does it make sense to stop on SIGPRIO?

does it make sense to stop on SIGPRIO?

[Joel Brobecker]

I've been looking at how we decide what to when we receive a signal.
We have some code that disables stop&printing for various signals
because these signals are used as part of normal thread operations.

  /* These signals are used internally by user-level thread
     implementations.  (See signal(5) on Solaris.)  Like the above
     signals, a healthy program receives and handles them as part of
     its normal operation.  */

We do the same for other signals, which are not error signals:

  /* Signals that are not errors should not normally enter the debugger.  */

On LynxOS, changing the priority of a thread automatically causes
a SIGPRIO signal to be raised.  I think that SIGPRIO falls more
into the second category (not a signal used to indicate an error).

Are there any known situations where we would want a SIGPRIO would
be indicating something abnormal, or significant enough that we would
want to stop?

Thanks,
-- 
Joel

Re: does it make sense to stop on SIGPRIO?

[Mark Kettenis]

> Date: Wed, 5 Jan 2011 11:22:45 +0400
> From: Joel Brobecker <brobecker@adacore.com>
> 
> I've been looking at how we decide what to when we receive a signal.
> We have some code that disables stop&printing for various signals
> because these signals are used as part of normal thread operations.
> 
>   /* These signals are used internally by user-level thread
>      implementations.  (See signal(5) on Solaris.)  Like the above
>      signals, a healthy program receives and handles them as part of
>      its normal operation.  */
> 
> We do the same for other signals, which are not error signals:
> 
>   /* Signals that are not errors should not normally enter the debugger.  */
> 
> On LynxOS, changing the priority of a thread automatically causes
> a SIGPRIO signal to be raised.  I think that SIGPRIO falls more
> into the second category (not a signal used to indicate an error).
> 
> Are there any known situations where we would want a SIGPRIO would
> be indicating something abnormal, or significant enough that we would
> want to stop?

Given that SIGPRIO seems to be something rather un-UNIXy (OpenBSD,
Linux and Solaris don't seem to have it), I think you can do here
whatever you like ;).

Re: does it make sense to stop on SIGPRIO?

[Andreas Schwab]

Joel Brobecker <brobecker@adacore.com> writes:

> Are there any known situations where we would want a SIGPRIO would
> be indicating something abnormal, or significant enough that we would
> want to stop?

If SIGPRIO is ignored by default then it makes sense to pass it through
silently.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: does it make sense to stop on SIGPRIO?

[Michael Snyder]

Joel Brobecker wrote:
> I've been looking at how we decide what to when we receive a signal.
> We have some code that disables stop&printing for various signals
> because these signals are used as part of normal thread operations.
> 
>   /* These signals are used internally by user-level thread
>      implementations.  (See signal(5) on Solaris.)  Like the above
>      signals, a healthy program receives and handles them as part of
>      its normal operation.  */
> 
> We do the same for other signals, which are not error signals:
> 
>   /* Signals that are not errors should not normally enter the debugger.  */
> 
> On LynxOS, changing the priority of a thread automatically causes
> a SIGPRIO signal to be raised.  I think that SIGPRIO falls more
> into the second category (not a signal used to indicate an error).
> 
> Are there any known situations where we would want a SIGPRIO would
> be indicating something abnormal, or significant enough that we would
> want to stop?
> 
> Thanks,


I think it might be peculiar to LynxOS.  Most google hits either refer 
to gdb or Lynx.

Re: does it make sense to stop on SIGPRIO?

[Michael Snyder]

Michael Snyder wrote:
> Joel Brobecker wrote:
>> I've been looking at how we decide what to when we receive a signal.
>> We have some code that disables stop&printing for various signals
>> because these signals are used as part of normal thread operations.
>>
>>   /* These signals are used internally by user-level thread
>>      implementations.  (See signal(5) on Solaris.)  Like the above
>>      signals, a healthy program receives and handles them as part of
>>      its normal operation.  */
>>
>> We do the same for other signals, which are not error signals:
>>
>>   /* Signals that are not errors should not normally enter the debugger.  */
>>
>> On LynxOS, changing the priority of a thread automatically causes
>> a SIGPRIO signal to be raised.  I think that SIGPRIO falls more
>> into the second category (not a signal used to indicate an error).
>>
>> Are there any known situations where we would want a SIGPRIO would
>> be indicating something abnormal, or significant enough that we would
>> want to stop?
>>
>> Thanks,
> 
> 
> I think it might be peculiar to LynxOS.  Most google hits either refer 
> to gdb or Lynx.
> 
> 

Meant to imply -- in which case you can do what you like.

;-)