Date: Sun, 30 May 2004 02:49:14 +0200 From: Marius Strobl <marius@alchemy.franken.de> To: freebsd-sparc64@freebsd.org Subject: hme(4) issues Message-ID: <20040530024914.A70602@newtrinity.zeist.de>
next in thread | raw e-mail | index | archive | help
According to the reports sent to this list from time to time there are two issues with hme(4) on FreeBSD/sparc64: 1) Autonegotiation mode fails to establish 100baseTX full-duplex links. 2) When using 100baseTX full-duplex links (by forcing it because of the former but maybe also otherwise?) bind(2) fails with EADDRNOTAVAIL when trying to bind to an IP address of a hme(4) but INADDR_ANY works. I hit 1) myself and tracked it down to a problem with the DP83840A PHY (nsphy) some hme(4) use. This combination is used on e.g. Ultra 2, Ultra AXi and AFAIK also on Ultra 5 and 10. The problem should be fixed with sys/dev/mii/nsphy.c revision 1.22 and sys/dev/hme/if_hme.c revision 1.23. The hme(4) using LU6612 (ukphy; used on Ultra AXe and Netra T1) or QS6612 (qsphy; used on Ultra 60) appear to be unaffected by 1). As for 2) I didn't manage to reproduce this and I have no idea what could cause it. Could those of you who experience this please update to the latest -current containing the above mentioned revisions, remove whatever workaround you have (e.g. forcing half-duplex) and test if it still is an issue? Does anyone see this also with other NICs (e.g. dc(4), fxp(4), gem(4) or xl(4)) or in another situation than described above (e.g. also in half-duplex mode)?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040530024914.A70602>