<P>(1) IP has the fragmentation function, ip should support it correctly
<P>(2) why windows98 can handle the matter correctly, but linux can't , maybe linux pci
<P>or sis900 ehternet driver has some bug
<P>(3) maybe linuxbios init sis900 so unusual that linux can't handle it correctly.
<P> <B><I>ollie lho <firstname.lastname@example.org></I></B> wrote:
<BLOCKQUOTE style="BORDER-LEFT: #1010ff 2px solid; MARGIN-LEFT: 5px; PADDING-LEFT: 5px">On Mon, 2002-10-14 at 22:30, zhu shi song wrote:<BR>> I met one strange error on sis900 ethernet driver. <BR>> I'm using cvs freebios on 6 Oct, 2002 , kernel 2.4.17,<BR>> mainboard is winfast6300. After booting using<BR>> linuxbios stored in DOC, I boot <BR>> my host linux system using normal bios. the host<BR>> linux<BR>> is the same system as in DOC. I test pings, the<BR>> result<BR>> is ( after linuxbios booting):<BR>> ping -s 2592 18.104.22.168 ( remote ip addr ) 0% lost<BR>> ping -s 2593 22.214.171.124 ( remote ip addr ) no<BR>> response, 100% lost<BR><BR>Why are your using a packet size of 259x bytes ?? It just makes<BR>no sense since the MTU of ethernet is about 1500.<BR><BR>> I can't make the mainboard work correctly until I have<BR>> booted<BR>> windows98. Then I use my host linux system to boot, <BR>> the thing is ok. (after windows98 booting )<BR>> ping -s 2592 126.96.36.199 0% lost<BR>> ping -s 2593 188.8.131.52 0% lost<BR>> <BR>> What's the problem, maybe something wrong with pci<BR>> ,or sis900 init <BR>> code.<BR><BR>Some early 630A/E chips has some kind of bugs due to semiconductor <BR>fab process. Probably you are just unlucky to get a bad one.<BR><BR>Ollie<BR><BR><BR>_______________________________________________<BR>Linuxbios mailing list<BR>Linuxbios@clustermatic.org<BR>http://www.clustermatic.org/mailman/listinfo/linuxbios</BLOCKQUOTE><p><br><hr size=1>Do you Yahoo!?<br>
<a href="http://faith.yahoo.com">Faith Hill</a> - Exclusive Performances, Videos, & more<br>