Date: Fri, 20 Dec 2002 22:35:57 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: Thomas Nystrom <thn@saeab.se> Cc: Mike Tancsa <mike@sentex.net>, stable@FreeBSD.ORG Subject: Re: Fix for hanging of vr interface (Rhine Ethernet) Message-ID: <200212210635.gBL6Zvc5035909@apollo.backplane.com> References: <5.2.0.9.0.20021220151835.03a72a48@marble.sentex.ca> <200212202229.gBKMTREh022639@apollo.backplane.com> <3E03B1C6.83A9671C@saeab.se>
next in thread | previous in thread | raw e-mail | index | archive | help
:I think they are 2K bytes each (don't have the datasheets with me). That
:message is printed out because the Rhine chip signals that a packet is
:dropped. On my system I could tweak it away by changing the RX-Threshold
:and the DMA block size. My theory is that the bus is busy making it
:impossible for the chip to store the received frame in memory.
:
:/thn
:
:--
:---------------------------------------------------------------
:Svensk Aktuell Elektronik AB Thomas Nyström
I messed around a bit with the DMA blocking and I also tried
enabling some bits in another Config register that turned on
receive fifo / transmit fifo interleaving. Nothing changes,
I still get 'rx packet lost' messages.
I have noticed that video takes appreciable ram bandwidth on
this EPIA5000. In 1024x768 @ 70KHz / 24bpp (32bpp frame buffer),
a busy cpu can glitch the screen.
I think the EPIA BIOS may not be setting up PCI latency timers
properly in my case. I'll bet workstations with faster memory
busses do better. There may not be anything that can be done
in the VR driver.
I'm still watching for long dropouts. So far none have occured.
-Matt
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200212210635.gBL6Zvc5035909>