This is the mail archive of the
mailing list for the Archer project.
Re: prototyping linux kernel-side gdbstub for userspace
- From: Tom Tromey <tromey at redhat dot com>
- To: Daniel Jacobowitz <drow at false dot org>
- Cc: archer at sourceware dot org, "\"\"Frank Ch. Eigler\"\"" <fche at redhat dot com>
- Date: Tue, 23 Jun 2009 12:19:34 -0600
- Subject: Re: prototyping linux kernel-side gdbstub for userspace
- References: <20090620023204.GK19576@redhat.com><20090621151606.GA10157@caradoc.them.org>
- Reply-to: Tom Tromey <tromey at redhat dot com>
Daniel> I can't say I'm thrilled to have this in kernel space.
What concerns do you have?
I was fairly surprised by this approach when I first heard about it
(in fact, I had joked about doing this before I found out about it
:-). I figured that the gdbserver protocol was probably pretty crufty
(I actually don't know much about it, so this is reasoning from
ignorance at its best) and would be unacceptable to kernel developers
But now it seems like a decent enough idea. IIUC, kgdb works this
way, so some gdbserver protocol implementation is already in-kernel.
This still may not be the ideal approach -- could we really advertise
this as successor to ptrace? -- but it does have the nice advantage of
not requiring changes to gdb to try it out.