Re: [TRIVIAL] [patch, 2.4] drivers_sound_ad1848.c doesn't release region on error

[Rusty Russell]

In message <[e-mail]> you write
:
> On Mon, 2002-08-12 at 04:35, Rusty Russell wrote:
> > [ Simple, and looks trivially correct from reading file. ]
> > From:  Marcus Alanen <[e-mail]>
> >   
> >   Doesn't check request_region and doesn't release region on failed 
> >   kmalloc.
> 
> I don't have hardware to test this, so on the basis its not tested I'd
> say no until someone with the hardware tests it. 

Can someone with an ad1848 please test this trivial patch?  Preferably
test that loading the module the second time works fine?

Thanks!
Rusty.
--
  Anyone who quotes me in their sig is an idiot. -- Rusty Russell.

===== drivers/sound/ad1848.c 1.8 vs edited =====
--- 1.8/drivers/sound/ad1848.c	Fri May 10 01:40:52 2002
+++ edited/drivers/sound/ad1848.c	Sun Jul 14 18:21:20 2002
@@ -1997,7 +1997,8 @@
 		sprintf(dev_name,
 			"Generic audio codec (%s)", devc->chip_name);
 
-	request_region(devc->base, 4, devc->name);
+	if (!request_region(devc->base, 4, devc->name))
+		return -1;
 
 	conf_printf2(dev_name, devc->base, devc->irq, dma_playback, dma_capture);
 
@@ -2013,8 +2014,10 @@
 	}
 
 	portc = (ad1848_port_info *) kmalloc(sizeof(ad1848_port_info), GFP_KERNEL);
-	if(portc==NULL)
+	if(portc==NULL) {
+		release_region(devc->base, 4);
 		return -1;
+	}
 
 	if ((my_dev = sound_install_audiodrv(AUDIO_DRIVER_VERSION,
 					     dev_name,
@@ -2026,8 +2029,8 @@
 					     dma_playback,
 					     dma_capture)) < 0)
 	{
+		release_region(devc->base, 4);
 		kfree(portc);
-		portc=NULL;
 		return -1;
 	}