On 9/17/19 5:21 AM, Greg Kurz wrote:
> Ensure that hints are added even if errp is &error_fatal or &error_abort.
>
> Signed-off-by: Greg Kurz <[email protected]>
> ---
> nbd/client.c | 24 +++++++++++++-----------
> nbd/server.c | 7 +++++--
> 2 files changed, 18 insertions(+), 13 deletions(-)
>
> diff --git a/nbd/client.c b/nbd/client.c
> index b9dc829175f9..c6e6e4046fd5 100644
> --- a/nbd/client.c
> +++ b/nbd/client.c
> @@ -154,6 +154,7 @@ static int nbd_handle_reply_err(QIOChannel *ioc,
> NBDOptionReply *reply,
> bool strict, Error **errp)
> {
> g_autofree char *msg = NULL;
> + Error *local_err = NULL;
I'd prefer 'err' here...
>
> if (!(reply->type & (1 << 31))) {
> return 1;
> @@ -161,14 +162,14 @@ static int nbd_handle_reply_err(QIOChannel *ioc,
> NBDOptionReply *reply,
>
> if (reply->length) {
> if (reply->length > NBD_MAX_BUFFER_SIZE) {
> - error_setg(errp, "server error %" PRIu32
> + error_setg(&local_err, "server error %" PRIu32
so that &err doesn't change line lengths.
> case NBD_REP_ERR_SHUTDOWN:
> - error_setg(errp, "Server shutting down before option %" PRIu32 "
> (%s)",
> + error_setg(&local_err, "Server shutting down before option %" PRIu32
> " (%s)",
For example, here, you went beyond 80 columns.
At any rate, I'm assuming this will probably go in through Markus' error
tree as part of the whole series, rather than me needing to pick just
this patch through my NBD tree.
Whether or not you shorten the local variable name,
Reviewed-by: Eric Blake <[email protected]>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization: qemu.org | libvirt.org
