I have been following the wiki on how to build a custom kernel, and it works great.
Except, I am doing ongoing work with the internals of the kernel, and it is really awkward to set up a change and then build it using the standard methodology (edit the file to change, diff it against the original to make a patch file, rebuild the whole thing).
Isn't there a way (and what is it) to play around with the kernel source files until they are in the shape I want them to be in, and THEN go through the whole build process again.
Thanks.
Mark Hull-Richter, Linux Kernel Engineer DATAllegro (www.datallegro.com) 85 Enterprise, Second Floor, Aliso Viejo, CA 92656 949-680-3082 - Office 949-330-7691 - fax
Mark Hull-Richter wrote:
I have been following the wiki on how to build a custom kernel, and it works great.
Except, I am doing ongoing work with the internals of the kernel, and it is really awkward to set up a change and then build it using the standard methodology (edit the file to change, diff it against the original to make a patch file, rebuild the whole thing).
Isn't there a way (and what is it) to play around with the kernel source files until they are in the shape I want them to be in, and THEN go through the whole build process again.
Thanks.
Probably, you should be on lkml - the Linux kernel mailing list.
_I_ would just do my editing in the source tree (and I'd not normally use a vendor's source tree) and regularly do "make && make install" etc then copy to my test system if it's not my build system.
There is no need, or even point, to create rpms for each dumb mistake.
If at all possible, you should be testing first in uml, xen or other virtual environment.
On 4/15/07, John Summerfield debian@herakles.homelinux.org wrote:
Probably, you should be on lkml - the Linux kernel mailing list.
I used to be, but that was ages ago. Guess I'll be rejoining when I find it....
There is no need, or even point, to create rpms for each dumb mistake.
Oh, I don't know - I could screw up more people who were dumb enough to take my rpms....
:-)
If at all possible, you should be testing first in uml, xen or other
virtual environment.
I have a dedicated test machine here at work. It builds nicely and so far I haven't built a kernel that won't boot, AND I have fallbacks for if I do....
Thanks.
mhr
Mark Hull-Richter wrote:
On 4/15/07, John Summerfield debian@herakles.homelinux.org wrote:
Probably, you should be on lkml - the Linux kernel mailing list.
I used to be, but that was ages ago. Guess I'll be rejoining when I find it....
There is no need, or even point, to create rpms for each dumb mistake.
Oh, I don't know - I could screw up more people who were dumb enough to take my rpms....
:-)
If at all possible, you should be testing first in uml, xen or other
virtual environment.
I have a dedicated test machine here at work. It builds nicely and so far I haven't built a kernel that won't boot, AND I have fallbacks for if I do....
Seems to me that typing linux ... or xm create testbox
is much easier than moving to another computer, pressing reset (if it's got one!) or cycling power, or even 'reboot -f -n'
Okay, so you don't move to another computer. Still, I've heard UML became pretty popular pretty quickly with kernel hackers.
On 4/15/07, John Summerfield debian@herakles.homelinux.org wrote:
Probably, you should be on lkml - the Linux kernel mailing list.
Egad - I just looked at it, and now I know why I'm not on it any more - WAYYYYYY too high traffic for me - I'd never get anything done....
Bummer - I'll probably wind up on it anyway.
(sigh)
Mark Hull-Richter wrote:
On 4/15/07, John Summerfield debian@herakles.homelinux.org wrote:
Probably, you should be on lkml - the Linux kernel mailing list.
Egad - I just looked at it, and now I know why I'm not on it any more - WAYYYYYY too high traffic for me - I'd never get anything done....
Bummer - I'll probably wind up on it anyway.
Last time I was on it (some years ago), I set up inn and fed the list into that. I don't see now why it should be so, but I found it easier to keep track of it with a newsreader, likely tin, and inn can be set to expire stuff sensibly.
On Sun, Apr 15, 2007 at 03:26:00PM -0700, Mark Hull-Richter wrote:
I have been following the wiki on how to build a custom kernel, and it works great.
Except, I am doing ongoing work with the internals of the kernel, and it is really awkward to set up a change and then build it using the standard methodology (edit the file to change, diff it against the original to make a patch file, rebuild the whole thing).
Isn't there a way (and what is it) to play around with the kernel source files until they are in the shape I want them to be in, and THEN go through the whole build process again.
make kernel, on the kernel sources creates rpm files.
-
John Summerfield -
Luciano Miguel Ferreira Rocha -
Mark Hull-Richter