Date: Mon, 27 Oct 1997 04:33:51 -0800 From: Don Lewis <Don.Lewis@tsc.tdk.com> To: Matt Thomas <matt@3am-software.com>, Charles Henrich <henrich@crh.cl.msu.edu>, Jaye Mathisen <mrcpu@cdsnet.net> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: de0 errors Message-ID: <199710271233.EAA22692@salsa.gv.tsc.tdk.com> In-Reply-To: Matt Thomas <matt@3am-software.com> "Re: de0 errors" (Oct 23, 7:51am)
next in thread | previous in thread | raw e-mail | index | archive | help
On Oct 23, 7:51am, Matt Thomas wrote: } Subject: Re: de0 errors } At 12:57 AM 10/23/97 -0400, Charles Henrich wrote: } >After moving to 2.2.5-RELEASE im seeing: } > } >Oct 23 00:32:12 msunews /kernel: ccd0-3: Concatenated disk drivers } >Oct 23 00:35:45 msunews /kernel: de0: abnormal interrupt: transmit underflow } >(raising TX threshold to 96|256) } >Oct 23 00:36:01 msunews /kernel: de0: abnormal interrupt: transmit underflow } >(raising TX threshold to 8|512) } >Oct 23 00:37:41 msunews /kernel: de0: abnormal interrupt: transmit underflow } >(raising TX threshold to 1024) } >Oct 23 00:55:14 msunews /kernel: de0: abnormal interrupt: transmit underflow } >(switching to store-and-forward mode) } > } >Any ideas? Are these informational, or are they bad? } } they are informational. In the sense that everything is working, but if you're seeing transmit underflows even when the TX threshold has been set to 1024, then I suspect you'll have a real problem getting anywhere near full performance out of the card. A full size ethernet packet contains 1514 bytes of data (including the source and destination address but not the trailing CRC), and data is transmitted at 12.5MB/sec in in 100Mb mode. In this mode it takes about 82 microseconds to send the 1024 bytes that were loaded into the transmit FIFO before the packet transmission is started. If it can't load the remaining 490 bytes of data in those 82 microseconds, that means the data transfer rate across the PCI bus to the NIC must be less than 6 MB/sec. That's not very good for a 132 MB/sec bus ... Hmn, I'd also expect to see receive overflows if the data transfer rate is truly this bad and there is significant network activity. In the de driver this only seems to increment a private counter as well as the interface ierrors counter. It doesn't get logged like alignment and crc input errors do.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199710271233.EAA22692>