Re: [PATCH] staging: gpib: Move free after the variable use has been completed

[Everest K.C.]

On Tue, Oct 15, 2024 at 4:55 PM Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On 10/15/24 16:48, Everest K.C. wrote:
> > On Tue, Oct 15, 2024 at 4:35 PM Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> wrote:
> >>
> >> On 10/15/24 15:51, Everest K.C. wrote:
> >>> The variable `in_data` is freed, but used later in the code.
> >>> Fix it by moving the freeing the memory after it use has been
> >>> completed.
> >>>
> >>> This issue was reported by Coverity Scan.
> >>> Report:
> >>> CID 1600783: (#1 of 1): Use after free (USE_AFTER_FREE)
> >>> 19. pass_freed_arg: Passing freed pointer in_data as an argument to
> >>> ni_usb_dump_raw_block.
> >>>
> >>> Fixes: 4e127de14fa7 ("staging: gpib: Add National Instruments USB GPIB driver")
> >>> Signed-off-by: Everest K.C. <everestkc@xxxxxxxxxxxxxxxx>
> >>> ---
> >>>    drivers/staging/gpib/ni_usb/ni_usb_gpib.c | 2 +-
> >>>    1 file changed, 1 insertion(+), 1 deletion(-)
> >>>
> >>> diff --git a/drivers/staging/gpib/ni_usb/ni_usb_gpib.c b/drivers/staging/gpib/ni_usb/ni_usb_gpib.c
> >>> index 1da263676f2a..75f39e1f3ed1 100644
> >>> --- a/drivers/staging/gpib/ni_usb/ni_usb_gpib.c
> >>> +++ b/drivers/staging/gpib/ni_usb/ni_usb_gpib.c
> >>> @@ -690,12 +690,12 @@ static int ni_usb_read(gpib_board_t *board, uint8_t *buffer, size_t length,
> >>>                kfree(in_data);
> >>>                return parse_retval;
> >>>        }
> >>> -     kfree(in_data);
> >>>        if (actual_length != length - status.count) {
> >>>                pr_err("%s: actual_length=%i expected=%li\n",
> >>>                       __func__, actual_length, (long)(length - status.count));
> >>>                ni_usb_dump_raw_block(in_data, usb_bytes_read);
> >>>        }
> >>> +     kfree(in_data);
> >>>        switch (status.error_code) {
> >>>        case NIUSB_NO_ERROR:
> >>>                retval = 0;
> >>
> >> Looks good to me. Isn't this on next though. Don't forget to
> >> indicate it is against next.
> > No, it was fixed in the linux-staging repo.
> >> Reviewed-by: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx>
> >>
>
> Okay - by the way the same problem is in ni_usb_write_registers().
> Did coverity catch that one?
No, there were no similar issues caught by coverity.
Also, I just checked the function. I didn't see the variable being
accessed after
it had been freed.
> Both problems could be fix in one patch - I will leave it up to the
> maintainers to make a call on two patches or a single patch.
>
> thanks,
> -- Shuah
>