From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <40A519F5.6010407@pp.iij4u.or.jp> Date: Sat, 15 May 2004 04:11:49 +0900 From: Yuji Nishimura MIME-Version: 1.0 To: Kumar Gala Cc: linuxppc-embedded , Stefan Nickl Subject: Re: Stability of network/io on MPC8540ADS? References: <1084462045.4683.68.camel@lucy.pep-kaufbeuren.de> <40A515E2.30401@pp.iij4u.or.jp> In-Reply-To: Content-Type: text/plain; charset=ISO-2022-JP Sender: owner-linuxppc-embedded@lists.linuxppc.org List-Id: Hi Kumar , You can check same patch on your internal SF server (Motorola internal). I uploaded patch on 5/6 or 5/7. Please check it... Best Regards, Yuji Nishimura -Tokyo Japan (rty922@email.sps.mot.com) Kumar Gala wrote: > Y.Nishimura , > > It looks like the patch was not included in your email. I would > appreciate seeing it and evaluating if it is something we should > include in the stock kernel. > > - kumar > > On May 14, 2004, at 1:54 PM, Yuji Nishimura wrote: > >> >> Helo Stefan, >> >> I evaluated MPC8540ADS and found out similar problem few weeks ago. I >> made tiny patch and it works well. >> I also evaluated with NEC ehci pci device. The result was fine. >> I wrote 1GB data to external usb-2.0 HDD via USB-2.0(ehci), that >> works well. >> I also try to use ftp(MPC8540=server) and transfer (R/W) 2047MB data via >> TSEC(gianfar) and it also works well. >> If you try to use this patch, please let me know. I think patch requires >> more evaluation... >> >> Best Regards, >> Y.Nishimura - Tokyo, Japan. >> >> >> >> Stefan Nickl wrote: >> >>> Hello, >>> >>> since the linuxppc-2.4 port to our custom 8540 board >>> has made some progress now, I went back to the ADS >>> board for some cross-checking. >>> >>> I've experienced several reproducible lockups that >>> are triggered by heavy IO and/or networking. >>> >>> Kernel is vanilla linuxppc-2.4.27-pre2, gcc 3.2.3, >>> root=/dev/nfs (Gentoo -Os -mcpu=603e) >>> >>> 1) Using netcat to write data over tcp: >>> cat /dev/zero | nc lucy 3000 >>> >>> "lucy" is running "nc -l -p 3000 |pipeworks > /dev/null" >>> This freezes the ADS after a few seconds. >>> >>> 2) Writing to the nfsroot server: >>> cat /dev/urandom > /test >>> >>> The same. Reading seems ok. >>> >>> 3) Writing to a USB2.0 HDD connected via PCI >>> interface board (OHCI compatible). >>> >>> Lockup after ~70MB of writing >>> >>> I haven't tested explicitly, but the kernel that >>> was shipped with the board (some 2.4.20?) didn't >>> seem to have such problems. >>> >>> Any ideas? >>> >>> -- >>> Stefan Nickl >>> Kontron Modular Computers >>> >>> >>> >> >> > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/