Red Hat Bugzilla – Bug 124312
Memory Detection - Install kernal boot failure
Last modified: 2007-11-30 17:10:43 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6)
Description of problem:
Target system is Compaq Proliant 5000 w/ 512MB
System hangs after displaying on console:
Uncompressing Linux... Ok, booting the kernel.
Version-Release number of selected component (if applicable):
Fedora core 2
Steps to Reproduce:
1.Boot from Fedora Core 2 CD 1
2.Reply "linux mem=exactmap mem=640k@0 mem=511m@1m" to boot:
Expected Results: The installation kernel should boot.
I'm fairly sure I did an install on one of these during FC2 test,
using just mem=512M does that work for you?
"linux mem=512m" fails also. linux can't boot from device <null>.
ah, but it gets further than before right ?
This new error sounds like something else, like it isn't finding your
root device for some reason.
The fact remains that the options that worked for FC1 do not work for
FD2: "linux mem=exactmap mem=640k@0 mem=511m@1m". Does the support org
have access to equipment to replicate?
When I do not provide any options,the kernel does not detect enough
memory to continue. This is consistant between FC1 & FC2. I conclude
that there is not a hardware or media problem.
If I try again by specifying the the mem option ("linux
mem=512m"),3what is the "root" option?
I just tried mem=512m on FC1 with similar results. The Kernel cannot
boot from the ramdisk.
it seems that this code got changed in 2.5 to reflect changes in 2.4,
and then got backed out and replaced with memmap= for the above
functionality. 2.4 never changed accordingly.
I'm trying to sort this mess out upstream right now on the
linux-kernel mailing list.
Glad I could be helpfull.
upstream opinion is that things aren't going to change.
so, for 2.6 kernels use memmap=, 2.4 use mem=
exactmap now belongs to the memmap argument.