On 04.12.2016 21:33, Joe Perches wrote:
>>> diff --git a/drivers/mtd/ubi/io.c b/drivers/mtd/ubi/io.c
> []
>>> @@ -1413,7 +1413,7 @@ int ubi_self_check_all_ff(struct ubi_device *ubi, int
>>> pnum, int offset, int len)
>>> buf = __vmalloc(len, GFP_NOFS, PAGE_KERNEL);
>>> if (!buf) {
>>> ubi_err(ubi, "cannot allocate memory to check for 0xFFs");
>>> - return 0;
>>> + return -ENOMEM;
>>
>> I wonder if you shouldn't also nuke the ubi_err() , because when you run
>> out of memory, printk() will likely also fail.
>
> No, not really. printk doesn't allocate memory.
>
> But the ubi_err should be removed because all memory
> allocations that fail without a specific GFP_NOWARN
> flag already have a dump_stack() call.
We should better think about how to get ubi_self_check_all_ff() fixed.
When enabled on a modern NAND, vmalloc() is likely to fail now and then
since len is the erase block size and can be up to a few mega bytes.
Thanks,
//richard
