Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: "Marcin Kościelnicki" <koriakin@0x04.net>
To: Yao Qi <qiyaoltc@gmail.com>
Cc: gdb-patches@sourceware.org
Subject: Re: [PATCH 0/4] Save target description in tfile.
Date: Fri, 12 Feb 2016 01:49:00 -0000	[thread overview]
Message-ID: <56BD3A20.5070409@0x04.net> (raw)
In-Reply-To: <86ziv7no7i.fsf@gmail.com>

On 11/02/16 18:36, Yao Qi wrote:
> Marcin Kościelnicki <koriakin@0x04.net> writes:
>
>> 1. target.xml is read from the target.
>> 2. Includes are processed, resulting in a single in-memory XML file
>>     containing all the data.
>> 3. The resulting file is stored in tfile header by prefixing every line
>>     with "tdesc ".  We may insert a spurious newline at the end of file
>>     with this encoding, but that won't matter for XML.
>> 4. When tfile is read, the XML is stored in an allocated buffer, and
>>     xfer for TARGET_OBJECT_AVAILABLE_FEATURES is implemented, reading
>>     from it.
>> 5. target_find_description is called to force reading it.
>
> The target description is a global state in trace file, so it doesn't
> work tracepoints have different target descriptions, for example in
> multi-inferior case, one is x86 and the other one is x86_64 (and arm vs
> aarch64 too), and we have tracepoints set in two processes.
>
> I know tracepoint doesn't support multi-inferior/multi-process, but we
> may support that in the future.  Since we are changing tracepoint file
> format, we'd better think about this a little bit.  Can we save multiple
> target descriptions in trace file, and associate the tracepoints to the
> right target description in trace file?
>
Well, that'd require rethinking the whole format - there's already the 
'R' line, which is also global and specifies register packet size, which 
is probably different for every target (it's effectively a very poor 
kind of tdesc).  There's also no way of assigning the tracepoints to 
inferiors, or for that matter to say what executable is running in what 
inferior (target tfile just picks the current one loaded in gdb).

In effect, we'd have to prefix every single line with the inferior id 
(or otherwise include it), and the result is probably no better than 
just bundling several tfile files eg. in a zip (I don't think we have 
support for merging trace buffers from several sources anyhow).


  reply	other threads:[~2016-02-12  1:49 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-06 15:39 Marcin Kościelnicki
2016-02-06 15:39 ` [PATCH 2/4] gdb.trace: Read XML target description from tfile Marcin Kościelnicki
2016-02-10 13:02   ` Pedro Alves
2016-02-10 20:19     ` [PATCH 2/3] " Marcin Kościelnicki
2016-02-10 22:20       ` Pedro Alves
2016-02-10 22:35         ` Marcin Kościelnicki
2016-02-06 15:39 ` [PATCH 3/4] gdb.trace: Use g packet order in tfile_fetch_registers Marcin Kościelnicki
2016-02-10 13:20   ` Pedro Alves
2016-02-10 13:21     ` Marcin Kościelnicki
2016-02-10 13:54       ` Pedro Alves
2016-02-10 14:12         ` [PATCH] " Marcin Kościelnicki
2016-02-10 14:21           ` Pedro Alves
2016-02-10 14:49             ` Marcin Kościelnicki
2016-02-06 15:39 ` [PATCH 1/4] gdb.trace: Save XML target description in tfile Marcin Kościelnicki
2016-02-10 13:02   ` Pedro Alves
2016-02-10 20:18     ` [PATCH 1/3] " Marcin Kościelnicki
2016-02-10 22:20       ` Pedro Alves
2016-02-10 22:35         ` Marcin Kościelnicki
2016-02-11 19:02           ` Simon Marchi
2016-02-11 21:28             ` Sergio Durigan Junior
2016-02-11 22:31             ` Marcin Kościelnicki
2016-02-11 22:56             ` [PATCH] gdb: Fix build failure in xml-tdesc.c without expat Marcin Kościelnicki
2016-02-12 10:06               ` Pedro Alves
2016-02-12 10:10                 ` Marcin Kościelnicki
2016-02-12 10:19                   ` Pedro Alves
2016-02-12 10:21                     ` Marcin Kościelnicki
2016-02-06 15:47 ` [PATCH 4/4] gdb.trace: Fix off-by-one in tfile_fetch_registers Marcin Kościelnicki
2016-02-10 13:22   ` Pedro Alves
2016-02-10 13:53     ` Marcin Kościelnicki
2016-02-06 20:54 ` [PATCH 5/6] gdb/x86: Implement ax_pseudo_register_collect hook Marcin Kościelnicki
2016-02-06 20:54   ` [PATCH 6/6] gdb.trace: Add a testcase for tdesc in tfile Marcin Kościelnicki
2016-02-10 13:50     ` Pedro Alves
2016-02-11 10:14       ` [PATCH] " Marcin Kościelnicki
2016-02-11 13:00         ` Pedro Alves
2016-02-11 14:17           ` Marcin Kościelnicki
2016-02-12 18:32             ` Antoine Tremblay
2016-02-12 18:40               ` Marcin Kościelnicki
2016-02-12 18:49                 ` Antoine Tremblay
2016-02-12 18:53                   ` Marcin Kościelnicki
2016-02-12 18:57                     ` Antoine Tremblay
2016-02-12 19:47                       ` [PATCH] gdb.trace/tfile-avx.c: Change ymm15 to xmm15 for old gcc Marcin Kościelnicki
2016-02-12 19:57                         ` Pedro Alves
2016-02-12 21:34                           ` Marcin Kościelnicki
2016-02-10 13:31   ` [PATCH 5/6] gdb/x86: Implement ax_pseudo_register_collect hook Pedro Alves
2016-02-10 13:34     ` Marcin Kościelnicki
2016-02-10 13:50       ` Pedro Alves
2016-02-10 14:27         ` [PATCH] " Marcin Kościelnicki
2016-02-10 14:28           ` Pedro Alves
2016-02-10 14:49             ` Marcin Kościelnicki
2016-02-10 14:24 ` [PATCH 0/4] Save target description in tfile Pedro Alves
2016-02-10 14:25   ` Marcin Kościelnicki
2016-02-10 20:19   ` [PATCH 3/3] gdb/doc: Add documentation for tfile description section lines Marcin Kościelnicki
2016-02-10 22:20     ` Pedro Alves
2016-02-17  9:50       ` Marcin Kościelnicki
2016-02-17 15:37         ` Pedro Alves
2016-02-17 15:38           ` Pedro Alves
2016-02-17 15:59         ` Eli Zaretskii
2016-02-17 16:04           ` Marcin Kościelnicki
2016-02-17 16:05             ` Pedro Alves
2016-02-18  8:28               ` Marcin Kościelnicki
2016-02-11 17:36 ` [PATCH 0/4] Save target description in tfile Yao Qi
2016-02-12  1:49   ` Marcin Kościelnicki [this message]
2016-02-12 12:13     ` Yao Qi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=56BD3A20.5070409@0x04.net \
    --to=koriakin@0x04.net \
    --cc=gdb-patches@sourceware.org \
    --cc=qiyaoltc@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox