On 04/02/2014 12:36 PM, Eric Blake wrote:
On 04/02/2014 09:54 AM, Cole Robinson wrote:
> Currently VolOpen notifies the user of a potentially non-fatal failure by
> returning -2 and logging a VIR_WARN or VIR_INFO. Unfortunately most
> callers treat -2 as fatal but don't actually report any message with
> the error APIs.
>
> Rename the VOL_OPEN_ERROR flag to VOL_OPEN_NOERROR. If NOERROR is specified,
> we preserve the current behavior of returning -2 (there's only one caller
> that wants this).
>
> However in the default case, only return -1, and actually use the error
> APIs. Fix up a couple callers as a result.
> ---
> src/storage/storage_backend.c | 77 ++++++++++++++++++++++++--------------
> src/storage/storage_backend.h | 7 ++--
> src/storage/storage_backend_fs.c | 28 +++++---------
> src/storage/storage_backend_scsi.c | 3 --
> 4 files changed, 62 insertions(+), 53 deletions(-)
> + * Returns -1 on error. If VIR_STORAGE_VOL_OPEN_NOERROR is passed, we
> + * return -2 if file mode is unexpected or the volume is a dangling
> + * symbolic link.
Seems fair; let's see how it works below...
> */
> int
> virStorageBackendVolOpen(const char *path, struct stat *sb,
> @@ -1288,9 +1289,10 @@ virStorageBackendVolOpen(const char *path, struct stat *sb,
> {
> int fd, mode = 0;
> char *base = last_component(path);
> + bool noerror = (flags * VIR_STORAGE_VOL_OPEN_NOERROR);
Whoops[1] - that returns true for any non-zero flags value. I think you
meant s/*/&/
Doh, sorry bout that. Fixed.
> } else {
> - VIR_WARN("ignoring unexpected type for file '%s'", path);
> VIR_FORCE_CLOSE(fd);
> - return -2;
> + if (noerror) {
> + VIR_WARN("ignoring unexpected type for file '%s'",
path);
> + return -2;
> + }
> + virReportError(VIR_ERR_INTERNAL_ERROR,
> + _("unexpected type for file '%s'"), path);
> + return -1;
...so far, so good (all cases of 'return -2' are guarded by noerror, and
an error is issued before returning -1).
> }
>
> if (virSetBlocking(fd, true) < 0) {
> + VIR_FORCE_CLOSE(fd);
> + if (noerror) {
> + VIR_WARN("unable to set blocking mode for '%s'",
path);
> + return -2;
> + }
But this one feels wrong[2]. If we get here, we KNOW the fd has the
expected type, and we successfully opened it. This is a case where the
system is hosed, and we should loudly and unconditionally fail with -1,
rather than returning -2 on noerror.
Agreed, I was just preserving the original behavior. Fixed to unconditionally
raise error and return -1.
> /* VolOpenCheckMode flags */
> enum {
> - VIR_STORAGE_VOL_OPEN_ERROR = 1 << 0, /* warn if unexpected type
> - * encountered */
> + VIR_STORAGE_VOL_OPEN_NOERROR = 1 << 0, /* don't error if unexpected
type
> + * encountered, just warn */
> VIR_STORAGE_VOL_OPEN_REG = 1 << 1, /* regular files okay */
> VIR_STORAGE_VOL_OPEN_BLOCK = 1 << 2, /* block files okay */
> VIR_STORAGE_VOL_OPEN_CHAR = 1 << 3, /* char files okay */
> VIR_STORAGE_VOL_OPEN_DIR = 1 << 4, /* directories okay */
Cosmetic, but alignment of = is now off [3].
Fixed.
> - if
(virStorageBackendUpdateVolTargetInfo(&vol->backingStore,
> - false,
> - VIR_STORAGE_VOL_OPEN_DEFAULT) < 0) {
> - /* The backing file is currently unavailable, the capacity,
> - * allocation, owner, group and mode are unknown. Just log the
> - * error and continue.
> - * Unfortunately virStorageBackendProbeTarget() might already
> - * have logged a similar message for the same problem, but only
> - * if AUTO format detection was used. */
> - virReportError(VIR_ERR_INTERNAL_ERROR,
> - _("cannot probe backing volume info: %s"),
> - vol->backingStore.path);
> - }
> + virStorageBackendUpdateVolTargetInfo(&vol->backingStore, false,
> + VIR_STORAGE_VOL_OPEN_DEFAULT);
> + /* If this failed, the backing file is currently unavailable,
> + * the capacity, allocation, owner, group and mode are unknown.
> + * An error message was raised, but we just continue. */
You may need an ignore_value() here to keep Coverity quiet about an
unchecked error return [4].
This touches code that I'm also working on, so I'm interested in getting
it in sooner to avoid rebase churn over repeated iterations. ACK if you
fix [1] and [2] above; and up to you whether changes are needed at [3]
and [4].
Fixed with ignore_value() and pushed now. Thanks Eric!
- Cole