linux-kernel - Re: [PATCH v10] MTD: LPC32xx MLC NAND driver

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: <1341066171.2115.10.camel@koala>
Date:	Sat, 30 Jun 2012 17:22:51 +0300
From:	Artem Bityutskiy <dedekind1@...il.com>
To:	Roland Stigge <stigge@...com.de>
Cc:	linux-mtd@...ts.infradead.org, linux-kernel@...r.kernel.org,
	linux-doc@...r.kernel.org, devicetree-discuss@...ts.ozlabs.org,
	dwmw2@...radead.org, kevin.wells@....com, srinivas.bakki@....com,
	linux-arm-kernel@...ts.infradead.org, hechtb@...glemail.com,
	lars@...afoo.de, b32955@...escale.com, leiwen@...vell.com,
	linux@....linux.org.uk,
	Alexandre Pereira da Silva <aletes.xgr@...il.com>
Subject: Re: [PATCH v10] MTD: LPC32xx MLC NAND driver

On Sat, 2012-06-30 at 14:32 +0200, Roland Stigge wrote:
> Hi Artem!
> 
> Thanks for the notes!
> 
> On 29/06/12 14:37, Artem Bityutskiy wrote:
> > Successfully built configuration
> > "arm-lpc32xx_defconfig,arm,arm-unknown-linux-gnueabi-", results:
> > 
> > --- before_patching.log +++ after_patching.log @@ @@ 
> > +drivers/mtd/nand/lpc32xx_mlc.c: In function 'lpc32xx_nand_probe': 
> > +drivers/mtd/nand/lpc32xx_mlc.c:679:10: warning: variable 'sr' set
> > but not used [-Wunused-but-set-variable] 
> > +drivers/mtd/nand/lpc32xx_mlc.c:561:24: error: bad constant
> > expression [sparse]
> 
> This is because sparse doesn't understand this local variable:
> 
> uint8_t buf[mtd->writesize];

mtd->writesize is typically 2048KiB (or even larger, there are MLCs with
4 and 8 KiB NAND page size). You cannot allocate that much on the stack,
linux kernel stack size is very small - only 8KiB, and there are people
who use even 4KiB stacks. So this have to be fixed.

> The above construction is quite convenient for this case and I'd
> prefer this to allocating another buffer dynamically.

Yes, in userspaces it is, but unfortunately in the kernel we cannot
afford allocating that much on the stack.

-- 
Best Regards,
Artem Bityutskiy

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)