Yes, SS4000E , based on EM-7210 card. I got the kernel sources from www.kernel.org I took version 2.6.25.1 then I patched it applying all the patches present in
http://git.rtp-net.org/?p=ss4000e.git;a=summary I DON'T applied the commented out patches: this is the series file: compiler.patch hack_machtypes.patch gpiodrv.patch iscsitgt.patch em7210_report_ac_status.patch #dma_test.patch try-dma-copy-to-user.patch #sdram.patch convert-copy-to-user.patch cheapen-get-user-pages.patch fix-nfs.patch defconfig.patch iop3xx_gpio.patch gpiolib_regs_show.patch f75111.patch add_f75111_pdata.patch em7210_add_missing_leds.patch gpio_polled_keys.patch em7210_add_gpio_keys.patch #add_triggers.patch crypto_cypher_buildfix2.patch so actually I didn't apply "dma_test.patch" Do you think try to recompile zImage after applying this patch could bring to better performance ? My second question was about disks status led: It could be fine to see the lighted when they are working. With the original software they lights on traffic, so I think that they are not broken. thanks, Andrea -----Messaggio originale----- Da: Arnaud Patard [mailto:[EMAIL PROTECTED] Inviato: martedì 18 novembre 2008 16.39 A: Andrea Lanza Cc: '[email protected]' Oggetto: Re: performance information Andrea Lanza <[EMAIL PROTECTED]> writes: > hi al, Hi, > I am completly new to this list and to Debian and arm too. > > Anyway I succeded in installing my Intel S4000e with kernel 2.6.25.1 and > latest lenny Debian. You meant SS4000e, right ? :) > I cross compiled my kernel using the patch I found on the net and then moved > the zImage > to the scratched (init-ialized) RedBoot Flash. > > I really have to thanks this site > http://em7210.kwaak.net/cgi-bin/trac.cgi/wiki because > without it I could never accomplish to this task > > Everything seems fine, I also installed nfs and samba joined to my domain. > > The question is: I scratched my SS400E because it was really slow, using > samba AND nfs > was almost impossible (no more than 1MB/s write speed) > > Now the performance are considerably better than before, but I think they are > still poor. which kernel sources are you using ? There are some experimental patches for dma offload which should help getting things better (look to the archives in this mailing-list. There should also be a bug open for the iirc) Arnaud -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
