On Mon, Feb 15, 2016 at 08:23:22PM +0100, Ard Biesheuvel wrote:
> Hi Thierry,
>
> On 15 February 2016 at 20:17, Thierry Reding <thierry.reding@xxxxxxxxx> wrote:
> > From: Thierry Reding <treding@xxxxxxxxxx>
> >
> > Changes introduced in the upstream version of libfdt pulled in by commit
> > 91feabc2e224 ("scripts/dtc: Update to upstream commit b06e55c88b9b") use
> > the strnlen() function, which isn't currently available to the EFI name-
> > space. Add it to the EFI namespace to avoid a linker error.
> >
> > Cc: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> > Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> > Cc: Will Deacon <will.deacon@xxxxxxx>
> > Cc: Rob Herring <robh@xxxxxxxxxx>
> > Signed-off-by: Thierry Reding <treding@xxxxxxxxxx>
> > ---
> > arch/arm64/kernel/image.h | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/arch/arm64/kernel/image.h b/arch/arm64/kernel/image.h
> > index c9c62cab25a4..8a1978f4a555 100644
> > --- a/arch/arm64/kernel/image.h
> > +++ b/arch/arm64/kernel/image.h
> > @@ -96,6 +96,8 @@ __efistub_strcmp = KALLSYMS_HIDE(__pi_strcmp);
> > __efistub_strncmp = KALLSYMS_HIDE(__pi_strncmp);
> > __efistub___flush_dcache_area = KALLSYMS_HIDE(__pi___flush_dcache_area);
> >
> > +__efistub_strnlen = KALLSYMS_HIDE(strnlen);
> > +
>
> To align with the existing code, we should use __pi_strnlen here, and
> declare the asm version of strnlen with ENDPIPROC()
Yes, you're absolutely right. For some reason I had missed the ARM64
implementation of strnlen() in the grep output. Sent a v2 now which is
consistent with the existing code.
Thierry
Attachment:
signature.asc
Description: PGP signature
