public inbox for gdb@sourceware.org
 help / color / mirror / Atom feed
From: Dave Trollope <trollope@lucent.com>
To: Daniel Jacobowitz <drow@false.org>
Cc: gdb@sources.redhat.com
Subject: Re: Linux Realtime Scheduling Option
Date: Tue, 22 Mar 2005 03:04:00 -0000	[thread overview]
Message-ID: <423F8B07.9010603@lucent.com> (raw)
In-Reply-To: <20050321193304.GA5205@nevyn.them.org>

Hi Daniel,

Daniel Jacobowitz wrote:

>On Mon, Mar 21, 2005 at 01:20:02PM -0600, David Steven Trollope wrote:
>  
>
>>Hi Daniel,
>>
>>Our application does change its own priority, but I was concerned with 
>>the priority of gdbserver/gdb. Which Linux tools are you referring to? 
>>I'll go take a look at them.
>>    
>>
>
>Search for 'rt' or 'chrt'; I do not recall which one is current.  I
>believe they are in the 'schedutils' distribution.
>  
>
I'll take a look.

>>In our environment gdb/gdbserver should always run realtime at a set 
>>priority. Help me understand why is it not a good idea to have 
>>gdb/gdbserver set its own priority based on an option in .gdbinit?
>>    
>>
>
>First of all, gdbserver doesn't parse an init file.  You would have to
>add a Linux-specific packet type to the remote protocol for GDB to
>communicate this to gdbserver.
>  
>
I wasn't thinking this was Linux specific. Doesn't Solaris also have 
realtime extensions that this would apply to?

I can certainly see that its not generic enough to be worthy of adding 
to the remote protocol.


>Secondly, because there are standalone tools to handle the problem.
>gdbserver is supposed to be simple; I don't want to add code specific
>to a particular, fairly uncommon debugging environment when existing
>tools handle it perfectly well.
>
>If you can come up with a reason why the standalone tools can not be
>used to solve the problem, then we can rediscuss :-)
>  
>
Its not a gdbserver specific issue. If you run gdb on any Linux machine 
while loading an application that sets its own real time priority 
problems will likely occur from the mismatch. I don't know if Solaris 
suffers from the same issue, but I would expect it to.

I think its really applicable to any unix variant where realtime 
extentions allow an application to change its own priority.

Perhaps gdb/gdbserver could have a command line option (instead of 
.gdbinit) that registers itself realtime, as the highest priority? Its 
likely you want the debugger to run at the highest priority to catch 
tight loops etc. This would also map to non-unix OSs like VxWorks.

Does this strategy sound more appealing?

I appreciate your time.

Cheers
Dave

  reply	other threads:[~2005-03-22  3:04 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1093622671.2836.ezmlm@sources.redhat.com>
2004-08-27 17:56 ` GDB/MI Output Syntax Jim Ingham
2004-08-27 19:12   ` Michael Chastain
2005-01-05 23:27     ` Bob Rossi
2005-01-06  4:48       ` Eli Zaretskii
2005-01-06 23:31         ` Bob Rossi
2005-01-07  0:36           ` Jim Ingham
2005-01-07  1:12             ` Bob Rossi
2005-01-07  3:12               ` Russell Shaw
2005-01-11 19:35                 ` Bob Rossi
2005-01-13  2:23                   ` Bob Rossi
2005-01-13  2:46                   ` Intrusive GDB Symbol Lookup when debugging remotely David Steven Trollope
2005-01-22  4:25                     ` Dave Trollope
2005-01-24 19:48                       ` Andrew Cagney
2005-01-24 19:54                         ` David Steven Trollope
2005-03-18 16:29                     ` Linux Realtime Scheduling Option David Steven Trollope
2005-03-18 18:12                       ` Daniel Jacobowitz
2005-03-21 19:21                         ` David Steven Trollope
2005-03-21 19:33                           ` Daniel Jacobowitz
2005-03-22  3:04                             ` Dave Trollope [this message]
2005-03-22  4:06                               ` Daniel Jacobowitz

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=423F8B07.9010603@lucent.com \
    --to=trollope@lucent.com \
    --cc=drow@false.org \
    --cc=gdb@sources.redhat.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).