cmd: bcb: Fix segfault on invalid block device

When blk_get_dev() fails, block NULL and gets de-referenced in
the error path by a printf(), resulting in a crash.

This can be reproduced on sandbox with:

$ ./u-boot --command "bcb load mmc 0 0"

Fix the message by using the functions arguments (iface, devnum)
instead.

Note: partition (being a global static initialized) can be used safely.

This issue has been reported by coverity [1]

[1] https://lore.kernel.org/all/20240723141844.GF989285@bill-the-cat/

Fixes: dfeb4f0d7935 ("cmd: bcb: extend BCB C API to allow read/write the fields")
Reviewed-by: Dmitrii Merkurev <dimorinny@google.com>
Link: https://lore.kernel.org/r/20240724-bcb-crash-v1-1-44caff15bce4@baylibre.com
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>

diff --git a/cmd/bcb.c b/cmd/bcb.c
index fe6d6cb2c3815b9c6980d10f3c1b10ef4a194033..97a96c009641cc094645607ef833575f3c03fe4b 100644 (file)
--- a/cmd/bcb.c
+++ b/cmd/bcb.c
@@ -172,8 +172,8 @@ static int __bcb_initialize(const char *iface, int devnum, const char *partp)
        return CMD_RET_SUCCESS;
 
 err_read_fail:
-       printf("Error: %d %d:%s read failed (%d)\n", block->uclass_id,
-              block->devnum, partition->name, ret);
+       printf("Error: %s %d:%s read failed (%d)\n", iface, devnum,
+              partition->name, ret);
        __bcb_reset();
        return CMD_RET_FAILURE;
 }