Re: [PATCH] module: Fix link failure due to invalid relocation on namespace offset

[Matthias Maennich]

On Wed, Sep 11, 2019 at 05:40:13PM +0100, Will Deacon wrote:
> On Wed, Sep 11, 2019 at 03:35:06PM +0200, Jessica Yu wrote:
> > +++ Will Deacon [11/09/19 13:26 +0100]:
> > > Commit 8651ec01daed ("module: add support for symbol namespaces.")
> > > broke linking for arm64 defconfig:
> > >
> > >  | lib/crypto/arc4.o: In function `__ksymtab_arc4_setkey':
> > >  | arc4.c:(___ksymtab+arc4_setkey+0x8): undefined reference to `no symbol'
> > >  | lib/crypto/arc4.o: In function `__ksymtab_arc4_crypt':
> > >  | arc4.c:(___ksymtab+arc4_crypt+0x8): undefined reference to `no symbol'
> > >
> > > This is because the dummy initialisation of the 'namespace_offset' field
> > > in 'struct kernel_symbol' when using EXPORT_SYMBOL on architectures with
> > > support for PREL32 locations uses an offset from an absolute address (0)
> > > in an effort to trick 'offset_to_pointer' into behaving as a NOP,
> > > allowing non-namespaced symbols to be treated in the same way as those
> > > belonging to a namespace.
> > >
> > > Unfortunately, place-relative relocations require a symbol reference
> > > rather than an absolute value and, although x86 appears to get away with
> > > this due to placing the kernel text at the top of the address space, it
> > > almost certainly results in a runtime failure if the kernel is relocated
> > > dynamically as a result of KASLR.
> > >
> > > Rework 'namespace_offset' so that a value of 0, which cannot occur for a
> > > valid namespaced symbol, indicates that the corresponding symbol does
> > > not belong to a namespace.
> > >
> > > Cc: Matthias Maennich <maennich@xxxxxxxxxx>
> > > Cc: Jessica Yu <jeyu@xxxxxxxxxx>
> > > Cc: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> > > Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> > > Fixes: 8651ec01daed ("module: add support for symbol namespaces.")
> > > Reported-by: kbuild test robot <lkp@xxxxxxxxx>
> > > Signed-off-by: Will Deacon <will@xxxxxxxxxx>
> > > ---
> > >
> > > Please note that I've not been able to test this at LPC, but it's been
> > > submitted to kernelci.
> >
> > Thanks for fixing this so quickly. I can confirm that this fixes the
> > build for arm64 defconfig and x86 built fine for me as well. I'll wait
> > a bit and apply this at the end of the day in case Matthias or anybody
> > else would like to confirm/test.
>
> FWIW, I've managed to boot arm64 Debian under QEMU and load/unload
> modules successfully with this patch applied on top of modules-next.

Thanks Will for fixing this so quickly! The patch looks good to me.

Feel free to add
Reviewed-by: Matthias Maennich <maennich@xxxxxxxxxx>
Tested-by: Matthias Maennich <maennich@xxxxxxxxxx>

Cheers,
Matthias