Re: [PATCH] fix PR 15180 "May only run agent-printf on the target"

[Pedro Alves <palves at redhat dot com>]

On 07/01/2013 09:51 AM, Hui Zhu wrote:
> On Mon, Jul 1, 2013 at 4:10 PM, Pedro Alves <palves@redhat.com> wrote:
>> On 06/30/2013 04:17 PM, Hui Zhu wrote:
>>>>> If dprintf style is agent, then the target doesn't report events
>>>>> for such hits.  If we happen to see a trap reported for the
>>>>> same address as another breakpoint, the trap is surely not explained
>>>>> by the dprintf.  IOW, we shouldn't have a bpstat a dprintf w/ style
>>>>> agent at all.  What is need is to override the breakpoint_hit method
>>>>> of dprintf_breakpoint_ops to always return false for agent style dprintfs.
>>>>>
>>>>> Or maybe I'm missing something.  I have a hard time going from
>>>>> the bug description in the PR to the patch.
>>> OK.  I add dprintf_check_status in the new patch according to your comments.
>>>
>>
>> But I suggested the breakpoint_ops->breakpoint_hit not breakpoint_ops->check_status.
>> An agent-style dprintf is just like a tracepoint.  It can't ever explain
>> a stop.  See tracepoint_breakpoint_hit.  If it didn't work, please explain why.
> 
> OK.  Post a new version according to your comments.
> Please help me review it.

So I've now picked the patches from
<http://sourceware.org/ml/gdb-patches/2013-06/msg00367.html>

and then applied these new dprint-same-addr patches on top.

They completely break gdbserver.  :-/  First I tried:

 $ make check RUNTESTFLAGS="--target_board=native-gdbserver dprintf-same-addr.exp"
 ...
 Running target native-gdbserver
 Using ../../../src/gdb/testsuite/boards/../boards/native-gdbserver.exp as board description file for target.
 Using ../../../src/gdb/testsuite/config/gdbserver.exp as tool-and-target-specific interface file.
 Running ../../../src/gdb/testsuite/gdb.base/dprintf-same-addr.exp ...

                 === gdb Summary ===

 /home/pedro/gdb/mygit/build/gdb/testsuite/../../gdb/gdb version  7.6.50.20130701-cvs -nw -nx -data-directory /home/pedro/gdb/mygit/build/gdb/testsuite/../data-directory

 make[1]: Leaving directory `/home/pedro/gdb/mygit/build/gdb/testsuite'
 $

No FAILs/PASSes/UNSUPPORTED/etc., no nothing, is odd.  Turns out
that the program isn't even stopping at main.  That should have
resulted in a FAIL.  (Please fix that.)

Trying some other existing test shows:

 Running target native-gdbserver
 Using ../../../src/gdb/testsuite/boards/../boards/native-gdbserver.exp as board description file for target.
 Using ../../../src/gdb/testsuite/config/gdbserver.exp as tool-and-target-specific interface file.
 Running ../../../src/gdb/testsuite/gdb.base/break.exp ...
 FAIL: gdb.base/break.exp: run until function breakpoint
 FAIL: gdb.base/break.exp: run until breakpoint set at a line number (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(6) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(5) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(4) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(3) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(2) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until file:function(1) breakpoint (the program is no longer running)
 FAIL: gdb.base/break.exp: run until quoted breakpoint (the program is no longer running)

...

I didn't bother to run the whole testsuite.

At least with native testing I got:

$ make check RUNTESTFLAGS="dprintf-same-addr.exp"
...
# of expected passes            29
# of unsupported tests          1
...

-- 
Pedro Alves