* Re: pc300too on a modern kernel? [not found] ` <1290442675.5515.92.camel@giskard.codewiz.org> @ 2010-11-22 21:20 ` Krzysztof Halasa 2010-11-23 14:44 ` Ward Vandewege 0 siblings, 1 reply; 2+ messages in thread From: Krzysztof Halasa @ 2010-11-22 21:20 UTC (permalink / raw) To: Bernie Innocenti; +Cc: Ward Vandewege, lkml, Jan Seiffert, netdev (added Cc: netdev) Bernie Innocenti <bernie@codewiz.org> writes: > Now the question is: why do we get so many spurious interrupts? Let me see... we call sca_tx_done() on (isr0 & 0x2020) which are DMIB3 and DMIB1, which in turn are (EOT & (EOTE = 0) | EOM & (EOME = 1)), i.e. the interrupt is generated on EOM (end of message = packet). It seems TN-PSC-339A/E is the answer: the interrupt is generated at the end of the last DMA access filling the TX buffer. Only then the status is written to the descriptor (=RAM). I guess it didn't make a difference on older, slower machines, with slower paths from PCI to CPU. Also I don't know if the descriptor status is being written in the same DMA transfer (between the chip and on-board SRAM) as the last data transfer. Perhaps it's another DMA request and arbitration, and perhaps the chip has to wait for another transfer to finish. > With this workaround applied, we're st seeing occasional clusters of > packet loss. We're working to graph the ping loss alongside traffic to > see if there's any correlation. That's interesting. I remember seeing some TX underruns at higher speeds, though nothing alike at 2 Mb/s. What bit rate are you using? Does "ifconfig hdlc0" show any errors? -- Krzysztof Halasa ^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: pc300too on a modern kernel? 2010-11-22 21:20 ` pc300too on a modern kernel? Krzysztof Halasa @ 2010-11-23 14:44 ` Ward Vandewege 0 siblings, 0 replies; 2+ messages in thread From: Ward Vandewege @ 2010-11-23 14:44 UTC (permalink / raw) To: Krzysztof Halasa; +Cc: Bernie Innocenti, lkml, Jan Seiffert, netdev On Mon, Nov 22, 2010 at 10:20:02PM +0100, Krzysztof Halasa wrote: > > With this workaround applied, we're st seeing occasional clusters of > > packet loss. We're working to graph the ping loss alongside traffic to > > see if there's any correlation. > > That's interesting. I remember seeing some TX underruns at higher > speeds, though nothing alike at 2 Mb/s. What bit rate are you using? > Does "ifconfig hdlc0" show any errors? This turned out to be caused by line saturation. We were not seeing this before we upgraded to the latest kernel because we have a set of qos rules that we forgot to install on our new box. Mea culpa... Thanks, Ward. -- Ward Vandewege <ward@gnu.org> ^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2010-11-23 14:44 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <20100902131531.GA19028@countzero.vandewege.net>
[not found] ` <m3mxrw1lg0.fsf@intrepid.localdomain>
[not found] ` <1289421869.9336.49.camel@giskard.codewiz.org>
[not found] ` <m3sjz34mka.fsf@intrepid.localdomain>
[not found] ` <1289944619.2677.22.camel@giskard.codewiz.org>
[not found] ` <m3tyjd7za9.fsf@intrepid.localdomain>
[not found] ` <1290442675.5515.92.camel@giskard.codewiz.org>
2010-11-22 21:20 ` pc300too on a modern kernel? Krzysztof Halasa
2010-11-23 14:44 ` Ward Vandewege
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).