Alvaro Herrera <alvherre@commandprompt.com> writes:
> Tom Lane escribi�:
>> Good joke, but to be serious: we expect that forward scans will result
>> in the kernel doing read-ahead, which will allow overlapping of
>> CPU work to process one page with the I/O to bring in the next page.
> I wonder if this is spoiled (or rather, the backwards case fixed) by the
> attempts to call posix_fadvise() on certain types of scan.
Yeah, I started wondering about that too after sending off the above.
The fadvise patch might eliminate the distinction ... on platforms where
fadvise exists and actually works well.
regards, tom lane