Messages in this thread | | | Date | Tue, 11 Jul 2000 13:22:36 -0400 | From | Jeff McNeil <> | Subject | Re: Problems with Compaq Proliant and 2.4.0-test series |
| |
<!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html> On the 1850R's atleast, setting OS to Unixware locks the SMP spec in at 1.1, configing anything other than Unixware 7 will set the SMP spec to 1.4. <p>You should have a "Smart Start" CD, which will give you more granularity when it comes to configuring your Compaq BIOS. <p>There are other changes as well that the "OS setting" controls, unfortunatly, I don't remember off hand what exactly they are. There are some PCI bus specifics in there, if I remember correctly... Something about "Hot Add" of resources perhaps? <p>Thanks, <p>Jeff <br> <br> <p>"Richard B. Johnson" wrote: <blockquote TYPE=CITE>On Tue, 11 Jul 2000, Torben Mathiasen wrote: <p>> On Mon, Jul 10 2000, Mario Mikocevic wrote: <br>> > Well, it seems that #1 and #2 are corelated. Now, here's even more fun things -> <br>> > <br>> > if I boot test2 with OS setting set to OTHER server boots as UP and eepro100 works, <br>> > <br>> > if I boot test3 (pre2, pre5 and pre7 tested) with OS on OTHER server boots as UP but <br>> > eepro100 spits 'eth0: card reports no RX buffers.' with endless slow stream of <br>> > 'eepro100: wait_for_cmd_done timeout!' !? <br>> > <br>> > if I boot any of those test[23] kernels with BIOS OS setting to Unixware (set as 7 <br>> > I don't know if it makes difference) server boots as SMP but eepro100 is deaf with <br>> > following errors : <br>> > NETDEV WATCHDOG: eth0: transmit timed out <br>> > eth0: Transmit timed out: status e050 0c00 at 0/28 command 0001a000. <p>Some BIOS(es) fail to put a usable address in the PCI registers for <br>some devices. For instance, if you have a shared-memory buffer that <br>is Nk in length, it must be on a Nk boundary! Some BIOSes ignore this <br>which may prevent the first portion of an allocated buffer from being <br>addressed. See if you can get a BIOS upgrade. If not, see if your driver <br>can take a parameter (using append=) from LILO. <p>Cheers, <br>Dick Johnson <p>Penguin : Linux version 2.2.15 on an i686 machine (797.90 BogoMips). <p>"Memory is like gasoline. You use it up when you are running. Of <br>course you get it all back when you reboot..."; Actual explanation <br>obtained from the Micro$oft help desk. <p>- <br>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in <br>the body of a message to majordomo@vger.rutgers.edu <br>Please read the FAQ at <a href="http://www.tux.org/lkml/">http://www.tux.org/lkml/</a></blockquote>
<pre>-- Jeff McNeil Sr. Systems Engineer Interland, Inc. (678)365-2807</pre> </html>
| |