On Tue, Jul 22, 2025 at 01:45:53AM -0700, Christoph Hellwig wrote:
> On Tue, Jul 22, 2025 at 01:49:11PM +0530, Anuj Gupta wrote:
> > + switch (bi->csum_type) {
> > + case BLK_INTEGRITY_CSUM_NONE:
> > + meta_cap.lbmd_guard_tag_type = LBMD_PI_CSUM_NONE;
> > + break;
> > + case BLK_INTEGRITY_CSUM_IP:
> > + meta_cap.lbmd_guard_tag_type = LBMD_PI_CSUM_IP;
> > + break;
> > + case BLK_INTEGRITY_CSUM_CRC:
> > + meta_cap.lbmd_guard_tag_type = LBMD_PI_CSUM_CRC16_T10DIF;
> > + break;
> > + case BLK_INTEGRITY_CSUM_CRC64:
> > + meta_cap.lbmd_guard_tag_type = LBMD_PI_CSUM_CRC64_NVME;
> > + break;
> > + default:
> > + break;
>
> This should catch and reject invalid values. Otherwise the patch looks
> fine.
I guess reject is wrong, as the kernel is the source, but I think
a bit of validation here still makes sense. I think if you just leave
the default out, the compiler would complain if we added a new value
to the union without handling it here, which should be enough.
