Messages in this thread | | | Date | Mon, 2 Aug 1999 11:25:48 +0200 (CEST) | From | Andrea Arcangeli <> | Subject | Re: Latency: Mingo's lowlatency patch still not perfect, but looking good |
| |
On Mon, 2 Aug 1999 yodaiken@chelm.cs.nmt.edu wrote:
>> Note that the write performance of writing a large file decreases by about 30% >> using Mingo's patch, and looking at the harddisk led, the led flashes
Try removing the additional run_task_queue() I was complaining about yesterday. They won't change the latency results at all.
@@ -1750,11 +1768,12 @@ { ndirty = 0; repeat: - bh = lru_list[nlist]; if(bh) for (i = nr_buffers_type[nlist]; i-- > 0 && ndirty < bdf_prm.b_un.ndirty; bh = next) { + conditional_schedule(); + run_task_queue(&tq_disk); /* We may have stalled while waiting for I/O to complete. */ if(bh->b_list != nlist) goto repeat; next = bh->b_next_free;
Also the above snapshot from Ingo's -N2 patch is buggy and may crash the kernel if the bh buffer gets released while sleeping. The right diff written by hand looks like this:
for(...) + bh->b_count++; + conditional_schedule() + bh->b_count--; /* We may have stalle...
Andrea
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |