On Sun, 02/09 10:48, Paolo Bonzini wrote:
> Signed-off-by: Paolo Bonzini <pbonz...@redhat.com>
> ---
> block/qcow.c | 16 ++++++----------
> 1 file changed, 6 insertions(+), 10 deletions(-)
>
> diff --git a/block/qcow.c b/block/qcow.c
> index 948b0c5..23bc691 100644
> --- a/block/qcow.c
> +++ b/block/qcow.c
> @@ -119,17 +119,19 @@ static int qcow_open(BlockDriverState *bs, QDict
> *options, int flags,
> if (header.version != QCOW_VERSION) {
> char version[64];
> snprintf(version, sizeof(version), "QCOW version %d",
> header.version);
> - qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
> - bs->device_name, "qcow", version);
> + error_set(errp, QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
> + bs->device_name, "qcow", version);
> ret = -ENOTSUP;
> goto fail;
> }
>
> if (header.size <= 1 || header.cluster_bits < 9) {
> + error_setg(errp, "invalid value in qcow header\n");
The convention is not adding "\n" in the end of error messages.
> ret = -EINVAL;
> goto fail;
> }
> if (header.crypt_method > QCOW_CRYPT_AES) {
> + error_setg(errp, "invalid encryption method in qcow header\n");
Same here.
Thanks,
Fam
