This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: fedora rawhide problems with kprobes/debuginfo addresses
- From: Vara Prasad <prasadav at us dot ibm dot com>
- To: vgoyal at in dot ibm dot com
- Cc: Roland McGrath <roland at redhat dot com>, Dave Jones <davej at redhat dot com>, "Frank Ch. Eigler" <fche at redhat dot com>, systemtap at sources dot redhat dot com
- Date: Mon, 05 Mar 2007 09:56:59 -0800
- Subject: Re: fedora rawhide problems with kprobes/debuginfo addresses
- References: <20070305040709.GA4880@in.ibm.com> <20070305063449.683141801C4@magilla.sf.frob.com> <20070305071641.GB4880@in.ibm.com>
Vivek Goyal wrote:
On Sun, Mar 04, 2007 at 10:34:49PM -0800, Roland McGrath wrote:
Ok, so you are talking about using magic arch knowledge to know what
virtual address range might include the kernel mapping and finding a memory
region there. That is worse than the kallsyms heuristic I'm doing, for
purposes of making things reasonable in userland.
I understand that using /proc/kallsyms is much easier but it does not work
if somebody built a custom kernel without kallsyms support.
How often customers build their own custom kernels? If they do build,
doesn't distro support contract gets void?
I personally think usefulness of kallsyms outweigh the risk of not
supporting custom kernels.
Thanks
Vivek