[Qemu-devel] Re: [PATCH 0/5] Debugger enhancements

[Jan Kiszka <jan.kiszka@web.de>]

[-- Attachment #1: Type: text/plain, Size: 1453 bytes --]

Fabrice Bellard wrote:
> Jan Kiszka wrote:
>> Fabrice Bellard wrote:
>>> Hi,
>>>
>>> I cannot accept the patches for several reasons:
>>>
>>> 1) You mix cosmetic and functional patches.
>> Do you have specific hunks in mind? I'm a bit blind ATM, not seeing
>> where I changed coding style or naming for cosmetic reasons.
> 
> You renammed mem_write_pc and mem_write_vaddr.

mem -> iomem dated back when I was confused about plain RAM access going
through handlers that are called io_read/write. Reverted this. But write
-> access is required as I changed the semantics of these variables.

> 
> BTW, why did you add 'len' and 'type' parameters to breakpoints ?
> 
> I don't think it is a good idea to say that breakpoints/watchpoints
> apply to all processors. Such behavior should be handled at a higher level.

Yes, I agree meanwhile. Now I introduced a host-injection layer to
gdbstub that handles this. All cpu_* services are per-CPU again.

> 
> It would also be interesting if the watchpoint/breakpoint implementation
> could be used to implement CPU watchpoints and breakpoints (I am
> thinking about the x86 DRx registers here).

Good point! And a hard argument to keep per-CPU breakpoint/watchpoint
lists. I've already prepared my patches to tell host-injected
breakpoints apart from guest-injected ones. Will now look into at least
basic debug register support of x86 (which was on my list anyway).

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 254 bytes --]