Hi,
I've had a few problems trying to build a fresh 2.6.10 or 2.6.11
kernel. Nothing fancy. With gcc-4.0 I get tons of warnings, and
eventually it barfs when it gets to:
CC drivers/eisa/eisa-bus.o
cc1: warnings being treated as errors
drivers/eisa/eisa-bus.c:421: warning: pointer targets
S.Salman Ahmed [EMAIL PROTECTED] writes:
Is there any way to pass the -j jobs option to make when building
kernels using make-kpkg ?
Yes. See kernel-pkg.conf(5) for details.
--
Leonard Stiles [EMAIL PROTECTED]
Steve Kowalik wrote:
On Sun, Jul 01, 2001 at 02:42:20PM -0500, S.Salman Ahmed uttered:
Is there any way to pass the -j jobs option to make when building
kernels using make-kpkg ?
Read the README for kernel-package.
Make a .kernel-pkg.conf in ~ as the user you compile it as.
[EMAIL
On Mon, Jul 02, 2001 at 12:51:17PM -0500, John Foster wrote:
If I put the CONCURRENCY_LEVEL := 2 in my root environment variable will
it be useful in allowing all of my compiled applications to run with
maximum efficiency on my dual processor system?
No, it will only affect the (one-time)
On Sun, Jul 01, 2001 at 02:42:20PM -0500, S.Salman Ahmed uttered:
Is there any way to pass the -j jobs option to make when building
kernels using make-kpkg ?
Read the README for kernel-package.
Make a .kernel-pkg.conf in ~ as the user you compile it as.
[EMAIL PROTECTED]:~$ grep -i concurr
On Sun, Jul 01, 2001 at 02:42:20PM -0500, S . Salman Ahmed wrote:
Is there any way to pass the -j jobs option to make when building
kernels using make-kpkg ?
RTFM. 'man kernel-pkg.conf' and search for CONCURRENCY_LEVEL or just
'-j'.
I am hoping that by passing the -j option to make, kernel
i love debian. seriously, apt is a work of genius and the entire
system is exactly how i want it - unlike SuSE or RedHat. since i do a
fair bit of developing and since i always want to have at least one
machine that's cutting edge, i do a whole lot of kernel compiles.
in the past, i have always
always want to have at least one
machine that's cutting edge, i do a whole lot of kernel compiles.
in the past, i have always used .debs unless a software was too old or
not available, in which case i beat the tarball around and installed
into /usr/local. by now, i do it the debian way, and use
a whole lot of kernel compiles.
in the past, i have always used .debs unless a software was too old or
not available, in which case i beat the tarball around and installed
into /usr/local. by now, i do it the debian way, and use
dpkg-buildpackage to create the .deb, which i then install. i
at least one
machine that's cutting edge, i do a whole lot of kernel compiles.
in the past, i have always used .debs unless a software was too old or
not available, in which case i beat the tarball around and installed
into /usr/local. by now, i do it the debian way, and use
dpkg-buildpackage
also sprach Nathan E Norman (on Fri, 22 Jun 2001 11:01:13AM -0500):
Well, for one thing, you can compile kernels on your 1GHz Athlon
instead of your old 486 :) Since kernel-package creates a package, it
can be installed anywhere.
but i usually choose the appropriate kernel architecture during
also sprach Sean Morgan (on Fri, 22 Jun 2001 12:05:19PM -0400):
zless /usr/share/doc/kernel-package/Rationale
exaclty what i wanted. thanks!
martin; (greetings from the heart of the sun.)
\ echo mailto: !#^.*|tr * mailto:; [EMAIL PROTECTED]
--
your fly might be open (but
Sean Morgan [EMAIL PROTECTED] wrote:
zless /usr/share/doc/kernel-package/Rationale
I haven't figured out if kernel-package includes the
patches Herbert puts into the kernel-source .debs.
I tried to patch in 2.4.5-ac17 yesterday over
kernel-source-2.4.5 and ran afoul of some conflicts.
Would
a whole lot of kernel compiles.
in the past, i have always used .debs unless a software was too old or
not available, in which case i beat the tarball around and installed
into /usr/local. by now, i do it the debian way, and use
dpkg-buildpackage to create the .deb, which i then install. i
On Fri, Jun 22, 2001 at 06:34:25PM +0200, MaD dUCK wrote:
also sprach Nathan E Norman (on Fri, 22 Jun 2001 11:01:13AM -0500):
Well, for one thing, you can compile kernels on your 1GHz Athlon
instead of your old 486 :) Since kernel-package creates a package, it
can be installed anywhere.
On Fri, Jun 22, 2001 at 06:34:25PM +0200, MaD dUCK wrote:
also sprach Nathan E Norman (on Fri, 22 Jun 2001 11:01:13AM -0500):
Well, for one thing, you can compile kernels on your 1GHz Athlon
instead of your old 486 :) Since kernel-package creates a package, it
can be installed anywhere.
On Fri, Jun 22, 2001 at 01:19:49PM -0400, David L. Craig wrote:
Sean Morgan [EMAIL PROTECTED] wrote:
zless /usr/share/doc/kernel-package/Rationale
I haven't figured out if kernel-package includes the
patches Herbert puts into the kernel-source .debs.
I tried to patch in 2.4.5-ac17
On Fri, Jun 22, 2001 at 06:34:25PM +0200, MaD dUCK wrote:
also sprach Nathan E Norman (on Fri, 22 Jun 2001 11:01:13AM -0500):
Well, for one thing, you can compile kernels on your 1GHz Athlon
instead of your old 486 :) Since kernel-package creates a package, it
can be installed anywhere.
also sprach Nathan E Norman (on Fri, 22 Jun 2001 12:42:37PM -0500):
I fail to see why you think compiling a kernel on an Athlon, but
optimising for a 486 cos you're installing on a 486 is a problem.
that's what i am doing btw. and sorry, i wasn't possibly thinking
about multiple .debs, just
also sprach Bob Nielsen (on Fri, 22 Jun 2001 10:42:23AM -0700):
True, but you can/should configure and compile separately for each
destination architecture, optimizing appropriately.
so yes, the argument that my athlon (thunderbird actually) will
outperform the others still holds. but i'd need
On Fri, 22 Jun 2001 20:17:14 +0200, MaD dUCK writes:
also sprach Bob Nielsen (on Fri, 22 Jun 2001 10:42:23AM -0700):
True, but you can/should configure and compile separately for each
destination architecture, optimizing appropriately.
so yes, the argument that my athlon (thunderbird actually)
also sprach Robert Waldner (on Fri, 22 Jun 2001 08:20:41PM +0200):
you don´t need to maintain seperate trees of the whole source. just
`make (menu|x|)config`, then backup the .config-file. that´s where the
information you entered/chose is kept.
i understand... but when i change the .config
On Fri, 22 Jun 2001 20:28:04 +0200, MaD dUCK writes:
also sprach Robert Waldner (on Fri, 22 Jun 2001 08:20:41PM +0200):
you don´t need to maintain seperate trees of the whole source. just
`make (menu|x|)config`, then backup the .config-file. that´s where the
information you entered/chose
also sprach Robert Waldner (on Fri, 22 Jun 2001 08:44:51PM +0200):
now do a `make clean` (just to get rid of the modules, mostly) and
re-do steps 1-3 for the second machine and the third and...
when you want to re-do for the first machine, just move the appropriate
.config back, and
On Fri, Jun 22, 2001 at 08:17:14PM +0200, MaD dUCK wrote:
also sprach Bob Nielsen (on Fri, 22 Jun 2001 10:42:23AM -0700):
True, but you can/should configure and compile separately for each
destination architecture, optimizing appropriately.
so yes, the argument that my athlon (thunderbird
On Fri, Jun 22, 2001 at 08:44:51PM +0200, Robert Waldner wrote:
now do a `make clean` (just to get rid of the modules, mostly) and
make-kpkg clean
re-do steps 1-3 for the second machine and the third and...
It will complain if you use machines' hostnames in
I installed slink on my laptop, and compiled my kernel with
everything I need. Now when I try loading my PCMCIA netcard, I
get some unresolved symbols. I compiled the ne.o net driver as a
module, and that reduced the number of symbols not found, but are
still complaining about 2 functions it
I installed slink on my laptop, and compiled my kernel with
everything I need. Now when I try loading my PCMCIA netcard, I
get some unresolved symbols. I compiled the ne.o net driver as a
module, and that reduced the number of symbols not found, but are
still complaining about 2
28 matches
Mail list logo