Pascal CHAPPERON <[EMAIL PROTECTED]> : [...] > That one works perfectly; i tried it in the same conditions as > the previous patch, and i don't notice a regression.
Nice. [...] > sis190-010.patch does not compile properly : Point taken. Thanks. [...] > But i get those traces in syslog, when the driver is > loaded and every time i use ethtool : > > [...] > scheduling while atomic: mii-tool/0x00000001/4699 spinlock + msleep in __mdio_cmd()... Completely stupid. The patchkit of the day should fix these issues: http://www.zoreil.com/~romieu/sis190/20050706-2.6.13-rc1/patches Tarball: http://www.zoreil.com/~romieu/sis190/20050706-2.6.13-rc1.tar.bz2 Please shake it. You should be able to modify the set of displayed messages with ethtool. I'd appreciate if you could check the allowed frame size range, say ping -s 1468 ... ping -s 1473 to compare with Lars's results. You can add something like a ping -q -l 48 -s 64 -f to your tests and increase the 48 and NUM_{RX/TX}_DESC but I am not sure that the remote 8139 will be able to go terribly far. If it performs well, pktgen could be useful too. Lars, can you describe the second host that you have used so far for the tests: - which nic ? - which kernel ? (on an unrelated note, something enabled the "send mail as html" checkbox in your mail user agent: you may consider removing it). -- Ueimor - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/