<br><font size=2 face="sans-serif"> Please don't be confused by the
numbers in the 4M column, they don't mean anything to the end user. We
had a bit better performance (approx. 720 Mbps) at one time but at 100%
CPU usage. These new patches lower the max. to about 700 Mbps but decrease
CPU usage down to about 30% which I think will help all around. I have
been testing this driver all night and have not found a problem yet. Consider
this my Acked-by: for all these patches.</font>
<br><font size=2 face="sans-serif"><br>
Jim Lewis<br>
Advisory Software Engineer<br>
IBM Linux Technology Center<br>
512-838-7754<br>
<br>
<br>
</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td width=40%><font size=1 face="sans-serif"><b>Arnd Bergmann <arnd@arndb.de></b>
</font>
<p><font size=1 face="sans-serif">10/11/2006 11:02 AM</font>
<td width=59%>
<table width=100%>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">To</font></div>
<td><font size=1 face="sans-serif">Linas Vepstas <linas@austin.ibm.com></font>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">cc</font></div>
<td><font size=1 face="sans-serif">akpm@osdl.org, jeff@garzik.org, James
K Lewis/Austin/IBM@ibmus, netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
linuxppc-dev@ozlabs.org</font>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">Subject</font></div>
<td><font size=1 face="sans-serif">Re: [PATCH 0/21]: powerpc/cell spidernet
bugfixes, etc.</font></table>
<br>
<table>
<tr valign=top>
<td>
<td></table>
<br></table>
<br>
<br>
<br><tt><font size=2>On Tuesday 10 October 2006 22:49, Linas Vepstas wrote:<br>
> Andrew, please apply/forward upstream.<br>
> <br>
> The following set of 21 patches (!) are all aimed at the the <br>
> spidernet ethernet device driver. The spidernet is an etherenet<br>
> controller built into the Toshiba southbridge for the PowerPC Cell<br>
> processor. (This is the only device in existance that with this<br>
> ethernet hardware in it).<br>
> <br>
> These patches re-package/re-order/re-cleanup a previous<br>
> set of patches I've previously mailed. Thus, some have<br>
> been previously Acked-by lines, most do not. Most of<br>
> these patches are tiny, and handle problems that cropped<br>
> up during testing. Sorry about there being so many of them.<br>
> <br>
> The first set of 12 patches fix a large variety of mostly <br>
> minor bugs. <br>
> <br>
> The important patches are 13 through 17: these overcome a <br>
> debilitating performance problem on transmit (6 megabits<br>
> per second !!) on transmit of patches 500 bytes or larger.<br>
> After applying these, I am able to get the following:<br>
> <br>
> pkt sz speed (100K buffs) speed (4M buffs)<br>
> ------ ----------------- ----------------<br>
> 1500 700 Mbits/sec
951 Mbits/sec<br>
> 1000 658 Mbits/sec
770<br>
> 800 600
648<br>
> 500 500
500<br>
> 300 372
372<br>
> 60 70
70<br>
> <br>
> Above buf size refers to /proc/sys/net/core/wmem_default<br>
<br>
Excellent work! I guess this the best tx performance we've<br>
seen so far on this hardware.<br>
<br>
Consider this as an Acked-by: for all the patches, I'll save<br>
the effort of replying to each one of them separately.<br>
<br>
Jeff, do you plan on merging these fixes for 2.6.19?<br>
<br>
Arnd <><<br>
</font></tt>
<br>