Re: [PATCH v2] linux-user: Add guest memory layout to exception dump

[Laurent Vivier <laurent@vivier.eu>]

Le 21/10/2022 à 18:21, Helge Deller a écrit :
> On 10/21/22 17:43, Laurent Vivier wrote:
>> Le 21/10/2022 à 16:57, Laurent Vivier a écrit :
>>> Le 27/09/2022 à 18:58, Helge Deller a écrit :
>>>> When the emulation stops with a hard exception it's very useful for
>>>> debugging purposes to dump the current guest memory layout (for an
>>>> example see /proc/self/maps) beside the CPU registers.
>>>>
>>>> The open_self_maps() function provides such a memory dump, but since
>>>> it's located in the syscall.c file, various changes (add #includes, make
>>>> this function externally visible, ...) are needed to be able to call it
>>>> from the existing EXCP_DUMP() macro.
>>>>
>>>> This patch takes another approach by re-defining EXCP_DUMP() to call
>>>> target_exception_dump(), which is in syscall.c, consolidates the log
>>>> print functions and allows to add the call to dump the memory layout.
>>>>
>>>> Beside a reduced code footprint, this approach keeps the changes across
>>>> the various callers minimal, and keeps EXCP_DUMP() highlighted as
>>>> important macro/function.
>>>>
>>>> Signed-off-by: Helge Deller <deller@gmx.de>
>>>>
>>>> ---
>>>>
>>>> v2:
>>>> Based on feedback by Philippe Mathieu-Daudé, renamed the two functions
>>>> to excp_dump_file() and target_exception_dump(), and #define'ed
>>>> EXCP_DUMP() to target_exception_dump().
>>>> I intentionally did not replace all occurences of EXCP_DUMP() by
>>>> target_exception_dump() as I think it's unneccesary and not beneficial.
>>>> If this is really wished, I will send a v3.
>>>>
>>>>
>>>> diff --git a/linux-user/cpu_loop-common.h b/linux-user/cpu_loop-common.h
>>>> index 36ff5b14f2..e644d2ef90 100644
>>>> --- a/linux-user/cpu_loop-common.h
>>>> +++ b/linux-user/cpu_loop-common.h
>>>> @@ -23,18 +23,9 @@
>>>>   #include "exec/log.h"
>>>>   #include "special-errno.h"
>>>>
>>>> -#define EXCP_DUMP(env, fmt, ...)                                        \
>>>> -do {                                                                    \
>>>> -    CPUState *cs = env_cpu(env);                                        \
>>>> -    fprintf(stderr, fmt , ## __VA_ARGS__);                              \
>>>> -    fprintf(stderr, "Failing executable: %s\n", exec_path);             \
>>>> -    cpu_dump_state(cs, stderr, 0);                                      \
>>>> -    if (qemu_log_separate()) {                                          \
>>>> -        qemu_log(fmt, ## __VA_ARGS__);                                  \
>>>> -        qemu_log("Failing executable: %s\n", exec_path);                \
>>>> -        log_cpu_state(cs, 0);                                           \
>>>> -    }                                                                   \
>>>> -} while (0)
>>>> +void target_exception_dump(CPUArchState *env, const char *fmt, int code);
>>>> +#define EXCP_DUMP(env, fmt, code) \
>>>> +    target_exception_dump(env, fmt, code)
>>>>
>>>>   void target_cpu_copy_regs(CPUArchState *env, struct target_pt_regs *regs);
>>>>   #endif
>>>> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
>>>> index 2e954d8dbd..7d29c4c396 100644
>>>> --- a/linux-user/syscall.c
>>>> +++ b/linux-user/syscall.c
>>>> @@ -158,6 +158,7 @@
>>>>   #include "qapi/error.h"
>>>>   #include "fd-trans.h"
>>>>   #include "tcg/tcg.h"
>>>> +#include "cpu_loop-common.h"
>>>>
>>>>   #ifndef CLONE_IO
>>>>   #define CLONE_IO                0x80000000      /* Clone io context */
>>>> @@ -8144,6 +8145,33 @@ static int is_proc_myself(const char *filename, const char *entry)
>>>>       return 0;
>>>>   }
>>>>
>>>> +static void excp_dump_file(FILE *logfile, CPUArchState *env,
>>>> +                      const char *fmt, int code)
>>>> +{
>>>> +    if (logfile) {
>>>> +        CPUState *cs = env_cpu(env);
>>>> +
>>>> +        fprintf(logfile, fmt, code);
>>>> +        fprintf(logfile, "Failing executable: %s\n", exec_path);
>>>> +        cpu_dump_state(cs, logfile, 0);
>>>> +        open_self_maps(env, fileno(logfile));
>>>> +    }
>>>> +}
>>>> +
>>>> +void target_exception_dump(CPUArchState *env, const char *fmt, int code)
>>>> +{
>>>> +    /* dump to console */
>>>> +    excp_dump_file(stderr, env, fmt, code);
>>>> +
>>>> +    /* dump to log file */
>>>> +    if (qemu_log_separate()) {
>>>> +        FILE *logfile = qemu_log_trylock();
>>>> +
>>>> +        excp_dump_file(logfile, env, fmt, code);
>>>> +        qemu_log_unlock(logfile);
>>>> +    }
>>>> +}
>>>> +
>>>>   #if HOST_BIG_ENDIAN != TARGET_BIG_ENDIAN || \
>>>>       defined(TARGET_SPARC) || defined(TARGET_M68K) || defined(TARGET_HPPA)
>>>>   static int is_proc(const char *filename, const char *entry)
>>>>
>>>
>>> Applied to my linux-user-for-7.2 branch.
>>
>> This breaks build with:
>>
>> .../linux-user/i386/cpu_loop.c: In function 'cpu_loop':
>> ...linux-user/i386/cpu_loop.c:312:39: error: macro "EXCP_DUMP" passed 4 arguments, but takes just 3
>>    312 |                       (long)pc, trapnr);
>>        |                                       ^
> 
> This is because of this line:
>             EXCP_DUMP(env, "qemu: 0x%08lx: unhandled CPU exception 0x%x - aborting\n",
>                        (long)pc, trapnr);
> 
> I wonder if it is ok to drop the pc value. It should be printed in the register dump
> anyway.
> If Ok, should I send a new v3 patch, or a patch in front of my v2 patch?

I think you can drop the PC value. Send a v3 please.

Thanks,
Laurent