Tom Lane <t...@sss.pgh.pa.us> wrote: > Antonin Houska <a...@cybertec.at> writes: > > Tom Lane <t...@sss.pgh.pa.us> wrote: > >> At one time, I think, readLen told how much data in readBuf was > >> actually valid. It seems not to be used for that anymore, but > >> I don't much like the idea that readBuf is only partially filled > >> but there is *no* persistent state indicating how much is valid. > >> The existing coding guarantees that the answer is "XLOG_BLCKSZ", > >> so that's fine, but this change would remove the guarantee. > > Oh, I withdraw that complaint --- readBuf isn't a persistent data > structure anymore, so there's no need for readLen to be persistent > either.
Well, there's yet a problem. Your concern about invalid data in readBuf reminded me of a problem that I reported a few years ago [1]. The problem is hidden as long as all XLOG reader callbacks fetch the whole page. However if readLen is used as an argument of read() in XLogPageRead() *and* if some other conditions that also hide the issue are removed (see [2]), the problem reported in [1] appears to be a live bug. [1] https://www.postgresql.org/message-id/2363.1428573952%40localhost [2] https://www.postgresql.org/message-id/32276.1516290849%40localhost -- Antonin Houska Cybertec Schönig & Schönig GmbH Gröhrmühlgasse 26, A-2700 Wiener Neustadt Web: https://www.cybertec-postgresql.com