Re: [RFC PATCH] disas/hppa: drop raw opcode dump

From: "Alex Bennée" <alex.bennee@linaro.org>
To: Helge Deller <deller@gmx.de>
Cc: qemu-devel@nongnu.org,  Richard Henderson <richard.henderson@linaro.org>
Subject: Re: [RFC PATCH] disas/hppa: drop raw opcode dump
Date: Thu, 29 Feb 2024 16:41:30 +0000	[thread overview]
Message-ID: <87wmqn6x51.fsf@draig.linaro.org> (raw)
In-Reply-To: <83e2986e-3c80-4cf3-93ae-5e578051c98b@gmx.de> (Helge Deller's message of "Thu, 29 Feb 2024 16:02:09 +0100")

Helge Deller <deller@gmx.de> writes:

> On 2/29/24 15:05, Alex Bennée wrote:
>> The hppa disassembly is different from the others due to leading with
>> the raw opcode data. This confuses plugins looking for instruction
>> prefixes to match instructions. For plugins like execlog there is
>> another mechanism for getting the instruction byte data.
>>
>> For the sake of consistently just present the instruction assembly
>> code.
>>
>> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
>
> This effectively reverts commit 2f926bfd5b79e6219ae65a1e530b38f37d62b384
> ("disas/hppa: Show hexcode of instruction along with disassembly").
>
> Sad, but Ok.
>
> Acked-by: Helge Deller <deller@gmx.de>

Well its an RFC for a reason. If this is useful we could just sneak a
flag into disassemble_info that is set by plugin_disas? Or maybe move
the insn stream to afterwards?

>
>
>> ---
>>   disas/hppa.c | 4 ----
>>   1 file changed, 4 deletions(-)
>>
>> diff --git a/disas/hppa.c b/disas/hppa.c
>> index 22dce9b41bb..dd34cce211b 100644
>> --- a/disas/hppa.c
>> +++ b/disas/hppa.c
>> @@ -1972,10 +1972,6 @@ print_insn_hppa (bfd_vma memaddr, disassemble_info *info)
>>
>>     insn = bfd_getb32 (buffer);
>>
>> -  info->fprintf_func(info->stream, " %02x %02x %02x %02x   ",
>> -                (insn >> 24) & 0xff, (insn >> 16) & 0xff,
>> -                (insn >>  8) & 0xff, insn & 0xff);
>> -
>>     for (i = 0; i < NUMOPCODES; ++i)
>>       {
>>         const struct pa_opcode *opcode = &pa_opcodes[i];

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro