NULL pointer in imx-sdma.c with from context arg in reject_firmware_nowait

[Alexandre Oliva]

On Apr 17, 2018, Xavi Drudis Ferran <xdrudis at tinet.cat> wrote:

> I think in previous versions, sdma_load_firmware didn't even get
> called.

Yeah.  I changed it in 4.16-gnu.  Sorry I missed your report and patch,
I wish I'd seen this in time for 4.17-gnu.  But, oh well... :-(

These patches to the affected deblobbing scripts should fix them,
producing the change to firmware.h that you posted.

I'll make sure the patch makes 4.18.  How bad is it for the affected
releases?  As in, should I make it a priority to backport it to stable
updates?

Index: deblob-4.16
===================================================================
--- deblob-4.16	(revision 14983)
+++ deblob-4.16	(working copy)
@@ -501,7 +501,7 @@
 	report_missing_free_firmware(dev_name(device), NULL);\
 	/* We assume NONFREE_FIRMWARE will not be found; how could it?  */\
 	return request_firmware_nowait(module, uevent, NONFREE_FIRMWARE,\
-				       device, gfp, NULL, cont);\
+				       device, gfp, context, cont);\
 }\
 static inline int\
 maybe_reject_firmware_nowait(struct module *module, int uevent,\

Index: deblob-4.17
===================================================================
--- deblob-4.17	(revision 15022)
+++ deblob-4.17	(working copy)
@@ -505,7 +505,7 @@
 	report_missing_free_firmware(dev_name(device), NULL);\
 	/* We assume NONFREE_FIRMWARE will not be found; how could it?  */\
 	return request_firmware_nowait(module, uevent, NONFREE_FIRMWARE,\
-				       device, gfp, NULL, cont);\
+				       device, gfp, context, cont);\
 }\
 static inline int\
 maybe_reject_firmware_nowait(struct module *module, int uevent,\



-- 
Alexandre Oliva, freedom fighter   https://FSFLA.org/blogs/lxo
Be the change, be Free!         FSF Latin America board member
GNU Toolchain Engineer                Free Software Evangelist