Re: [PATCH v4 2/4] sysemu/os-win32: fix setjmp/longjmp on windows-arm64

["Philippe Mathieu-Daudé" <philmd@linaro.org>]

On 22/2/23 17:08, Pierrick Bouvier wrote:
> On 2/21/23 23:27, Philippe Mathieu-Daudé wrote:
>> On 21/2/23 16:30, Pierrick Bouvier wrote:
>>> Windows implementation of setjmp/longjmp is done in
>>> C:/WINDOWS/system32/ucrtbase.dll. Alas, on arm64, it seems to *always*
>>> perform stack unwinding, which crashes from generated code.
>>>
>>> By using alternative implementation built in mingw, we avoid doing stack
>>> unwinding and this fixes crash when calling longjmp.
>>>
>>> Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
>>> Acked-by: Richard Henderson <richard.henderson@linaro.org>
>>> ---
>>>    include/sysemu/os-win32.h | 28 ++++++++++++++++++++++++----
>>>    meson.build               | 21 +++++++++++++++++++++
>>>    2 files changed, 45 insertions(+), 4 deletions(-)
>>
>>
>>> -#if defined(_WIN64)
>>> -/* On w64, setjmp is implemented by _setjmp which needs a second 
>>> parameter.
>>> +#if defined(__aarch64__)
>>> +/*
>>> + * On windows-arm64, setjmp is available in only one variant, and 
>>> longjmp always
>>> + * does stack unwinding. This crash with generated code.
>>> + * Thus, we use another implementation of setjmp (not windows one), 
>>> coming from
>>> + * mingw, which never performs stack unwinding.
>>> + */
>>> +#undef setjmp
>>> +#undef longjmp
>>> +/*
>>> + * These functions are not declared in setjmp.h because __aarch64__ 
>>> defines
>>> + * setjmp to _setjmpex instead. However, they are still defined in 
>>> libmingwex.a,
>>> + * which gets linked automatically.
>>> + */
>>> +extern int __mingw_setjmp(jmp_buf);
>>> +extern void __attribute__((noreturn)) __mingw_longjmp(jmp_buf, int);
>>> +#define setjmp(env) __mingw_setjmp(env)
>>> +#define longjmp(env, val) __mingw_longjmp(env, val)
>>> +#elif defined(_WIN64)
>>> +/*
>>> + * On windows-x64, setjmp is implemented by _setjmp which needs a 
>>> second parameter.
>>>     * If this parameter is NULL, longjump does no stack unwinding.
>>>     * That is what we need for QEMU. Passing the value of register 
>>> rsp (default)
>>> - * lets longjmp try a stack unwinding which will crash with 
>>> generated code. */
>>> + * lets longjmp try a stack unwinding which will crash with 
>>> generated code.
>>> + */
>>>    # undef setjmp
>>>    # define setjmp(env) _setjmp(env, NULL)
>>> -#endif
>>> +#endif /* __aarch64__ */
>>
>> This comment is confusing, the previous if ladder is about i86. Maybe
>> better not add any comment?
> 
> If I am not mistaken, before we had:
> 
> #if x64
> define setjmp as _setjmp(env, 0)
> #endif
> // nothing done for x86
> 
> and now we have:
> 
> #if aarch64
> define setjmp as __mingw_setjmp
> define longjmp as __mingw_longjmp
> #elif x64
> define setjmp as _setjmp(env, 0)
> #endif
> // nothing done for x86
> 
> Maybe the patch format is confusing, or I missed what you pointed.

Oh OK, we are good then, sorry :)