Messages in this thread | | | Date | Tue, 29 Feb 2000 12:34:25 -0700 (MST) | From | "Ronald G. Minnich" <> | Subject | Re: Load linux...from linux? (fwd) |
| |
continuing the linux-loads-linux discussion. Werner's new code is quite cool. I'm somewhat concerned about turning off PTEs during the process. I did a bit of work in LOBOS to leave all the VM hardware on. My worries are probably groundless ...
as for linuxbios, I'm now into debugging start_kernel. I haven't got interrupts turned on right (it's somewhat harder from a hard reset) but hope to have that worked out later today.
ron
---------- Forwarded message ---------- Date: Tue, 29 Feb 2000 19:35:16 +0100 (MET) From: Werner Almesberger <almesber@lrc.epfl.ch> To: Ronald G. Minnich <rminnich@lanl.gov> Subject: Re: Load linux...from linux?
Ronald G. Minnich wrote:
> Your code is very nice.
Thanks ! ;-) BTW, I've made a new version which should raise the limit for loaded data to 2GB:
ftp://icaftp.epfl.ch/pub/people/almesber/misc/bootimg-1.tar.gz
Now I need to build a kernel >> 1 GB to actually test this ...
> Question: If I have 1 GB or less there's no issue with addressing all of > physical memory from the current kernel virtual address space.
Yep, your approach is limited to the kernel address space. Not sure if this is really a problem, though:
- vmalloc should always give you pages you can reach from the kernel - I don't think any architecture puts boot code beyond the kernel address space (it could make a little bit of sense for things like initrd, but then, any solution requiring multi-gigabyte initrds is at least suspicious, I'd say ;-) - 1 GB for kernel, initrd, parameters, and friends should be sufficient for most practical purposes.
> Also I'd like to leave some hook in there for continuing with the page > table usage because the next step is the Alpha.
What are actually the requirements on the Alpha ? Does the kernel expect to be started with page tables already set up ? If not, is there no way to simply turn off paging ? In my approach, one could add some page tables to the "sacred" pages, if necessary, but it's a little ugly.
- Werner
-- _________________________________________________________________________ / Werner Almesberger, ICA, EPFL, CH werner.almesberger@ica.epfl.ch / /_IN_N_032__Tel_+41_21_693_6621__Fax_+41_21_693_6610_____________________/
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |