On 20. 4. 8. 오전 12:55, Sami Tolvanen wrote: > On Mon, Apr 6, 2020 at 11:54 PM Sunwook Eom <speed.eom@xxxxxxxxxxx> wrote: >> Even if block type is metadata, >> block in verity_fec_decode() has already the right block number. >> So there is no need to add data_blocks to block. > Is this also true if the hashes are on a separate block device? > > The idea here is that the error correction data was computed over both > data and hash blocks, as if they were concatenated, and we want to > calculate the logical block number based on that. I agree that the > code doesn't look quite right though. Should we use something like > this instead? > > if (type == DM_VERITY_BLOCK_TYPE_METADATA) > block = block - v->hash_start + v->data_blocks; > > Sami > > You're right. I missed the case that hashes are on a separate block device. And, the code you wrote seems to be correct. If you don't mind, I'll send a new version of this patch. Thank you for the review. Sunwook -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
