On 06/26/2018 05:40 PM, Paolo Bonzini wrote:
> Let management know if there were any problems communicating with
> qemu-pr-helper. The event is edge-triggered, and is sent every time
> the connection status of the pr-manager-helper object changes.
>
> Signed-off-by: Paolo Bonzini <pbonz...@redhat.com>
> ---
> qapi/block.json | 24 ++++++++++++++++++++++++
> scsi/pr-manager-helper.c | 25 +++++++++++++++++++------
> 2 files changed, 43 insertions(+), 6 deletions(-)
>
> diff --git a/qapi/block.json b/qapi/block.json
> index dc3323c954..1882a8d107 100644
> --- a/qapi/block.json
> +++ b/qapi/block.json
> @@ -334,6 +334,30 @@
> { 'event': 'DEVICE_TRAY_MOVED',
> 'data': { 'device': 'str', 'id': 'str', 'tray-open': 'bool' } }
>
> +##
> +# @PR_MANAGER_STATUS_CHANGED:
> +#
> +# Emitted whenever the connected status of a persistent reservation
> +# manager changes.
> +#
> +# @id: The QOM path of the PR manager object
> +#
> +# @connected: true if the PR manager is connected to a backend
> +#
> +# Since: 2.12
> +#
> +# Example:
> +#
> +# <- { "event": "PR_MANAGER_STATUS_CHANGED",
> +# "data": { "id": "/object/pr-helper0",
> +# "connected": true
> +# },
> +# "timestamp": { "seconds": 1519840375, "microseconds": 450486 } }
> +#
> +##
> +{ 'event': 'PR_MANAGER_STATUS_CHANGED',
> + 'data': { 'id': 'str', 'connected': 'bool' } }
> +
> ##
> # @QuorumOpType:
> #
> diff --git a/scsi/pr-manager-helper.c b/scsi/pr-manager-helper.c
> index b11481be9e..6643caf4cf 100644
> --- a/scsi/pr-manager-helper.c
> +++ b/scsi/pr-manager-helper.c
> @@ -17,6 +17,7 @@
> #include "io/channel.h"
> #include "io/channel-socket.h"
> #include "pr-helper.h"
> +#include "qapi/qapi-events-block.h"
>
> #include <scsi/sg.h>
>
> @@ -33,6 +34,7 @@ typedef struct PRManagerHelper {
> PRManager parent;
>
> char *path;
> + char *qom_path;
>
> QemuMutex lock;
> QIOChannel *ioc;
> @@ -127,6 +129,8 @@ static int pr_manager_helper_initialize(PRManagerHelper
> *pr_mgr,
> goto out_close;
> }
>
> + qapi_event_send_pr_manager_status_changed(pr_mgr->qom_path, true,
> + &error_abort);
> return 0;
Also, can we emit this earlier? For instance at the two places where we
unset pr_mgr->ioc:
diff --git i/scsi/pr-manager-helper.c w/scsi/pr-manager-helper.c
index 6643caf4cf..4ca15dff3f 100644
--- i/scsi/pr-manager-helper.c
+++ w/scsi/pr-manager-helper.c
@@ -48,6 +48,8 @@ static int pr_manager_helper_read(PRManagerHelper *pr_mgr,
if (r < 0) {
object_unref(OBJECT(pr_mgr->ioc));
+ qapi_event_send_pr_manager_status_changed(pr_mgr->qom_path, false,
+ &error_abort);
pr_mgr->ioc = NULL;
return -EINVAL;
}
@@ -74,6 +76,8 @@ static int pr_manager_helper_write(PRManagerHelper *pr_mgr,
assert(n_written != QIO_CHANNEL_ERR_BLOCK);
object_unref(OBJECT(pr_mgr->ioc));
pr_mgr->ioc = NULL;
+ qapi_event_send_pr_manager_status_changed(pr_mgr->qom_path, false,
+ &error_abort);
return n_written < 0 ? -EINVAL : 0;
}
@@ -150,7 +154,6 @@ static int pr_manager_helper_run(PRManager *p,
int ret;
int expected_dir;
int attempts;
- bool was_connected = pr_mgr->ioc != NULL;
uint8_t cdb[PR_HELPER_CDB_SIZE] = { 0 };
if (!io_hdr->cmd_len || io_hdr->cmd_len > PR_HELPER_CDB_SIZE) {
@@ -232,11 +235,6 @@ static int pr_manager_helper_run(PRManager *p,
return ret;
out:
- if (was_connected) {
- qapi_event_send_pr_manager_status_changed(pr_mgr->qom_path, false,
- &error_abort);
- }
-
sense_len = scsi_build_sense(io_hdr->sbp, SENSE_CODE(LUN_COMM_FAILURE));
io_hdr->driver_status = SG_ERR_DRIVER_SENSE;
io_hdr->sb_len_wr = MIN(io_hdr->mx_sb_len, sense_len);
@@ -261,7 +259,7 @@ static void pr_manager_helper_complete(UserCreatable *uc,
Error **errp)
{
PRManagerHelper *pr_mgr = PR_MANAGER_HELPER(uc);
- pr_mgr->qom_path = object_get_canonical_path(OBJECT(pr_mgr));
+ pr_mgr->qom_path = object_get_canonical_path_component(OBJECT(pr_mgr));
qemu_mutex_lock(&pr_mgr->lock);
pr_manager_helper_initialize(pr_mgr, errp);
With your approach the event is emitted only after the 5 reconnect
attempts (which makes little sense IMO). With my approach the event is
emitted a bit sooner so that reconnect has at least chance of
succeeding.
I've written some patches for libvirt and with these changes it works
well.
Michal
