@@ -2876,6 +2876,7 @@ static void mmc_blk_remove_debugfs(struct mmc_card *card,
static int mmc_blk_probe(struct mmc_card *card)
{
struct mmc_blk_data *md, *part_md;
+ int ret = 0;
/*
* Check that the card supports the command class(es) we need.
@@ -2893,19 +2894,24 @@ static int mmc_blk_probe(struct mmc_card *card)
}
md = mmc_blk_alloc(card);
- if (IS_ERR(md))
- return PTR_ERR(md);
+ if (IS_ERR(md)) {
+ ret = PTR_ERR(md);
+ goto out_free;
+ }
- if (mmc_blk_alloc_parts(card, md))
+ ret = mmc_blk_alloc_parts(card, md);
+ if (ret)
goto out;
dev_set_drvdata(&card->dev, md);
- if (mmc_add_disk(md))
+ ret = mmc_add_disk(md);
+ if (ret)
goto out;
list_for_each_entry(part_md, &md->part, part) {
- if (mmc_add_disk(part_md))
+ ret = mmc_add_disk(part_md);
+ if (ret)
goto out;
}
@@ -2926,10 +2932,12 @@ static int mmc_blk_probe(struct mmc_card *card)
return 0;
- out:
+out:
mmc_blk_remove_parts(card, md);
mmc_blk_remove_req(md);
- return 0;
+out_free:
+ destroy_workqueue(card->complete_wq);
+ return ret;
}
static void mmc_blk_remove(struct mmc_card *card)
Returning zero to indicate success, when we actually have failed to probe is wrong. As a matter of fact, it leads to that mmc_blk_remove() gets called at a card removal and then triggers "NULL pointer dereference" splats. This is because mmc_blk_remove() relies on data structures and pointers to be setup from mmc_blk_probe(), of course. There have been no errors reported about this, which is most likely because mmc_blk_probe() never fails like this. Nevertheless, let's fix the code by propagating the error codes correctly and prevent us from leaking memory by calling also destroy_workqueue() in the error path. Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> --- drivers/mmc/core/block.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-) -- 2.25.1