Patchwork mtd_concat + m25p80 fails

login
register
mail settings
Submitter Andreas Fenkart
Date Feb. 3, 2009, 8:52 p.m.
Message ID <20090203205245.GB29085@CCCP.sue.net>
Download mbox | patch
Permalink /patch/21791/
State New
Headers show

Comments

Andreas Fenkart - Feb. 3, 2009, 8:52 p.m.
Hi all

I successfully merged two m25p80 flash devices into a single 'virtual' device. Thanks for your efforts. 

There is one problem though, when running flash_erase. Actually at the end of it, the kernel crashes in concat_sync. Since m25p80 does not define the 'sync' method concat_sync crashes while dereferencing the NULL pointer. 

static void concat_sync(struct mtd_info *mtd)
{ 
	struct mtd_concat *concat = CONCAT(mtd);
	int i;

	for (i = 0; i < concat->num_subdev; i++) {
		struct mtd_info *subdev = concat->subdev[i];
		subdev->sync(subdev);
	}
}
               

My question is, if the following the 'hack' is valid or whether I run into troubles, by not defining sync/lock/unlock when the subdevices do not define them.


Andi

Patch

diff --git a/drivers/mtd/mtdconcat.c b/drivers/mtd/mtdconcat.c
index d563dcd..8c2384f 100644
--- a/drivers/mtd/mtdconcat.c
+++ b/drivers/mtd/mtdconcat.c
@@ -793,11 +793,16 @@  struct mtd_info *mtd_concat_create(struct mtd_info *subdev[],     /* subdevices to c
        concat->mtd.erase = concat_erase;
        concat->mtd.read = concat_read;
        concat->mtd.write = concat_write;
-       concat->mtd.sync = concat_sync;
+       if (subdev[0]->sync)
+               concat->mtd.sync = concat_sync;
+       if (subdev[0]->lock)
        concat->mtd.lock = concat_lock;
+       if (subdev[0]->unlock)
        concat->mtd.unlock = concat_unlock;
-       concat->mtd.suspend = concat_suspend;
-       concat->mtd.resume = concat_resume;
+       if (subdev[0]->suspend)
+               concat->mtd.suspend = concat_suspend;
+       if (subdev[0]->resume)
+               concat->mtd.resume = concat_resume;
 

thank you