Re: [Qemu-devel] [PATCH 4/4] ivshmem: check the value returned by fstat()

[zhanghailiang <zhang.zhanghailiang@huawei.com>]

On 2014/8/5 14:50, Levente Kurusa wrote:
> ----- Original Message -----
>> Hi Michael,
>> Thanks for your review of this patch!
>>
>>> On Mon, Aug 04, 2014 at 04:25:44PM +0800, zhanghailiang wrote:
>>>> The function fstat() may fail, so check its return value.
>>>>
>>>> Signed-off-by: zhanghailiang<zhang.zhanghailiang@huawei.com>
>>>> ---
>>>>    hw/misc/ivshmem.c | 5 ++++-
>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/hw/misc/ivshmem.c b/hw/misc/ivshmem.c
>>>> index 768e528..2667e9f 100644
>>>> --- a/hw/misc/ivshmem.c
>>>> +++ b/hw/misc/ivshmem.c
>>>> @@ -324,7 +324,10 @@ static int check_shm_size(IVShmemState *s, int fd) {
>>>>
>>>>        struct stat buf;
>>>>
>>>> -    fstat(fd,&buf);
>>>> +    if (fstat(fd,&buf)<   0) {
>>>> +        fprintf(stderr, "Cannot stat IVSHMEM: %s\n", strerror(errno));
>>>> +        return -1;
>>>> +    }
>>>>
>>>
>>> That's a confusing error message:
>>> 1. You don't stat ivshmem. You stat a shmem fd. Also best to print fd #.
>> I will add this in next version of the patch.
>>> 2. Tell the user what action was taken, e.g. IVSHMEM failed to start.
>>>
>>>>        if (s->ivshmem_size>   buf.st_size) {
>>>>            fprintf(stderr,
>>>> --
>> I have check the places of calling this function, And found that, if
>> this function return -1, qemu will call exit(-1). One of the callers is
>> ivshmem_read(), the purpose of check_shm_size() is to forbid guest to
>> map more memory than the object has allocated. So here is it suitable to
>> return -1 if fstat() failed? Or just give a warning message and return 0?
>> what's your opinion? Thanks.
>
> Yes, please return -1. All the errors fstat(2) can return in this scenario
> will be fatal to ivshmem.
>
> Exiting is probably not the best way to go, but I'm working on a fix for
> that at the moment, and for now, let's just exit, like all the other error
> paths do.
>
> Thanks,
> Levente Kurusa
>

OK,thanks!

Best regards,
zhanghailiang