Re: memory leak in amdisplay_attach()

[Claudio Jeker]

On Tue, Sep 18, 2018 at 03:49:15PM +1000, Jonathan Gray wrote:
> Index: amdisplay.c
> ===================================================================
> RCS file: /cvs/src/sys/arch/armv7/omap/amdisplay.c,v
> retrieving revision 1.7
> diff -u -p -r1.7 amdisplay.c
> --- amdisplay.c       25 Oct 2017 14:34:22 -0000      1.7
> +++ amdisplay.c       18 Sep 2018 05:12:41 -0000
> @@ -272,6 +272,7 @@ amdisplay_attach(struct device *parent, 
>  
>       if (rasops_init(&sc->sc_ro, 200, 200)) {
>               printf("%s: no rasops\n", DEVNAME(sc));
> +             free(edid_buf, M_DEVBUF, EDID_LENGTH);
>               amdisplay_detach(self, 0);
>               return;
>       }
> 

I think it is better to free the edid_buf further up in that function
since it is unused after calling edid_parse(edid_buf, &sc->sc_edid) on
line 215. So currently there is still a leak at the end of the function.

-- 
:wq Claudio

Index: arch/armv7/omap/amdisplay.c
===================================================================
RCS file: /cvs/src/sys/arch/armv7/omap/amdisplay.c,v
retrieving revision 1.7
diff -u -p -r1.7 amdisplay.c
--- arch/armv7/omap/amdisplay.c 25 Oct 2017 14:34:22 -0000      1.7
+++ arch/armv7/omap/amdisplay.c 18 Sep 2018 06:32:43 -0000
@@ -219,6 +219,8 @@ amdisplay_attach(struct device *parent, 
                return;
        }
 
+       free(edid_buf, M_DEVBUF, EDID_LENGTH);
+
 #ifdef LCD_DEBUG
        edid_print(&sc->sc_edid);
 #endif
@@ -246,7 +248,6 @@ amdisplay_attach(struct device *parent, 
        /* configure DMA framebuffer */
        if (amdisplay_setup_dma(sc)) {
                printf("%s: couldn't allocate DMA framebuffer\n", DEVNAME(sc));
-               free(edid_buf, M_DEVBUF, EDID_LENGTH);
                amdisplay_detach(self, 0);
                return;
        }