On Mon, 29 Jul 2019, Jia-Ju Bai wrote:
> In sddr55_transport(), there is an if statement on line 836 to check
> whether info->lba_to_pba is NULL:
> if (info->lba_to_pba == NULL || ...)
>
> When info->lba_to_pba is NULL, it is used on line 948:
> pba = info->lba_to_pba[lba];
>
> Thus, a possible null-pointer dereference may occur.
>
> To fix this bug, info->lba_to_pba is checked before being used.
>
> This bug is found by a static analysis tool STCheck written by us.
This is not the right way to fix the bug.
The code already contains a test on line 938:
if (lba >= info->max_log_blks) {
If this test fails, the driver doesn't try to dereference
info->lba_to_pba.
The problem is that info->max_log_blks may be set even though
info->lba_to_pba is NULL, because the READ_CAPACITY case in
sddr55_transport() doesn't check the return code from
sddr55_read_map(). _That_ is what needs to be fixed.
Alan Stern
> Signed-off-by: Jia-Ju Bai <[email protected]>
> ---
> v2:
> * Avoid uninitialized access of pba.
> Thank Oliver for helpful advice.
>
> ---
> drivers/usb/storage/sddr55.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/usb/storage/sddr55.c b/drivers/usb/storage/sddr55.c
> index b8527c55335b..d23aff16091e 100644
> --- a/drivers/usb/storage/sddr55.c
> +++ b/drivers/usb/storage/sddr55.c
> @@ -945,7 +945,7 @@ static int sddr55_transport(struct scsi_cmnd *srb, struct
> us_data *us)
> return USB_STOR_TRANSPORT_FAILED;
> }
>
> - pba = info->lba_to_pba[lba];
> + pba = info->lba_to_pba ? info->lba_to_pba[lba] : 0;
>
> if (srb->cmnd[0] == WRITE_10) {
> usb_stor_dbg(us, "WRITE_10: write block %04X (LBA %04X)
> page %01X pages %d\n",
>
