On Sat, Jan 03, 2009 at 10:39:57AM +0100, Gerhard Wiesinger wrote: > Hello! > > The ethernet driver performance especially on TX is very limited and > takes 100% system resources. For details see below. > > Source can be found at: > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/linux/brcm/linux.v24_2/drivers/net/et > > Especially the following files are relevant: > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/linux/brcm/linux.v24_2/drivers/net/et/etc.c > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/linux/brcm/linux.v24_2/drivers/net/et/et_linux.c > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/linux/brcm/linux.v24_2/drivers/net/et/Makefile > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/include.v24/hnddma.h > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/shared/hnddma.c > > I think the problem comes from: > http://svn.dd-wrt.com:8000/dd-wrt/browser/src/linux/brcm/linux.v24_2/drivers/net/et/et_linux.c#L422 > /* > * Yeah, queueing the packets on a tx queue instead of throwing them > * directly into the descriptor ring in the case of dma is kinda lame, > * but this results in a unified transmit path for both dma and pio > * and localizes/simplifies the netif_*_queue semantics, too. > */ > > Can someone help me to improve the driver performance especially on TX? Good luck. IMHO, a better option would be to move to a 2.6-based kernel and to use the b43 and b44 drivers. John -- John W. Linville Linux should be at the core linville@xxxxxxxxxxxxx of your literate lifestyle. -- To unsubscribe from this list: send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
