This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch, testsuite] gdb.base/savedregs.exp: SIGSEGV -> SIGILL


On 06/09/2011 07:41 PM, Pedro Alves wrote:
> On Thursday 09 June 2011 12:17:27, Mark Kettenis wrote:
>>> Date: Thu, 09 Jun 2011 17:28:09 +0800
>>> From: Yao Qi <yao@codesourcery.com>
>>>
>>> In current gdb.base/savedregs.exp, signal handler is installed for
>>> signal SIGSEGV, and SIGSEGV is trigger by `*(char *)0 = 0;'.  However,
>>> on non-mmu uclinux system, writing to an address 0x0 doesn't trigger
>>> SIGSEGV.
>>>
>>> In my patch, SIGILL is chosen to replace SIGSEGV.  One assumption here
>>> is that 0xffff is an invalid instruction on all ports.
>>
>> Please don't do this.  You're changing the test significantly.  And
>> there is no guarantee that 0xffff is an invalid instruction.  Heck
>> most platforms don't even have 16-bit instructions.
> 
> If backtracing through 0 is important to this test (haven't looked),
> and replacing the write to 0 by raise(SIGSEGV) won't cut it,
> then you can use the same trick sigbpt.exp, signest.exp, signull.exp
> use to skip the test on targets without an MMU.
> 

Pedro,
Thanks for this trick, which I don't know before.  SIGSEGV usage in
savedregs.exp is a little bit different from them.  In savedregs.exp,
SIGSEGV is used to trigger invocation of signal handler, and check
registers contents in signal trampoline frame.  If my understand is
correct, I can't see any reason that we can't replace SIGSEGV by SIGILL,
or other signal.

Of course, we can skip savedregs.exp with the same trick, but it is
imperfect that we skip a test which can be run naturally.

-- 
Yao (éå)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]