This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFC][PATCH] Allow JIT unwinder provide symbol information
- From: Alexander Smundak <asmundak at google dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: gdb-patches <gdb-patches at sourceware dot org>
- Date: Mon, 19 May 2014 14:30:54 -0700
- Subject: Re: [RFC][PATCH] Allow JIT unwinder provide symbol information
- Authentication-results: sourceware.org; auth=none
- References: <20131226183618 dot D264CA18A0 at sasha2 dot mtv dot corp dot google dot com> <5359101B dot 1080603 at redhat dot com> <CAHQ51u40jadvuYqmYBqs4yFnFy96HZ46GmKpj3=_V3G9NSF12w at mail dot gmail dot com> <535FC3AC dot 6030200 at redhat dot com> <CAHQ51u6sOVSqwFCfiq6FOOZUrp4ag02kmGrgvSCzPGD0ijFqMA at mail dot gmail dot com>
Ping.
On Fri, May 2, 2014 at 9:58 AM, Alexander Smundak <asmundak@google.com> wrote:
> On Tue, Apr 29, 2014 at 8:22 AM, Pedro Alves <palves@redhat.com> wrote:
>> But it doesn't actually provide symbol information. Not in the sense
>> that it hooks with GDB's symbol lookup mechanisms.
> I cannot find a better way to hook into GDB symbols lookup machinery.
> It looks to me that GDB is operating on assumption that symbol information
> is static (except for shared objects loading), and it not frame-specific.
> The code emitted by JIT does not conform to these assumptions.
> IMHO what I am proposing has the beginnings of the approach to
> overcome this difficulty by allowing unwinder to have its own symbol handler.
> It this approach looks reasonable, obviously more changes are needed,
> but I'd like to have a confirmation that the approach is reasonable.