linux-kernel - Re: [PATCH v10 3/4] x86, mce: Add __mcsafe

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 for Android: free password hash cracker in your pocket

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

Date:	Sun, 7 Feb 2016 21:54:06 +0100
From:	Richard Weinberger <richard@....at>
To:	Borislav Petkov <bp@...en8.de>, Tony Luck <tony.luck@...el.com>
Cc:	Ingo Molnar <mingo@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Andy Lutomirski <luto@...nel.org>,
	Dan Williams <dan.j.williams@...el.com>, elliott@....com,
	Brian Gerst <brgerst@...il.com>, linux-kernel@...r.kernel.org,
	linux-mm@...ck.org, linux-nvdimm@...1.01.org, x86@...nel.org
Subject: Re: [PATCH v10 3/4] x86, mce: Add __mcsafe_copy()

Am 07.02.2016 um 17:55 schrieb Borislav Petkov:
> due to those
> 
>> +     _ASM_EXTABLE_FAULT(0b,30b)
>> +     _ASM_EXTABLE_FAULT(1b,31b)
>> +     _ASM_EXTABLE_FAULT(2b,32b)
>> +     _ASM_EXTABLE_FAULT(3b,33b)
>> +     _ASM_EXTABLE_FAULT(4b,34b)
> 
> things below and that's because ex_handler_fault() is defined in
> arch/x86/mm/extable.c and UML doesn't include that file in the build. It
> takes kernel/extable.c and lib/extable.c only.
> 
> Richi, what's the usual way to address that in UML? I.e., make an
> x86-only symbol visible to the UML build too? Define a dummy one, just
> so that it builds?

As discussed on IRC with Boris, UML offers only minimal extable support.
To get rid of the said #ifndef, UML would have to provide its own
extable.c (mostly copy&paste from arch/x86) and an advanced
struct exception_table_entry which includes the trap number.
This implies also that UML can no longer use uaccess.h from asm-generic
or has to add a new ifdef into uaccess.h to whiteout the minimal
struct exception_table_entry from there.

So, I'd vote to keep the #ifndef CONFIG_UML in memcpy_64.S.
As soon you need another #ifndef please ping me and I'll happily
bite the bullet and implement the advanced extable stuff for UML.
Deal?

Thanks,
//richard