Date: Sun, 19 Dec 2004 22:55:18 +0100 From: "Heinz Knocke" <knockefreebsd@o2.pl> To: "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net> Cc: freebsd-net@freebsd.org Subject: Re: Marvell 88E8001 on sk0 and RELENG_5_3 - big problems Message-ID: <004e01c4e615$70234d80$df5561d9@ALFA> References: <003501c4e228$5f2cd780$df5561d9@ALFA> <Pine.BSF.4.53.0412142202360.931@e0-0.zab2.int.zabbadoz.net> <006301c4e2f4$26a67fc0$df5561d9@ALFA> <Pine.BSF.4.53.0412161343000.931@e0-0.zab2.int.zabbadoz.net>
next in thread | previous in thread | raw e-mail | index | archive | help
----- Original Message -----=20 From: "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net> To: "Heinz Knocke" <knockefreebsd@o2.pl> Cc: <freebsd-stable@freebsd.org> Sent: Thursday, December 16, 2004 2:45 PM Subject: Re: Marvell 88E8001 on sk0 and RELENG_5_3 - big problems > would you mind trying this patch. It should only activate jumbo frames > conditional on interface MTU like we do for xmacii. >=20 >=20 > Index: sys/pci/if_sk.c > (...) I tried this one and seems to work corectly :) Jumbograms were used, = data stream was smooth,=20 I haven't seen any error messages in system log.=20 In rough tests I had a little bit better speeds than with this one:=20 http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/pci/if_sk.c.diff?r1=3D1= .51&r2=3D1.52 But I only tried netperf with maximal message send size on similar = boxes, if you'd like some other tests,=20 let me know the details.=20 By the way - do you know if sk(4) driver supports device polling (man 4 = polling)? I know there were some plans to make it work, but I couldn't found any up2date information.=20 hk
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?004e01c4e615$70234d80$df5561d9>