linux-kernel - Re: Build error due to commit 8ab2ae655b ("default exported asm symbols to zero")

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Mon, 5 Dec 2016 14:12:33 -0800
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Guenter Roeck <linux@...ck-us.net>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Arnd Bergmann <arnd@...db.de>,
        Adam Borowski <kilobyte@...band.pl>
Subject: Re: Build error due to commit 8ab2ae655b ("default exported asm
 symbols to zero")

On Mon, Dec 5, 2016 at 1:23 PM, Guenter Roeck <linux@...ck-us.net> wrote:
>
> commit 8ab2ae655bf ("default exported asm symbols to zero") results
> in the following build error when building alpha:allmodconfig.
>
> arch/alpha/lib/callback_srm.S: Assembler messages:
> arch/alpha/lib/callback_srm.S:97: Warning:
>         Tried to .set unrecognized mode `__crc_callback_getenv'
> arch/alpha/lib/callback_srm.S:97: Error:
>         junk at end of line, first unrecognized character is `,'

Christ on a stick!

This binutils modversion problem is just _incredibly_ annoying.

I think the problem is that on alpha, ".set" is used for various
random attributes, like "noat" and "noreorder".

Hmm. Looking at what the C __kcrctab_ output is, it looks like this
for me (picking one at random):

          .section        ___kcrctab+blk_set_runtime_active,"a",@progbits
          .align 8
          .type   __kcrctab_blk_set_runtime_active, @object
          .size   __kcrctab_blk_set_runtime_active, 8
  __kcrctab_blk_set_runtime_active:
          .quad   __crc_blk_set_runtime_active

and I wonder if we should just try to emulate that as closely as possible?

For example, why does the asm code use ".weak" for the kcrctab entry?
None of the other entries do that. Maybe just removing the .weak (and
the .set) would fix things?

Over to Arnd, who presumably knows more than he ever really wanted to
know about those nasty crc entries.. Ideas?

              Linus