On Tue, 14 Dec 2004 17:37:41 +0100, Juan Cespedes <[EMAIL PROTECTED]> said:
> On Tue, Dec 14, 2004 at 08:03:15AM -0800, Manoj Srivastava wrote:
>> > Being able to inspect the kernel variables and the kernel content
>> > with:
>> >> gdb /boot/vmlinux /proc/kcore
>>
>> I disagree. People who want the bare vmlinux can use the
>> configuration option install_vmlinux to get the debugging symbols
>> they need
> I would like to be able to have a copy of the vmlinux used to
> generate the kernel in kernel-image-* packages. Maybe it will
> overbloat the package, but we could add it to anothe package or at
> least leave it at some URL... I don't want to debug the kernel just
> for fun, but to be able to know what is going on when some weird
> things happen.
How is this any different from asking for debugging
information in all the packages all the time? There is a mechanism
for a user to get the debug symbols on a recompile, which is not very
dofferent from any other binary package in Debian.
> I could recompile my own kernel to create my vmlinux, but unless I
> use exactly the same enviroment (gcc, etc), the resulting vmlinux
> will not match the kernels shipped with Debian, and that's what I
> need.
The same argument holds for any other binary package as well,
I do not see why kernel images should be treated differently.
> I don't know what is the best solution for this, but I would really
> like to have access to the vmlinux used to generate our kernels...
Right. However, many users do not want to further bloat kernel
packages, so we have conflicting desires.
>> Just like other packages, it is easy enough to recreate the kernel
>> image package with debugging symbols built in, and that should be
>> all that is required.
> But unlike other packages, it is not easy to compile other kernel
> and just run it... in some situations, stopping a machine to try a
> new kernel is not acceptable.
There are production machines that are restrictred, yes. I
don't allow gcc and gdb on my server boxes, which makes debugging
hard. However, this is a local policy decision, and not a global
invariant; this should not be used to impose one set of preferences
on everyone else.
manoj
--
Be careful! UGLY strikes 9 out of 10!
Manoj Srivastava <[EMAIL PROTECTED]> <http://www.debian.org/%7Esrivasta/>
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
