From: Richard Henderson <rth@twiddle.net>
To: Peter Maydell <peter.maydell@linaro.org>,
Peter Crosthwaite <crosthwaitepeter@gmail.com>
Cc: "claudio.fontana" <claudio.fontana@gmail.com>,
Peter Crosthwaite <crosthwaite.peter@gmail.com>,
QEMU Developers <qemu-devel@nongnu.org>,
Alexander Graf <agraf@suse.de>,
Edgar Iglesias <edgari@xilinx.com>,
"Edgar E. Iglesias" <edgar.iglesias@gmail.com>
Subject: Re: [Qemu-devel] [PATCH 6/7] monitor: "i": Add ARM specifics
Date: Tue, 05 May 2015 10:43:30 -0700 [thread overview]
Message-ID: <55490142.5030901@twiddle.net> (raw)
In-Reply-To: <CAFEAcA94orSvGJ8BQ2=MeXGwEytKhYzmWmzPoBHZqByMVZSX4w@mail.gmail.com>
On 05/05/2015 10:19 AM, Peter Maydell wrote:
> On 5 May 2015 at 05:45, Peter Crosthwaite <crosthwaitepeter@gmail.com> wrote:
>> Add the ARM specific disassembly flags setup, so ARM can be correctly
>> disassembled from the monitor.
>>
>> Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
>> ---
>> monitor.c | 11 +++++++++++
>> 1 file changed, 11 insertions(+)
>>
>> diff --git a/monitor.c b/monitor.c
>> index d831d98..9d9f1e2 100644
>> --- a/monitor.c
>> +++ b/monitor.c
>> @@ -1217,6 +1217,17 @@ static void memory_dump(Monitor *mon, int count, int format, int wsize,
>> int flags;
>> flags = 0;
>> env = mon_get_cpu();
>> +#ifdef TARGET_ARM
>> + if (env->thumb) {
>> + flags |= 1;
>> + }
>> + if (env->bswap_code) {
>> + flags |= 2;
>> + }
>> + if (env->aarch64) {
>> + flags |= 4;
>> + }
>> +#endif
>
> monitor.c has no business poking around in the CPU state
> internals like this... You probably want a CPU method
> get_disas_flags() or something.
>
> -- PMM
>
While this patch set does improve the current dismal state of affairs, I think
the ideal solution is a cpu method that takes care of all the disassembly info
setup.
Indeed, the flags setup becomes less obscure when, instead of
#ifdef TARGET_I386
if (wsize == 2) {
flags = 1;
} else if (wsize == 4) {
flags = 0;
} else {
/* as default we use the current CS size */
flags = 0;
if (env) {
#ifdef TARGET_X86_64
if ((env->efer & MSR_EFER_LMA) &&
(env->segs[R_CS].flags & DESC_L_MASK))
flags = 2;
else
#endif
if (!(env->segs[R_CS].flags & DESC_B_MASK))
flags = 1;
}
}
in one place and
#if defined(TARGET_I386)
if (flags == 2) {
s.info.mach = bfd_mach_x86_64;
} else if (flags == 1) {
s.info.mach = bfd_mach_i386_i8086;
} else {
s.info.mach = bfd_mach_i386_i386;
}
print_insn = print_insn_i386;
in another, we merge the two so that we get
s.info.mach = bfd_mach_i386_i8086;
if (env->hflags & (1U << HF_CS32_SHIFT)) {
s.info.mach = bfd_mach_i386_i386;
}
#ifdef TARGET_X86_64
if (env->hflags & (1U << HF_CS64_SHIFT)) {
s.info.mach = bfd_mach_x86_64;
}
#endif
I'm not sure what the right interface for this is, exactly. But I'd think that
the CPUDebug structure would be initialized in the caller -- target or monitor
-- while the mach and whatnot get filled in by the cpu hook. Maybe even have
the cpu hook return the print_insn function to use.
r~
next prev parent reply other threads:[~2015-05-05 17:43 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-05 4:44 [Qemu-devel] [PATCH 0/7] disas: Unify target_disas and monitor_disas Peter Crosthwaite
2015-05-05 4:44 ` [Qemu-devel] [PATCH 1/7] disas: Create factored out fn for monitor and target disas Peter Crosthwaite
2015-05-05 14:45 ` Claudio Fontana
2015-05-05 4:44 ` [Qemu-devel] [PATCH 2/7] disas: microblaze: Migrate setup to common code Peter Crosthwaite
2015-05-05 4:45 ` [Qemu-devel] [PATCH 3/7] disas: cris: Fix 0 buffer length case Peter Crosthwaite
2015-05-05 4:45 ` [Qemu-devel] [PATCH 4/7] disas: cris: Migrate setup to common code Peter Crosthwaite
2015-05-05 4:45 ` [Qemu-devel] [PATCH 5/7] disas: arm-a64: Make printfer and stream variable Peter Crosthwaite
2015-05-05 14:41 ` Claudio Fontana
2015-05-05 17:22 ` Richard Henderson
2015-05-05 4:45 ` [Qemu-devel] [PATCH 6/7] monitor: "i": Add ARM specifics Peter Crosthwaite
2015-05-05 14:40 ` Claudio Fontana
2015-05-05 17:19 ` Peter Maydell
2015-05-05 17:43 ` Richard Henderson [this message]
2015-05-06 6:57 ` Peter Crosthwaite
2015-05-06 13:57 ` Richard Henderson
2015-05-06 7:06 ` Peter Crosthwaite
2015-05-06 14:12 ` Richard Henderson
2015-05-06 14:17 ` Paolo Bonzini
2015-05-06 14:32 ` Stefano Stabellini
2015-05-06 15:44 ` Peter Maydell
2015-05-06 18:24 ` Richard Henderson
2015-05-06 18:39 ` Peter Maydell
2015-05-05 4:45 ` [Qemu-devel] [PATCH 7/7] disas: arm: Use target_disas impl for monitor Peter Crosthwaite
2015-05-05 14:38 ` Claudio Fontana
2015-05-05 16:48 ` Peter Crosthwaite
2015-05-05 17:18 ` [Qemu-devel] [PATCH 0/7] disas: Unify target_disas and monitor_disas Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55490142.5030901@twiddle.net \
--to=rth@twiddle.net \
--cc=agraf@suse.de \
--cc=claudio.fontana@gmail.com \
--cc=crosthwaite.peter@gmail.com \
--cc=crosthwaitepeter@gmail.com \
--cc=edgar.iglesias@gmail.com \
--cc=edgari@xilinx.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.