linux-kernel - Re: [PATCH] cs46xx memory management fixes for cs46xx_dsp_spos_create() - make sure we free and don't do pointless work.

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: <201010301406.12961.linux@rainbow-software.org>
Date:	Sat, 30 Oct 2010 14:06:07 +0200
From:	Ondrej Zary <linux@...nbow-software.org>
To:	Benny Sjöstrand <benny@...tmobility.com>
Cc:	Jesper Juhl <jj@...osbits.net>, linux-kernel@...r.kernel.org,
	Jaroslav Kysela <perex@...ex.cz>, Takashi Iwai <tiwai@...e.de>,
	alsa-devel@...a-project.org
Subject: Re: [PATCH] cs46xx memory management fixes for cs46xx_dsp_spos_create() - make sure we free and don't do pointless work.

On Saturday 30 October 2010 00:18:57 Benny Sjöstrand wrote:
> Hello again!
>
> Just doing a reply-all.
> It's has been many years since I did anything to the cs46xx driver, so
> I'm wondering
> if there's anyone out there still using a cs46xx sound card?
> I think the changes look's correct, but I can't test it, I do not have a
> cs46xx hardware anymore.

Sure, there are many of these cards "out there". I just got a "new" Terratec 
DMX Xfire 1024 (CS4624). I'll test the patch.

>
> /Benny
>
> Jesper Juhl wrote:
> > Hi,
> >
> > When reading through sound/pci/cs46xx/dsp_spos.c I noticed a couple of
> > things in cs46xx_dsp_spos_create().
> >
> > It seems to me that we don't always free the various memory buffers we
> > allocate and we also do some work (structure member assignment) early,
> > that is completely pointless if some of the memory allocations fail and
> > we end up just aborting the whole thing.
> >
> > I don't have hardware to test, so the patch below is compile tested only,
> > but it makes the following changes:
> >
> > - Make sure we always free all allocated memory on failures.
> > - Don't do pointless work assigning to structure members before we know
> >   all memory allocations, that may abort progress, have completed
> >   successfully.
> > - Remove some trailing whitespace.
> >
> > If it looks ok, please merge, otherwise I'd be interested in knowing
> > what's wrong so I can improve it.
> >
> >
> > Signed-off-by: Jesper Juhl <jj@...osbits.net>
> > ---
> >  dsp_spos.c |   33 +++++++++++----------------------
> >  1 file changed, 11 insertions(+), 22 deletions(-)
> >
> > diff --git a/sound/pci/cs46xx/dsp_spos.c b/sound/pci/cs46xx/dsp_spos.c
> > index 3e5ca8f..e377287 100644
> > --- a/sound/pci/cs46xx/dsp_spos.c
> > +++ b/sound/pci/cs46xx/dsp_spos.c
> > @@ -225,39 +225,25 @@ struct dsp_spos_instance *cs46xx_dsp_spos_create
> > (struct snd_cs46xx * chip) {
> >  	struct dsp_spos_instance * ins = kzalloc(sizeof(struct
> > dsp_spos_instance), GFP_KERNEL);
> >
> > -	if (ins == NULL)
> > +	if (ins == NULL)
> >  		return NULL;
> >
> >  	/* better to use vmalloc for this big table */
> > -	ins->symbol_table.nsymbols = 0;
> >  	ins->symbol_table.symbols = vmalloc(sizeof(struct dsp_symbol_entry) *
> >  					    DSP_MAX_SYMBOLS);
> > -	ins->symbol_table.highest_frag_index = 0;
> > -
> > -	if (ins->symbol_table.symbols == NULL) {
> > +	ins->code.data = kmalloc(DSP_CODE_BYTE_SIZE, GFP_KERNEL);
> > +	ins->modules = kmalloc(sizeof(struct dsp_module_desc) *
> > DSP_MAX_MODULES, GFP_KERNEL); +	if (!ins->symbol_table.symbols ||
> > !ins->code.data || !ins->modules) { cs46xx_dsp_spos_destroy(chip);
> >  		goto error;
> >  	}
> > -
> > +	ins->symbol_table.nsymbols = 0;
> > +	ins->symbol_table.highest_frag_index = 0;
> >  	ins->code.offset = 0;
> >  	ins->code.size = 0;
> > -	ins->code.data = kmalloc(DSP_CODE_BYTE_SIZE, GFP_KERNEL);
> > -
> > -	if (ins->code.data == NULL) {
> > -		cs46xx_dsp_spos_destroy(chip);
> > -		goto error;
> > -	}
> > -
> >  	ins->nscb = 0;
> >  	ins->ntask = 0;
> > -
> >  	ins->nmodules = 0;
> > -	ins->modules = kmalloc(sizeof(struct dsp_module_desc) *
> > DSP_MAX_MODULES, GFP_KERNEL); -
> > -	if (ins->modules == NULL) {
> > -		cs46xx_dsp_spos_destroy(chip);
> > -		goto error;
> > -	}
> >
> >  	/* default SPDIF input sample rate
> >  	   to 48000 khz */
> > @@ -271,8 +257,8 @@ struct dsp_spos_instance *cs46xx_dsp_spos_create
> > (struct snd_cs46xx * chip)
> >
> >  	/* set left and right validity bits and
> >  	   default channel status */
> > -	ins->spdif_csuv_default =
> > -		ins->spdif_csuv_stream =
> > +	ins->spdif_csuv_default =
> > +		ins->spdif_csuv_stream =
> >  	 /* byte 0 */  ((unsigned int)_wrap_all_bits( 
> > (SNDRV_PCM_DEFAULT_CON_SPDIF        & 0xff)) << 24) | /* byte 1 */ 
> > ((unsigned int)_wrap_all_bits( ((SNDRV_PCM_DEFAULT_CON_SPDIF >> 8) &
> > 0xff)) << 16) | /* byte 3 */   (unsigned int)_wrap_all_bits( 
> > (SNDRV_PCM_DEFAULT_CON_SPDIF >> 24) & 0xff) | @@ -281,6 +267,9 @@ struct
> > dsp_spos_instance *cs46xx_dsp_spos_create (struct snd_cs46xx * chip)
> > return ins;
> >
> >  error:
> > +	kfree(ins->modules);
> > +	kfree(ins->code.data);
> > +	vfree(ins->symbol_table.symbols);
> >  	kfree(ins);
> >  	return NULL;
> >  }
>
> --
> 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/



-- 
Ondrej Zary
--
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/