Message ID | 20150428172920.30e53b06@kryten |
---|---|
State | Accepted |
Headers | show |
Anton Blanchard <anton@samba.org> writes: > - prerror("SURV: Failed to allocate hbeat msg\n"); > + prerror("SURV: Failed to allocate heartbeat msg\n"); > return; > } > if (fsp_queue_msg(msg, fsp_surv_ack)) { > fsp_freemsg(msg); > - prerror("SURV: Failed to queue hbeat msg\n"); > + prerror("SURV: Failed to queue heartbeat msg\n"); > } else { Are you getting a lot of these two? That would kind of concern me as you shouldn't ordinarily get these. Not on a larger P8 system by any chance?
Hi Stewart, > Are you getting a lot of these two? That would kind of concern me as > you shouldn't ordinarily get these. > > Not on a larger P8 system by any chance? I'm not, I'm just getting these: [20405211202285,7] SURV: Sending the hearbeat command to FSP [20405220455009,7] SURV: Received heartbeat acknowledge from FSP [20436391277825,7] SURV: Sending the hearbeat command to FSP [20436400534449,7] SURV: Received heartbeat acknowledge from FSP But I figured I'd clean all the log messages in the area. Anton
On Wed, Apr 29, 2015 at 09:38:23AM +1000, Stewart Smith wrote: > Anton Blanchard <anton@samba.org> writes: > > - prerror("SURV: Failed to allocate hbeat msg\n"); > > + prerror("SURV: Failed to allocate heartbeat msg\n"); > > return; > > } > > if (fsp_queue_msg(msg, fsp_surv_ack)) { > > fsp_freemsg(msg); > > - prerror("SURV: Failed to queue hbeat msg\n"); > > + prerror("SURV: Failed to queue heartbeat msg\n"); > > } else { > > Are you getting a lot of these two? That would kind of concern me as you > shouldn't ordinarily get these. > > Not on a larger P8 system by any chance? I guess Anton's using 810.x? The less chatty version is only in the 830 stream. Ananth
Anton Blanchard <anton@samba.org> writes: > I'm not, I'm just getting these: Phew. I was worried for a minute :) > [20405211202285,7] SURV: Sending the hearbeat command to FSP > [20405220455009,7] SURV: Received heartbeat acknowledge from FSP > [20436391277825,7] SURV: Sending the hearbeat command to FSP > [20436400534449,7] SURV: Received heartbeat acknowledge from FSP I'm kind of keen to get rid of these, instead maybe having something in a well known data structure and maybe a watchdog... While useful for seeing if *everything* has wedged, they're often enough that all the rest of the skiboot log vanishes if you don't reboot your system as often as firmware developers do.
diff --git a/hw/fsp/fsp-surveillance.c b/hw/fsp/fsp-surveillance.c index 138ff3f..b94eb1b 100644 --- a/hw/fsp/fsp-surveillance.c +++ b/hw/fsp/fsp-surveillance.c @@ -112,15 +112,15 @@ static void fsp_surv_hbeat(void) (tb_compare(now, surv_timer) == TB_AAFTERB) || (tb_compare(now, surv_timer) == TB_AEQUALB)) { prlog(PR_DEBUG, - "SURV: Sending the hearbeat command to FSP\n"); + "SURV: Sending the heartbeat command to FSP\n"); msg = fsp_mkmsg(FSP_CMD_SURV_HBEAT, 1, 120); if (!msg) { - prerror("SURV: Failed to allocate hbeat msg\n"); + prerror("SURV: Failed to allocate heartbeat msg\n"); return; } if (fsp_queue_msg(msg, fsp_surv_ack)) { fsp_freemsg(msg); - prerror("SURV: Failed to queue hbeat msg\n"); + prerror("SURV: Failed to queue heartbeat msg\n"); } else { fsp_surv_ack_pending = true; surv_timer = now + secs_to_tb(60);
I get thousands of these in my OPAL message log so they may as well be spelled correctly. While here, make the other messages consistent by changing hbeat to heartbeat. Signed-off-by: Anton Blanchard <anton@samba.org> ---