linux-kernel - Re: [PATCH] powerpc/booke: Avoid link stack corruption in several places

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]

Message-ID: <20210823201219.GZ1583@gate.crashing.org>
Date:   Mon, 23 Aug 2021 15:12:19 -0500
From:   Segher Boessenkool <segher@...nel.crashing.org>
To:     Christophe Leroy <christophe.leroy@...roup.eu>
Cc:     Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] powerpc/booke: Avoid link stack corruption in several places

On Mon, Aug 23, 2021 at 07:05:38PM +0200, Christophe Leroy wrote:
> Le 23/08/2021 à 17:58, Segher Boessenkool a écrit :
> >On Mon, Aug 23, 2021 at 07:53:01AM +0000, Christophe Leroy wrote:
> >>  /* Be careful, this will clobber the lr register. */
> >>  #define LOAD_REG_ADDR_PIC(reg, name)		\
> >>-	bl	0f;				\
> >>+	bcl	20,31,0f			\
> >>  0:	mflr	reg;				\
> >>  	addis	reg,reg,(name - 0b)@ha;		\
> >>  	addi	reg,reg,(name - 0b)@l;
> >
> >The code ended each line with a semicolon before, for absolutely no
> >reason that I can see, but still.  Fixing that would be nice, but only
> >doing it on one line isn't good.
> 
> Sure, forgetting the semicolon broke the build. That's because the 
> backslash removes the newline.

Ah right, one of the surprises you get from using the C preprocessor on
non-C code :-)

> The cleanest way I found to fix that quite of stuff is by using GAS macro, 
> as I did for LOAD_REG_IMMEDIATE() some time ago.

Yeah, good plan.  You can use loops and saner parameters etc. as well if
you do :-)

> >Btw.  Both the 7450 and the modern cores implementing this really need
> >this to be $+4, so it is a lot clearer to write that instead of 1f or
> >a named label.
> 
> I like that, removing unneeded labels will make it smoother and clearer. 
> I'll do it.

Cool, thanks!


Segher