Re: [alsa-devel] [RFC PATCH 4/7] sound: core: Avoid using timespec for struct snd_rawmidi_status

Hi Arnd,

On 21 September 2017 at 20:56, Arnd Bergmann arnd@arndb.de wrote:

On Thu, Sep 21, 2017 at 8:18 AM, Baolin Wang baolin.wang@linaro.org wrote:

...

•   case SNDRV_RAWMIDI_IOCTL_STATUS:
  

+#if __BITS_PER_LONG == 32

•   case SNDRV_RAWMIDI_IOCTL_STATUS32:
  

•   {
  

•           int err = 0;
  

•           struct snd_rawmidi_status32 __user *status = argp;
  

•           struct snd_rawmidi_status32 status32;
  

•           struct snd_rawmidi_status64 status64;
  

•           if (copy_from_user(&status32, argp,
  

•                              sizeof(struct snd_rawmidi_status32)))
  

•                   return -EFAULT;
  

•           switch (status32.stream) {
  

•           case SNDRV_RAWMIDI_STREAM_OUTPUT:
  

•                   if (rfile->output == NULL)
  

•                           return -EINVAL;
  

•                   err = snd_rawmidi_output_status(rfile->output, &status64);
  

•                   break;
  

•           case SNDRV_RAWMIDI_STREAM_INPUT:
  

•                   if (rfile->input == NULL)
  

•                           return -EINVAL;
  

•                   err = snd_rawmidi_input_status(rfile->input, &status64);
  

•                   break;
  

•           default:
  

•                   return -EINVAL;
  

•           }
  

•           if (err < 0)
  

•                   return err;
  

•           if (put_user(status64.stream, &status->stream) ||
  

•               put_user(status64.tstamp.tv_sec, &status->tstamp.tv_sec) ||
  

•               put_user(status64.tstamp.tv_nsec, &status->tstamp.tv_nsec) ||
  

•               put_user(status64.avail, &status->avail) ||
  

•               put_user(status64.xruns, &status->xruns))
  

•                   return -EFAULT;
  

•           return 0;
  

•   }
  

This follows the existing coding style for the other functions, but I think it would be nicer to express the last part as

   status32 = (struct snd_rawmidi_status32) {
           .stream = status->stream,
           .tstamp.tv_sec, &status->tstamp.tv_sec,
           .tstamp.tv_nsec, &status->tstamp.tv_nsec,
           .avail, &status->avail,
           .xruns, &status->xruns,
   };
   if (copy_to_user(status, &status32, sizeof(*status))
          return -EFAULT;
   return 0;

It's completely equivalent, I just find my version easier to read, and it should produce slightly better object code.

Maybe the maintainers have a preference, or there might be a good reason to use the series of put_user() instead.

I just saw there are not many put_user() will be used in this function, but I agree with you and I like to change as you suggested.