"Hiroshi Inoue" <Inoue(at)tpf(dot)co(dot)jp> writes:
>> (at least since 6.3): bufmgr.c expects that I/O errors will result in
>> an SM_FAIL return code from the smgr.c routines, but smgr.c does no
>> such thing: it does elog(ERROR) if it sees a failure. All of the
> except smgropen().
Right. I'm mainly looking at the block read/write/flush calls,
which have a lot of now-useless error recovery code after them.
> I also prefer the latter. Even though smgr returns SM_FAIL,md stuff
> already calls elog(ERROR) in many places.
Good point, and the fd.c level may have some elogs too...
regards, tom lane