Re: buffercache/bgwriter - Mailing list pgsql-performance

Hi Brad,

yes. that's the question....
in the source code in freelist.c there is something that I don't understand.

This is the first try to get a free page. The second try scans used buffers.
What makes me wonder is the why postgres is checking for <<buf->usage_count == 0>>
where usage_count is supposed to be NULL initially.

    while (StrategyControl->firstFreeBuffer >= 0)
    {
        buf = &BufferDescriptors[StrategyControl->firstFreeBuffer];
        Assert(buf->freeNext != FREENEXT_NOT_IN_LIST);

        /* Unconditionally remove buffer from freelist */
        StrategyControl->firstFreeBuffer = buf->freeNext;
        buf->freeNext = FREENEXT_NOT_IN_LIST;

        /*
         * If the buffer is pinned or has a nonzero usage_count, we cannot use
         * it; discard it and retry.  (This can only happen if VACUUM put a
         * valid buffer in the freelist and then someone else used it before
         * we got to it.  It's probably impossible altogether as of 8.3, but
         * we'd better check anyway.)
         */
        LockBufHdr(buf);
        if (buf->refcount == 0 && buf->usage_count == 0)
        {
            if (strategy != NULL)
                AddBufferToRing(strategy, buf);
            return buf;
        }
        UnlockBufHdr(buf);
    }


Best...
Uwe