Re: Integration of timed breakpoints into GDB

[logitech <supriya dot rao17 at gmail dot com>]

Hi,
This breakpoint functions as any normal breakpoint in gdb with the only
added advantage that it can be invoked at a particular time. If in a
blocking syscall, the program execution would be interrupted at the
specified time as it would receive a SIGINT signal. Basically it would
function the same as having a normal break statement in gdb.

The use of this feature is in Real time systems which have a strict time
constraint and if we want to know how much the program has executed in a
specified time. It can also help in sampling data at a particular frequency.
Cheers.

Abhijit Halder wrote:
> 
> On Wed, Sep 7, 2011 at 11:12 PM, logitech <supriya.rao17@gmail.com> wrote:
>>
>> Hello,
>> I have attached a patch file which integrates timed breakpoints into gdb.
>> http://old.nabble.com/file/p32418104/patchfile.patch patchfile.patch
>> The format of the command is "break 'x's 'y'us" where 's' and 'us' stand
>> for
>> seconds and microseconds respectively. Eg. break 1s 0us
>> When the program execution begins, it breaks after 1sec. On continuing,
>> the
>> program resumes again and breaks after 1 sec.
>> To continuously break the program after a specific time interval a
>> separate
>> script can be written as shown :
>>
>> break 0s 500us
>> run
>> while(1)
>> p $pc
>> c
>> end
>> quit
>>
>> Further details about the implementation can be found here
>> http://www.youtube.com/watch?v=PAFQlxqI7qs&feature=player_embedded#! GDB
>> Timed breakpoint
>>
>> I would highly appreciate comments and feedback from your side regarding
>> the
>> same.
>> Thank you.
>>
>> --
>> View this message in context:
>> http://old.nabble.com/Integration-of-timed-breakpoints-into-GDB-tp32418104p32418104.html
>> Sent from the Sourceware - gdb-patches mailing list archive at
>> Nabble.com.
>>
>>
> 
> I have not yet seen the patch, but before that I would like to get
> answer of few questions:
>       1. What happen if your program execution is stuck inside a
> blocking syscall when it should break? I believe the time you are
> talking about is not a hard bound limit.
> 
>       2. What is the usecase of this feature?
> 
> -Abhijit
> 
> 

-- 
View this message in context: http://old.nabble.com/Integration-of-timed-breakpoints-into-GDB-tp32418104p32423483.html
Sent from the Sourceware - gdb-patches mailing list archive at Nabble.com.