On 5/12/19 11:46 AM, Wensheng Tang wrote:
> (The last email does not include signed-off-by line. Please ignoreit)
>
> We should keep the error handling consistent. ENOMEDIUM is more meaningful than ENOENT a when driver cannot be loaded.
>
Consistent with what?
Why is it more meaningful?
I might argue that any one of ENODEV, EINVAL, ENOSYS, ENOTSUP, or
EPROTONOSUPPORT might be more meaningful than ENOENT or ENOMEDIUM, but
it depends on what people expect and why.
> Signed-off-by: Wensheng Tang <sp3478@gmail.com>
> ---
> block.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/block.c b/block.c
> index 5c2c6aa..6685be7 100644
> --- a/block.c
> +++ b/block.c
> @@ -770,7 +770,7 @@ static int find_image_format(BlockBackend *file, const char *filename,
> if (!drv) {
> error_setg(errp, "Could not determine image format: No compatible "
> "driver found");
> - ret = -ENOENT;
> + ret = -ENOMEDIUM;
> }
> *pdrv = drv;
> return ret;
> @@ -1619,7 +1619,7 @@ static int bdrv_fill_options(QDict **options, const char *filename,
> drv = bdrv_find_format(drvname);
> if (!drv) {
> error_setg(errp, "Unknown driver '%s'", drvname);
> - return -ENOENT;
> + return -ENOMEDIUM;
> }
> /* If the user has explicitly specified the driver, this choice should
> * override the BDRV_O_PROTOCOL flag */
> @@ -1655,7 +1655,7 @@ static int bdrv_fill_options(QDict **options, const char *filename,
> if (filename) {
> drv = bdrv_find_protocol(filename, parse_filename, errp);
> if (!drv) {
> - return -EINVAL;
> + return -ENOMEDIUM;
> }
>
> drvname = drv->format_name;
>