linux-kernel - Re: [PATCH] x86: add phys addr validity check for /dev/mem mmap

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

Message-ID: <20130402205215.GD3314@dhcp-26-164.brq.redhat.com>
Date:	Tue, 2 Apr 2013 22:52:15 +0200
From:	Frantisek Hrbata <fhrbata@...hat.com>
To:	"H. Peter Anvin" <hpa@...or.com>
Cc:	linux-kernel@...r.kernel.org, linux-mm@...ck.org,
	tglx@...utronix.de, mingo@...hat.com, x86@...nel.org,
	oleg@...hat.com, kamaleshb@...ibm.com, hechjie@...ibm.com
Subject: Re: [PATCH] x86: add phys addr validity check for /dev/mem mmap

On Tue, Apr 02, 2013 at 01:29:12PM -0700, H. Peter Anvin wrote:
> On 04/02/2013 12:10 PM, Frantisek Hrbata wrote:
> > 
> > Hi, this is exactly what the patch is doing imho. Note that the
> > valid_phys_addr_range(), which is using the high_memory, is the same as the
> > default one in drivers/char/mem.c(#ifndef ARCH_HAS_VALID_PHYS_ADDR_RANGE). I
> > just added x86 specific check for valid_mmap_phys_addr_range and moved both
> > functions to arch/x86/mm/mmap.c, rather then modifying the default generic ones.
> > This is how other archs(arm) are doing it.
> > 
> > Also valid_phys_addr_range is used just in read|write_mem and
> > valid_mmap_phys_addr_range is checked in mmap_mem and it calls phys_addr_valid
> > 
> > static inline int phys_addr_valid(resource_size_t addr)
> > {
> > #ifdef CONFIG_PHYS_ADDR_T_64BIT
> > 	return !(addr >> boot_cpu_data.x86_phys_bits);
> > #else
> >         return 1;
> > #endif
> > }                          
> > 
> > I for sure could overlooked something, but this seems right to me.
> > 
> 
> OK, this is really confusing ... which isn't a *huge* surprise (the
> entire /dev/mem code has some gigantic bugs in it.)
> 
> I think I need to do more of an in-depth review.  The other question is
> why we don't call phys_addr_valid() everywhere.

I'm not going to pretend I understand the code, but IMHO the
valid_phys_addr_range and valid_mmap_phys_addr_range in drivers/char/mem.c are
generic for all archs. If some arch wants specific version of those functions it
defines them in the arch specific code and define ARCH_HAS_VALID_PHYS_ADDR_RANGE.
The phys_addr_valid is x86 specific defined in arch/x86/mm/physaddr.h, so IMHO
it cannot be used in the generic checks. For example ARM has it's specific
checks in arch/arm/mm/mmap.c.

I reused phys_addr_valid because it is already used in ioremap(__ioremap_caller)
for the same purpose imho.

Thank you for looking into this.

> 
> 	-hpa
> 
> 

-- 
Frantisek Hrbata
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/