Messages in this thread | | | Subject | Re: [RFC, PATCH] Extensible AIO interface | From | "Martin K. Petersen" <> | Date | Tue, 02 Oct 2012 18:34:05 -0400 |
| |
>>>>> "Kent" == Kent Overstreet <koverstreet@google.com> writes:
>> Hmm, careful here. I think that in DIF/DIX the checksums are >> per-sector, not per IO, right? That'd mean that the PAGE_SIZE attr >> limit in this patch would be magically creating different max IO size >> limits on different architectures. That doesn't seem great.
Kent> Not just per sector, Per hardware sector.
Per logical block (or for some devices less).
Kent> For passing around checksums userspace would have to find out the Kent> hardware sector size and checksum type/size via a different Kent> interface,
The relevant information is already exported in sysfs. Including the format, how many bytes of integrity metadata go with how many bytes of data, etc.
-- Martin K. Petersen Oracle Linux Engineering
| |