Date: Mon, 24 Nov 2008 13:58:46 +0900 From: Pyun YongHyeon <pyunyh@gmail.com> To: albri <galbrig@googlemail.com> Cc: freebsd-current@freebsd.org Subject: Re: Call for testers: Atheros AR8121(L1E)/AR8113/AR8114(L2E) ethernet Message-ID: <20081124045846.GM78954@cdnetworks.co.kr> In-Reply-To: <46d45f030811171014i2ae5df78mbbebc367ef2ca7d4@mail.gmail.com> References: <46d45f030811160642m2dff1481g457f1fa1a4ac1372@mail.gmail.com> <20081117010558.GD50872@cdnetworks.co.kr> <46d45f030811171014i2ae5df78mbbebc367ef2ca7d4@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Nov 17, 2008 at 07:14:12PM +0100, albri wrote: > hello, > > On 11/17/08, Pyun YongHyeon <pyunyh@gmail.com> wrote: > > On Sun, Nov 16, 2008 at 03:42:16PM +0100, albri wrote: > > > hello, > > > > > > I have issues transfering big or many files over ethernet on 1000H, also. > > > Using yongari's ale(4) driver > > > http://people.freebsd.org/~yongari/ale/ale.20081114.tar.gz I did not > > > have any problems while I surf or download 37MB sourcecode from inet. > > > Then scp(1)'ing source from 1000H to desktop PC showed a transfer rate > > > with maximum 86kB/s regardless to which direction is copied. The > > > ethernet NIC, while copying, is switched off regularily then. No > > > copies possible after three megabytes. > > > > Try turning off TSO and let me know how it goes. > > ("#ifconfig ale0 -tso" will do the job.) > > > > this helps a little bit with two effects. > Copying source-tree with scp(1): > Now I can see transfers with up to approx. 500kB/s - inaccurate > measured with scp(1), > but relation counts. > Transfer stalles after different data volumes with DMA-error on tty0, > but networking > port is not turned off. This happens after 30-80MB data transfers. > You can restart the whole copy at once again. > > Copying source-tree with nc(1) tar-gzipped: > Transfer stops and starts with DMA-error every approx. 7MB with turning off NIC. I still can't reproduce this and I have no idea how to solve it even if I can reproduce that on my box. :-( There could be a wrong in DMA configuration or some mis-programmed registers but I still see no errors in these area. [...] > > > > > > - ifconfig shows: > > > 1 ale0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu > > 1500 > > > 2 options=319b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,TSO4> > > > > It seems that your hardware doesn't have WOL capability. But I > > guess EeePC 1000H does support WOL. You've disabled some > > functionality of ethernet controller in BIOS? > > > Ethernet controller is turned on in BIOS, but booting via ROM is turned off. > Maybe did I not compile WOL to my (nearly) monolithic kernel? And if > so, I do not know the switch. Maybe you're not using latest stable/7. [...] > Did you remember my second post?: > > >I want to make one addition: > >doing a > ># sockstat -46l > >I get following error: > >sockstat: struct xtcpcb size mismatch > >sockstat: struct xtcpcb size mismatch > Looks like your kernel is out of sync with world. > Is there a relation to DMA-error? > It may not be related with DMA error. ale(4) was MFCed to stable/7 so it would be better to rebuild your box with latest stable/7. -- Regards, Pyun YongHyeon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20081124045846.GM78954>