Hi,
Does anyone know which model of the bios savior is needed for via eden motherboards? I would like to try out LinuxBIOS, but would like to have an easy way back if anything goes wrong.
Jody
On my motherboard the stock BIOS was 256K. So I just used:
dumpdevmem 0xfffc 0x40000 > stockbios
Dumpdevmem was just some tool that came with linuxbios and did a mmap from /dev/mem.
I was using a Grammar Engine ROM ICE for my development, so this was handy. I could easily go back and forth between the linuxbios and stock by changing the load I placed into the ICE.
Kevin
On Tuesday 29 October 2002 19:42, Jody Belka wrote:
Hi,
Does anyone know which model of the bios savior is needed for via eden motherboards? I would like to try out LinuxBIOS, but would like to have an easy way back if anything goes wrong.
Jody
Linuxbios mailing list Linuxbios@clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios
Jody,
Does anyone know which model of the bios savior is needed for via eden motherboards? I would like to try out LinuxBIOS, but would like to have an easy way back if anything goes wrong.
RD1-PL is fine.
-Andrew
Hi,
Does anyone know what util/nrv2b/nrv2b.c is supposed to be? I did a CVS update to pull in the latest linuxbios code (so I can send in a patch). However, it seems that the build now needs this tool (somehow related to the new BIOS compression?).
Did someone forget to check this file in?
Kevin
Never mind - my cvs install was on crack.
;-)
On Wednesday 30 October 2002 11:34, Kevin Hester wrote:
Hi,
Does anyone know what util/nrv2b/nrv2b.c is supposed to be? I did a CVS update to pull in the latest linuxbios code (so I can send in a patch). However, it seems that the build now needs this tool (somehow related to the new BIOS compression?).
Did someone forget to check this file in?
Kevin
Kevin Hester kevinh@ispiri.com writes:
Hi,
Does anyone know what util/nrv2b/nrv2b.c is supposed to be? I did a CVS update to pull in the latest linuxbios code (so I can send in a patch). However, it seems that the build now needs this tool (somehow related to the new BIOS compression?).
That is the compressor.
Did someone forget to check this file in?
Yes, but not this last round.
do a cvs update -d? To get the new directories. Recommended is to always use cvs update -d -P which will also prune out empty directories.
Eric
you did a cvs update -d, right? it's in there in my tree.
ron
Hi,
I recall reading a white paper about Linuxbios and patches so that the kernel can boot other kernels? Does anyone know where this code is? I've poked around with no luck.
For our application, I'd like to place a minimal kernel into the FLASH with the BIOS. That kernel would either use IP or IDE to read in the 'real' kernel.
I considered chaining into Etherboot, but I need to boot from IDE drives in addition to Ethernet. The Etherboot CAN_BOOT_DISK flag seems to require IDE read funtionality in the BIOS (a no-no with linuxbios?).
Thanks for any advice. I've only been futzing with linuxbios since last Friday so I'm still new at this.
Kevin
On Thu, 31 Oct 2002, Kevin Hester wrote:
I recall reading a white paper about Linuxbios and patches so that the kernel can boot other kernels? Does anyone know where this code is? I've poked around with no luck.
just go with kexec, it's the standard now for 2.5. There are other packages but my inclination is to go with whatever comes with the base kernel.
For our application, I'd like to place a minimal kernel into the FLASH with the BIOS. That kernel would either use IP or IDE to read in the 'real' kernel.
that's how we do it here. Works quite nicely.
I considered chaining into Etherboot, but I need to boot from IDE drives in addition to Ethernet. The Etherboot CAN_BOOT_DISK flag seems to require IDE read funtionality in the BIOS (a no-no with linuxbios?).
No, there is an ide driver in etherboot. But you get no filesystems.
There is a point of disagreement in this community: is linux too heavy to be a boot loader, but is its complexity needed; is etherboot a nice and light and fast bootloader, but is it too simplistic. That is the disagreement described in its broadest form.
For now, here at lanl, we come down on the side of using Linux wherever possible and we have been happy with that decision. Just yesterday I had a problem with etherboot that was solved by using Linux as the bootloader instead. Overall I prefer using Linux as my bootloader, but I am also extremely impressed with etherboot's growing capabilities.
ron
Ronald G Minnich rminnich@lanl.gov writes:
On Thu, 31 Oct 2002, Kevin Hester wrote:
I recall reading a white paper about Linuxbios and patches so that the kernel can boot other kernels? Does anyone know where this code is? I've poked around with no luck.
just go with kexec, it's the standard now for 2.5. There are other packages but my inclination is to go with whatever comes with the base kernel.
Ron that logic unfortunately does not work. So far Linus has just ignored kexec. kexec is the only one available for 2.5.x though.
For our application, I'd like to place a minimal kernel into the FLASH with the BIOS. That kernel would either use IP or IDE to read in the 'real' kernel.
that's how we do it here. Works quite nicely.
I considered chaining into Etherboot, but I need to boot from IDE drives in addition to Ethernet. The Etherboot CAN_BOOT_DISK flag seems to require IDE read funtionality in the BIOS (a no-no with linuxbios?).
No, there is an ide driver in etherboot. But you get no filesystems.
There is a point of disagreement in this community: is linux too heavy to be a boot loader, but is its complexity needed; is etherboot a nice and light and fast bootloader, but is it too simplistic. That is the disagreement described in its broadest form.
For now, here at lanl, we come down on the side of using Linux wherever possible and we have been happy with that decision. Just yesterday I had a problem with etherboot that was solved by using Linux as the bootloader instead. Overall I prefer using Linux as my bootloader, but I am also extremely impressed with etherboot's growing capabilities.
If I could seriously get the attention of the kernel developers. And stabalize the kernel so it properly shutdown the hardware I would be more persuaded. At least until I can get kexec into the kernel or we start a our own kernel tree for network booting and develop from there I am not especially interested in using the Linux kernel because it does not work correctly.
Eric
-
Andrew Ip -
ebiederman＠lnxi.com -
Jody Belka -
Kevin Hester -
Ronald G Minnich