Re: [PATCH] Fix regression on aarch64-linux gdbserver

[Tom de Vries <tdevries@suse.de>]

On 5/3/24 00:07, Mark Wielaard wrote:
> Hi,
> 
> On Mon, Apr 22, 2024 at 07:51:49AM -0600, Tom Tromey wrote:
>> diff --git a/gdbserver/linux-aarch64-low.cc b/gdbserver/linux-aarch64-low.cc
>> index 5df67fccd08..da5c1fd0629 100644
>> --- a/gdbserver/linux-aarch64-low.cc
>> +++ b/gdbserver/linux-aarch64-low.cc
>> @@ -576,41 +576,9 @@ aarch64_target::low_stopped_data_address ()
>>   
>>     /* Check if the address matches any watched address.  */
>>     state = aarch64_get_debug_reg_state (pid_of (current_thread));
>> -  for (i = aarch64_num_wp_regs - 1; i >= 0; --i)
>> -    {
>> -      const unsigned int offset
>> -	= aarch64_watchpoint_offset (state->dr_ctrl_wp[i]);
>> -      const unsigned int len = aarch64_watchpoint_length (state->dr_ctrl_wp[i]);
>> -      const CORE_ADDR addr_watch = state->dr_addr_wp[i] + offset;
>> -      const CORE_ADDR addr_watch_aligned = align_down (state->dr_addr_wp[i], 8);
>> -      const CORE_ADDR addr_orig = state->dr_addr_orig_wp[i];
>> -
>> -      if (state->dr_ref_count_wp[i]
>> -	  && DR_CONTROL_ENABLED (state->dr_ctrl_wp[i])
>> -	  && addr_trap >= addr_watch_aligned
>> -	  && addr_trap < addr_watch + len)
>> -	{
>> -	  /* ADDR_TRAP reports the first address of the memory range
>> -	     accessed by the CPU, regardless of what was the memory
>> -	     range watched.  Thus, a large CPU access that straddles
>> -	     the ADDR_WATCH..ADDR_WATCH+LEN range may result in an
>> -	     ADDR_TRAP that is lower than the
>> -	     ADDR_WATCH..ADDR_WATCH+LEN range.  E.g.:
>> -
>> -	     addr: |   4   |   5   |   6   |   7   |   8   |
>> -				   |---- range watched ----|
>> -		   |----------- range accessed ------------|
>> -
>> -	     In this case, ADDR_TRAP will be 4.
>> -
>> -	     To match a watchpoint known to GDB core, we must never
>> -	     report *ADDR_P outside of any ADDR_WATCH..ADDR_WATCH+LEN
>> -	     range.  ADDR_WATCH <= ADDR_TRAP < ADDR_ORIG is a false
>> -	     positive on kernels older than 4.10.  See PR
>> -	     external/20207.  */
>> -	  return addr_orig;
>> -	}
>> -    }
>> +  CORE_ADDR result;
>> +  if (aarch64_stopped_data_address (state, addr_trap, &result))
>> +    return result;
>>   
>>     return (CORE_ADDR) 0;
>>   }
> 
> This broke the build on the gdb-fedora-arm64 buildbot:
> https://builder.sourceware.org/buildbot/#/builders/181/builds/6402
> 
> Fix attached.
> 

Hi Mark,

thanks for the fix.

I ran into the build breaker on an aarch64-linux system, and managed to 
complete the build after applying your fix.

Pushed.

Thanks,
- Tom