* [PATCH v2 RESEND] qga: Introduce disk smart
@ 2022-03-07 6:20 zhenwei pi
2022-03-07 7:18 ` Marc-André Lureau
0 siblings, 1 reply; 2+ messages in thread
From: zhenwei pi @ 2022-03-07 6:20 UTC (permalink / raw)
To: philippe.mathieu.daude, michael.roth
Cc: Keith Busch, armbru, zhenwei pi, qemu-devel
After assigning a NVMe/SCSI controller to guest by VFIO, we lose
everything on the host side. A guest uses these devices exclusively,
we usually don't care the actions on these devices. But there is a
low probability that hitting physical hardware warning, we need a
chance to get the basic smart log info.
Introduce disk smart, and implement NVMe smart on linux.
CC: Keith Busch <kbusch@kernel.org>
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
---
qga/commands-posix.c | 78 ++++++++++++++++++++++++++++++++++++++++++++
qga/qapi-schema.json | 50 +++++++++++++++++++++++++++-
2 files changed, 127 insertions(+), 1 deletion(-)
diff --git a/qga/commands-posix.c b/qga/commands-posix.c
index 4ec83bbfbc..1fa089e9c8 100644
--- a/qga/commands-posix.c
+++ b/qga/commands-posix.c
@@ -27,6 +27,7 @@
#include "qemu/base64.h"
#include "qemu/cutils.h"
#include "commands-common.h"
+#include "block/nvme.h"
#ifdef HAVE_UTMPX
#include <utmpx.h>
@@ -49,6 +50,7 @@ extern char **environ;
#include <sys/socket.h>
#include <net/if.h>
#include <sys/statvfs.h>
+#include <linux/nvme_ioctl.h>
#ifdef CONFIG_LIBUDEV
#include <libudev.h>
@@ -1390,6 +1392,81 @@ static GuestDiskInfoList *get_disk_partitions(
return ret;
}
+static void get_disk_smart(GuestDiskInfo *disk)
+{
+ if (disk->has_address
+ && (disk->address->bus_type == GUEST_DISK_BUS_TYPE_NVME)) {
+ int fd;
+ GuestDiskSmart *smart;
+ NvmeSmartLog log = {0};
+ struct nvme_admin_cmd cmd = {
+ .opcode = NVME_ADM_CMD_GET_LOG_PAGE,
+ .nsid = NVME_NSID_BROADCAST,
+ .addr = (uint64_t)&log,
+ .data_len = sizeof(log),
+ .cdw10 = NVME_LOG_SMART_INFO | (1 << 15) /* RAE bit */
+ | (((sizeof(log) >> 2) - 1) << 16)
+ };
+
+ fd = qemu_open_old(disk->name, O_RDONLY);
+ if (fd == -1) {
+ g_debug("Failed to open device: %s", disk->name);
+ return;
+ }
+ if (ioctl(fd, NVME_IOCTL_ADMIN_CMD, &cmd)) {
+ g_debug("Failed to get smart: %s", disk->name);
+ close(fd);
+ return;
+ }
+
+ smart = g_new0(GuestDiskSmart, 1);
+ smart->type = GUEST_DISK_BUS_TYPE_NVME;
+ smart->u.nvme.critical_warning = log.critical_warning;
+ /* unaligned 'temperature' field */
+ smart->u.nvme.temperature = lduw_le_p(&log.temperature);
+ smart->u.nvme.available_spare = log.available_spare;
+ smart->u.nvme.available_spare_threshold = log.available_spare_threshold;
+ smart->u.nvme.percentage_used = log.percentage_used;
+ smart->u.nvme.data_units_read_lo = le64_to_cpu(log.data_units_read[0]);
+ smart->u.nvme.data_units_read_hi = le64_to_cpu(log.data_units_read[1]);
+ smart->u.nvme.data_units_written_lo =
+ le64_to_cpu(log.data_units_written[0]);
+ smart->u.nvme.data_units_written_hi =
+ le64_to_cpu(log.data_units_written[1]);
+ smart->u.nvme.host_read_commands_lo =
+ le64_to_cpu(log.host_read_commands[0]);
+ smart->u.nvme.host_read_commands_hi =
+ le64_to_cpu(log.host_read_commands[1]);
+ smart->u.nvme.host_write_commands_lo =
+ le64_to_cpu(log.host_write_commands[0]);
+ smart->u.nvme.host_write_commands_hi =
+ le64_to_cpu(log.host_write_commands[1]);
+ smart->u.nvme.controller_busy_time_lo =
+ le64_to_cpu(log.controller_busy_time[0]);
+ smart->u.nvme.controller_busy_time_hi =
+ le64_to_cpu(log.controller_busy_time[1]);
+ smart->u.nvme.power_cycles_lo = le64_to_cpu(log.power_cycles[0]);
+ smart->u.nvme.power_cycles_hi = le64_to_cpu(log.power_cycles[1]);
+ smart->u.nvme.power_on_hours_lo = le64_to_cpu(log.power_on_hours[0]);
+ smart->u.nvme.power_on_hours_hi = le64_to_cpu(log.power_on_hours[1]);
+ smart->u.nvme.unsafe_shutdowns_lo =
+ le64_to_cpu(log.unsafe_shutdowns[0]);
+ smart->u.nvme.unsafe_shutdowns_hi =
+ le64_to_cpu(log.unsafe_shutdowns[1]);
+ smart->u.nvme.media_errors_lo = le64_to_cpu(log.media_errors[0]);
+ smart->u.nvme.media_errors_hi = le64_to_cpu(log.media_errors[1]);
+ smart->u.nvme.number_of_error_log_entries_lo =
+ le64_to_cpu(log.number_of_error_log_entries[0]);
+ smart->u.nvme.number_of_error_log_entries_hi =
+ le64_to_cpu(log.number_of_error_log_entries[1]);
+
+ disk->has_smart = true;
+ disk->smart = smart;
+
+ close(fd);
+ }
+}
+
GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
{
GuestDiskInfoList *ret = NULL;
@@ -1463,6 +1540,7 @@ GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
}
get_disk_deps(disk_dir, disk);
+ get_disk_smart(disk);
ret = get_disk_partitions(ret, de->d_name, disk_dir, dev_name);
}
diff --git a/qga/qapi-schema.json b/qga/qapi-schema.json
index 8f73770210..8bb8731ce4 100644
--- a/qga/qapi-schema.json
+++ b/qga/qapi-schema.json
@@ -888,6 +888,53 @@
'*serial': 'str', '*dev': 'str',
'*ccw-address': 'GuestCCWAddress'} }
+##
+# @GuestNVMeSmart:
+#
+# NVMe smart informations, base on NVMe specification
+#
+# Since: 7.0
+##
+{ 'struct': 'GuestNVMeSmart',
+ 'data': {'critical-warning': 'int',
+ 'temperature': 'int',
+ 'available-spare': 'int',
+ 'available-spare-threshold': 'int',
+ 'percentage-used': 'int',
+ 'data-units-read-lo': 'uint64',
+ 'data-units-read-hi': 'uint64',
+ 'data-units-written-lo': 'uint64',
+ 'data-units-written-hi': 'uint64',
+ 'host-read-commands-lo': 'uint64',
+ 'host-read-commands-hi': 'uint64',
+ 'host-write-commands-lo': 'uint64',
+ 'host-write-commands-hi': 'uint64',
+ 'controller-busy-time-lo': 'uint64',
+ 'controller-busy-time-hi': 'uint64',
+ 'power-cycles-lo': 'uint64',
+ 'power-cycles-hi': 'uint64',
+ 'power-on-hours-lo': 'uint64',
+ 'power-on-hours-hi': 'uint64',
+ 'unsafe-shutdowns-lo': 'uint64',
+ 'unsafe-shutdowns-hi': 'uint64',
+ 'media-errors-lo': 'uint64',
+ 'media-errors-hi': 'uint64',
+ 'number-of-error-log-entries-lo': 'uint64',
+ 'number-of-error-log-entries-hi': 'uint64' } }
+
+##
+# @GuestDiskSmart:
+#
+# Smart of disk
+# - @nvme: NVMe disk smart
+#
+# Since: 7.0
+##
+{ 'union': 'GuestDiskSmart',
+ 'base': { 'type': 'GuestDiskBusType' },
+ 'discriminator': 'type',
+ 'data': { 'nvme': 'GuestNVMeSmart' } }
+
##
# @GuestDiskInfo:
#
@@ -904,7 +951,8 @@
##
{ 'struct': 'GuestDiskInfo',
'data': {'name': 'str', 'partition': 'bool', '*dependencies': ['str'],
- '*address': 'GuestDiskAddress', '*alias': 'str'} }
+ '*address': 'GuestDiskAddress', '*alias': 'str',
+ '*smart': 'GuestDiskSmart'} }
##
# @guest-get-disks:
--
2.20.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH v2 RESEND] qga: Introduce disk smart
2022-03-07 6:20 [PATCH v2 RESEND] qga: Introduce disk smart zhenwei pi
@ 2022-03-07 7:18 ` Marc-André Lureau
0 siblings, 0 replies; 2+ messages in thread
From: Marc-André Lureau @ 2022-03-07 7:18 UTC (permalink / raw)
To: zhenwei pi
Cc: Keith Busch, Michael Roth, Philippe Mathieu-Daudé,
Markus Armbruster, QEMU
[-- Attachment #1: Type: text/plain, Size: 7935 bytes --]
Hi
On Mon, Mar 7, 2022 at 10:24 AM zhenwei pi <pizhenwei@bytedance.com> wrote:
> After assigning a NVMe/SCSI controller to guest by VFIO, we lose
> everything on the host side. A guest uses these devices exclusively,
> we usually don't care the actions on these devices. But there is a
> low probability that hitting physical hardware warning, we need a
> chance to get the basic smart log info.
>
> Introduce disk smart, and implement NVMe smart on linux.
>
> CC: Keith Busch <kbusch@kernel.org>
> Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
> ---
> qga/commands-posix.c | 78 ++++++++++++++++++++++++++++++++++++++++++++
> qga/qapi-schema.json | 50 +++++++++++++++++++++++++++-
> 2 files changed, 127 insertions(+), 1 deletion(-)
>
> diff --git a/qga/commands-posix.c b/qga/commands-posix.c
> index 4ec83bbfbc..1fa089e9c8 100644
> --- a/qga/commands-posix.c
> +++ b/qga/commands-posix.c
> @@ -27,6 +27,7 @@
> #include "qemu/base64.h"
> #include "qemu/cutils.h"
> #include "commands-common.h"
> +#include "block/nvme.h"
>
> #ifdef HAVE_UTMPX
> #include <utmpx.h>
> @@ -49,6 +50,7 @@ extern char **environ;
> #include <sys/socket.h>
> #include <net/if.h>
> #include <sys/statvfs.h>
> +#include <linux/nvme_ioctl.h>
>
> #ifdef CONFIG_LIBUDEV
> #include <libudev.h>
> @@ -1390,6 +1392,81 @@ static GuestDiskInfoList *get_disk_partitions(
> return ret;
> }
>
> +static void get_disk_smart(GuestDiskInfo *disk)
> +{
> + if (disk->has_address
> + && (disk->address->bus_type == GUEST_DISK_BUS_TYPE_NVME)) {
>
to avoid extra indentation, you could return early.
> + int fd;
> + GuestDiskSmart *smart;
> + NvmeSmartLog log = {0};
> + struct nvme_admin_cmd cmd = {
> + .opcode = NVME_ADM_CMD_GET_LOG_PAGE,
> + .nsid = NVME_NSID_BROADCAST,
> + .addr = (uint64_t)&log,
> + .data_len = sizeof(log),
> + .cdw10 = NVME_LOG_SMART_INFO | (1 << 15) /* RAE bit */
> + | (((sizeof(log) >> 2) - 1) << 16)
> + };
> +
> + fd = qemu_open_old(disk->name, O_RDONLY);
> + if (fd == -1) {
> + g_debug("Failed to open device: %s", disk->name);
>
g_strerror(errno) could help
> + return;
> + }
> + if (ioctl(fd, NVME_IOCTL_ADMIN_CMD, &cmd)) {
> + g_debug("Failed to get smart: %s", disk->name);
>
same
> + close(fd);
> + return;
> + }
> +
> + smart = g_new0(GuestDiskSmart, 1);
> + smart->type = GUEST_DISK_BUS_TYPE_NVME;
> + smart->u.nvme.critical_warning = log.critical_warning;
> + /* unaligned 'temperature' field */
> + smart->u.nvme.temperature = lduw_le_p(&log.temperature);
> + smart->u.nvme.available_spare = log.available_spare;
> + smart->u.nvme.available_spare_threshold =
> log.available_spare_threshold;
> + smart->u.nvme.percentage_used = log.percentage_used;
> + smart->u.nvme.data_units_read_lo =
> le64_to_cpu(log.data_units_read[0]);
> + smart->u.nvme.data_units_read_hi =
> le64_to_cpu(log.data_units_read[1]);
> + smart->u.nvme.data_units_written_lo =
> + le64_to_cpu(log.data_units_written[0]);
> + smart->u.nvme.data_units_written_hi =
> + le64_to_cpu(log.data_units_written[1]);
> + smart->u.nvme.host_read_commands_lo =
> + le64_to_cpu(log.host_read_commands[0]);
> + smart->u.nvme.host_read_commands_hi =
> + le64_to_cpu(log.host_read_commands[1]);
> + smart->u.nvme.host_write_commands_lo =
> + le64_to_cpu(log.host_write_commands[0]);
> + smart->u.nvme.host_write_commands_hi =
> + le64_to_cpu(log.host_write_commands[1]);
> + smart->u.nvme.controller_busy_time_lo =
> + le64_to_cpu(log.controller_busy_time[0]);
> + smart->u.nvme.controller_busy_time_hi =
> + le64_to_cpu(log.controller_busy_time[1]);
> + smart->u.nvme.power_cycles_lo = le64_to_cpu(log.power_cycles[0]);
> + smart->u.nvme.power_cycles_hi = le64_to_cpu(log.power_cycles[1]);
> + smart->u.nvme.power_on_hours_lo =
> le64_to_cpu(log.power_on_hours[0]);
> + smart->u.nvme.power_on_hours_hi =
> le64_to_cpu(log.power_on_hours[1]);
> + smart->u.nvme.unsafe_shutdowns_lo =
> + le64_to_cpu(log.unsafe_shutdowns[0]);
> + smart->u.nvme.unsafe_shutdowns_hi =
> + le64_to_cpu(log.unsafe_shutdowns[1]);
> + smart->u.nvme.media_errors_lo = le64_to_cpu(log.media_errors[0]);
> + smart->u.nvme.media_errors_hi = le64_to_cpu(log.media_errors[1]);
> + smart->u.nvme.number_of_error_log_entries_lo =
> + le64_to_cpu(log.number_of_error_log_entries[0]);
> + smart->u.nvme.number_of_error_log_entries_hi =
> + le64_to_cpu(log.number_of_error_log_entries[1]);
> +
> + disk->has_smart = true;
> + disk->smart = smart;
> +
> + close(fd);
> + }
> +}
> +
> GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
> {
> GuestDiskInfoList *ret = NULL;
> @@ -1463,6 +1540,7 @@ GuestDiskInfoList *qmp_guest_get_disks(Error **errp)
> }
>
> get_disk_deps(disk_dir, disk);
> + get_disk_smart(disk);
> ret = get_disk_partitions(ret, de->d_name, disk_dir, dev_name);
> }
>
> diff --git a/qga/qapi-schema.json b/qga/qapi-schema.json
> index 8f73770210..8bb8731ce4 100644
> --- a/qga/qapi-schema.json
> +++ b/qga/qapi-schema.json
> @@ -888,6 +888,53 @@
> '*serial': 'str', '*dev': 'str',
> '*ccw-address': 'GuestCCWAddress'} }
>
> +##
> +# @GuestNVMeSmart:
> +#
> +# NVMe smart informations, base on NVMe specification
>
based
> +#
> +# Since: 7.0
> +##
> +{ 'struct': 'GuestNVMeSmart',
> + 'data': {'critical-warning': 'int',
> + 'temperature': 'int',
> + 'available-spare': 'int',
> + 'available-spare-threshold': 'int',
> + 'percentage-used': 'int',
> + 'data-units-read-lo': 'uint64',
> + 'data-units-read-hi': 'uint64',
> + 'data-units-written-lo': 'uint64',
> + 'data-units-written-hi': 'uint64',
> + 'host-read-commands-lo': 'uint64',
> + 'host-read-commands-hi': 'uint64',
> + 'host-write-commands-lo': 'uint64',
> + 'host-write-commands-hi': 'uint64',
> + 'controller-busy-time-lo': 'uint64',
> + 'controller-busy-time-hi': 'uint64',
> + 'power-cycles-lo': 'uint64',
> + 'power-cycles-hi': 'uint64',
> + 'power-on-hours-lo': 'uint64',
> + 'power-on-hours-hi': 'uint64',
> + 'unsafe-shutdowns-lo': 'uint64',
> + 'unsafe-shutdowns-hi': 'uint64',
> + 'media-errors-lo': 'uint64',
> + 'media-errors-hi': 'uint64',
> + 'number-of-error-log-entries-lo': 'uint64',
> + 'number-of-error-log-entries-hi': 'uint64' } }
> +
> +##
> +# @GuestDiskSmart:
> +#
> +# Smart of disk
> +# - @nvme: NVMe disk smart
> +#
> +# Since: 7.0
> +##
> +{ 'union': 'GuestDiskSmart',
> + 'base': { 'type': 'GuestDiskBusType' },
> + 'discriminator': 'type',
> + 'data': { 'nvme': 'GuestNVMeSmart' } }
> +
> ##
> # @GuestDiskInfo:
> #
>
You should update the doc about the new smart field
> @@ -904,7 +951,8 @@
> ##
> { 'struct': 'GuestDiskInfo',
> 'data': {'name': 'str', 'partition': 'bool', '*dependencies': ['str'],
> - '*address': 'GuestDiskAddress', '*alias': 'str'} }
> + '*address': 'GuestDiskAddress', '*alias': 'str',
> + '*smart': 'GuestDiskSmart'} }
>
> ##
> # @guest-get-disks:
> --
> 2.20.1
>
>
>
--
Marc-André Lureau
[-- Attachment #2: Type: text/html, Size: 11144 bytes --]
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2022-03-07 7:29 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-03-07 6:20 [PATCH v2 RESEND] qga: Introduce disk smart zhenwei pi
2022-03-07 7:18 ` Marc-André Lureau
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.