On Fri, Oct 25, 2024 at 11:14 AM Rong Xu <xur@xxxxxxxxxx> wrote:
> >
> > I can reproduce the build error.
> >
> > The reason for the error is because in CONFIG_SPARC64, the symbol
> > swapper_tsb needs to have a fixed address of 0x0000000000408000.
> >
> > In arch/sparc/kernel/vmlinux.lds.S
> >  37 #ifdef CONFIG_SPARC64
> >  38 ASSERT((swapper_tsb == 0x0000000000408000), "Error: sparc64 early
> > assembler too large")
> >  39 #endif
> >
> > This symbol resides in the .text section. The patch relocates "cold"
> > functions (tagged by .text.unlikely) to precede other symbols within
> > the .text section. This pushes swapper_tsb to a higher memory address,
> > and we have the assertion.
> >
> > However, this sparc code appears incompatible with the original linker
> > script. The original script prioritized .text.hot symbols, placing
> > them before others in the .text section. Currently, no functions are
> > marked as "hot" in the source code.  If a function were marked with
> > __attribute__((hot)), the build would also fail.
> >
> > I don't know a good solution other than adding a branch to the
> > include/asm-generic/vmlinux.lds.h.
> >
> > @Masahiro Yamada: I'd appreciate any further suggestions you might have.
> >
> > Best regards,
> >
> > -Rong
> >
> > On Thu, Oct 24, 2024 at 11:47 PM kernel test robot <lkp@xxxxxxxxx> wrote:
> > >
> > > Hi Rong,
> > >
> > > kernel test robot noticed the following build errors:
> > >
> > > [auto build test ERROR on eb952c47d154ba2aac794b99c66c3c45eb4cc4ec]
> > >
> > > url:    https://github.com/intel-lab-lkp/linux/commits/Rong-Xu/Add-AutoFDO-support-for-Clang-build/20241024-064725
> > > base:   eb952c47d154ba2aac794b99c66c3c45eb4cc4ec
> > > patch link:    https://lore.kernel.org/r/20241023224409.201771-4-xur%40google.com
> > > patch subject: [PATCH v5 3/7] Change the symbols order when --ffunction-sections is enabled
> > > config: sparc-sparc64_defconfig (https://download.01.org/0day-ci/archive/20241025/202410251411.eRNArTOZ-lkp@xxxxxxxxx/config)
> > > compiler: sparc64-linux-gcc (GCC) 13.3.0
> > > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241025/202410251411.eRNArTOZ-lkp@xxxxxxxxx/reproduce)
> > >
> > > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > > the same patch/commit), kindly add following tags
> > > | Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > > | Closes: https://lore.kernel.org/oe-kbuild-all/202410251411.eRNArTOZ-lkp@xxxxxxxxx/
> > >
> > > All errors (new ones prefixed by >>):
> > >
> > > >> sparc64-linux-ld: Error: sparc64 early assembler too large
> > >
> > > --
> > > 0-DAY CI Kernel Test Service
> > > https://github.com/intel/lkp-tests/wiki





[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