* Re: [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization
2022-09-21 18:39 [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization Stanislav Kinsburskii
@ 2022-09-22 17:19 ` Wei Liu
2022-09-22 20:07 ` Dexuan Cui
2022-09-22 21:59 ` Michael Kelley (LINUX)
2022-09-23 10:39 ` Wei Liu
2 siblings, 1 reply; 5+ messages in thread
From: Wei Liu @ 2022-09-22 17:19 UTC (permalink / raw)
To: Stanislav Kinsburskii
Cc: skinsburskii, Stanislav Kinsburskiy, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Wei Liu, Dexuan Cui,
linux-hyperv, linux-kernel, Michael Kelley
On Wed, Sep 21, 2022 at 06:39:05PM +0000, Stanislav Kinsburskii wrote:
> From: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
>
> Waiting to 5 seconds in case of missing VMBUS ACPI device is redundant as the
> device is either present already or won't be available at all.
> This patch enforces synchronous probing to make sure the bus traversal,
> happening upon driver registering will either find the device (if present) or
> not spend any additional time if device is absent.
>
> Signed-off-by: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> CC: "K. Y. Srinivasan" <kys@microsoft.com>
> CC: Haiyang Zhang <haiyangz@microsoft.com>
> CC: Stephen Hemminger <sthemmin@microsoft.com>
> CC: Wei Liu <wei.liu@kernel.org>
> CC: Dexuan Cui <decui@microsoft.com>
> CC: linux-hyperv@vger.kernel.org
> CC: linux-kernel@vger.kernel.org
Looks good to me. I will wait for a few days for others to chime in.
CC Michael.
> ---
> drivers/hv/vmbus_drv.c | 13 ++++---------
> 1 file changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
> index 7b9f3fc3adf7..32d0009631a6 100644
> --- a/drivers/hv/vmbus_drv.c
> +++ b/drivers/hv/vmbus_drv.c
> @@ -46,8 +46,6 @@ struct vmbus_dynid {
>
> static struct acpi_device *hv_acpi_dev;
>
> -static struct completion probe_event;
> -
> static int hyperv_cpuhp_online;
>
> static void *hv_panic_page;
> @@ -2468,7 +2466,6 @@ static int vmbus_acpi_add(struct acpi_device *device)
> ret_val = 0;
>
> acpi_walk_err:
> - complete(&probe_event);
> if (ret_val)
> vmbus_acpi_remove(device);
> return ret_val;
> @@ -2647,6 +2644,7 @@ static struct acpi_driver vmbus_acpi_driver = {
> .remove = vmbus_acpi_remove,
> },
> .drv.pm = &vmbus_bus_pm,
> + .drv.probe_type = PROBE_FORCE_SYNCHRONOUS,
> };
>
> static void hv_kexec_handler(void)
> @@ -2719,7 +2717,7 @@ static struct syscore_ops hv_synic_syscore_ops = {
>
> static int __init hv_acpi_init(void)
> {
> - int ret, t;
> + int ret;
>
> if (!hv_is_hyperv_initialized())
> return -ENODEV;
> @@ -2727,8 +2725,6 @@ static int __init hv_acpi_init(void)
> if (hv_root_partition)
> return 0;
>
> - init_completion(&probe_event);
> -
> /*
> * Get ACPI resources first.
> */
> @@ -2737,9 +2733,8 @@ static int __init hv_acpi_init(void)
> if (ret)
> return ret;
>
> - t = wait_for_completion_timeout(&probe_event, 5*HZ);
> - if (t == 0) {
> - ret = -ETIMEDOUT;
> + if (!hv_acpi_dev) {
> + ret = -ENODEV;
> goto cleanup;
> }
>
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread* RE: [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization
2022-09-22 17:19 ` Wei Liu
@ 2022-09-22 20:07 ` Dexuan Cui
0 siblings, 0 replies; 5+ messages in thread
From: Dexuan Cui @ 2022-09-22 20:07 UTC (permalink / raw)
To: Wei Liu, Stanislav Kinsburskii
Cc: Stanislav Kinsburskii, Stanislav Kinsburskiy, KY Srinivasan,
Haiyang Zhang, Stephen Hemminger, linux-hyperv, linux-kernel,
Michael Kelley (LINUX)
> From: Wei Liu <wei.liu@kernel.org>
> Sent: Thursday, September 22, 2022 10:20 AM
> ...
> On Wed, Sep 21, 2022 at 06:39:05PM +0000, Stanislav Kinsburskii wrote:
> > From: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> >
> > Waiting to 5 seconds in case of missing VMBUS ACPI device is redundant as
> the
> > device is either present already or won't be available at all.
> > This patch enforces synchronous probing to make sure the bus traversal,
> > happening upon driver registering will either find the device (if present) or
> > not spend any additional time if device is absent.
> >
> > Signed-off-by: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> > CC: "K. Y. Srinivasan" <kys@microsoft.com>
> > CC: Haiyang Zhang <haiyangz@microsoft.com>
> > CC: Stephen Hemminger <sthemmin@microsoft.com>
> > CC: Wei Liu <wei.liu@kernel.org>
> > CC: Dexuan Cui <decui@microsoft.com>
> > CC: linux-hyperv@vger.kernel.org
> > CC: linux-kernel@vger.kernel.org
>
> Looks good to me. I will wait for a few days for others to chime in.
>
> CC Michael.
Looks good to me.
Reviewed-by: Dexuan Cui <decui@microsoft.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization
2022-09-21 18:39 [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization Stanislav Kinsburskii
2022-09-22 17:19 ` Wei Liu
@ 2022-09-22 21:59 ` Michael Kelley (LINUX)
2022-09-23 10:39 ` Wei Liu
2 siblings, 0 replies; 5+ messages in thread
From: Michael Kelley (LINUX) @ 2022-09-22 21:59 UTC (permalink / raw)
To: Stanislav Kinsburskii, Stanislav Kinsburskii
Cc: Stanislav Kinsburskiy, KY Srinivasan, Haiyang Zhang,
Stephen Hemminger, Wei Liu, Dexuan Cui, linux-hyperv,
linux-kernel
From: Stanislav Kinsburskii <skinsburskii@linux.microsoft.com> Sent: Wednesday, September 21, 2022 11:39 AM
>
> Waiting to 5 seconds in case of missing VMBUS ACPI device is redundant as the
A nit: We've been all over the map on the capitalization of "VMBus". Somebody
outside of Microsoft recently did a census in the Linux code and commits, and
"VMbus" is the most popular, closely followed by "VMBus". The preferred approach
in Microsoft public documentation is "VMBus", so that's what we should try to
use going forward. Not "VMBUS" and not "vmbus", though the latter is fine in
variable and function names.
> device is either present already or won't be available at all.
> This patch enforces synchronous probing to make sure the bus traversal,
> happening upon driver registering will either find the device (if present) or
> not spend any additional time if device is absent.
>
> Signed-off-by: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> CC: "K. Y. Srinivasan" <kys@microsoft.com>
> CC: Haiyang Zhang <haiyangz@microsoft.com>
> CC: Stephen Hemminger <sthemmin@microsoft.com>
> CC: Wei Liu <wei.liu@kernel.org>
> CC: Dexuan Cui <decui@microsoft.com>
> CC: linux-hyperv@vger.kernel.org
> CC: linux-kernel@vger.kernel.org
> ---
> drivers/hv/vmbus_drv.c | 13 ++++---------
> 1 file changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
> index 7b9f3fc3adf7..32d0009631a6 100644
> --- a/drivers/hv/vmbus_drv.c
> +++ b/drivers/hv/vmbus_drv.c
> @@ -46,8 +46,6 @@ struct vmbus_dynid {
>
> static struct acpi_device *hv_acpi_dev;
>
> -static struct completion probe_event;
> -
> static int hyperv_cpuhp_online;
>
> static void *hv_panic_page;
> @@ -2468,7 +2466,6 @@ static int vmbus_acpi_add(struct acpi_device *device)
> ret_val = 0;
>
> acpi_walk_err:
> - complete(&probe_event);
> if (ret_val)
> vmbus_acpi_remove(device);
> return ret_val;
> @@ -2647,6 +2644,7 @@ static struct acpi_driver vmbus_acpi_driver = {
> .remove = vmbus_acpi_remove,
> },
> .drv.pm = &vmbus_bus_pm,
> + .drv.probe_type = PROBE_FORCE_SYNCHRONOUS,
> };
>
> static void hv_kexec_handler(void)
> @@ -2719,7 +2717,7 @@ static struct syscore_ops hv_synic_syscore_ops = {
>
> static int __init hv_acpi_init(void)
> {
> - int ret, t;
> + int ret;
>
> if (!hv_is_hyperv_initialized())
> return -ENODEV;
> @@ -2727,8 +2725,6 @@ static int __init hv_acpi_init(void)
> if (hv_root_partition)
> return 0;
>
> - init_completion(&probe_event);
> -
> /*
> * Get ACPI resources first.
> */
> @@ -2737,9 +2733,8 @@ static int __init hv_acpi_init(void)
> if (ret)
> return ret;
>
> - t = wait_for_completion_timeout(&probe_event, 5*HZ);
> - if (t == 0) {
> - ret = -ETIMEDOUT;
> + if (!hv_acpi_dev) {
> + ret = -ENODEV;
> goto cleanup;
> }
>
>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization
2022-09-21 18:39 [PATCH] Drivers: hv: vmbus: Don't wait for the ACPI device upon initialization Stanislav Kinsburskii
2022-09-22 17:19 ` Wei Liu
2022-09-22 21:59 ` Michael Kelley (LINUX)
@ 2022-09-23 10:39 ` Wei Liu
2 siblings, 0 replies; 5+ messages in thread
From: Wei Liu @ 2022-09-23 10:39 UTC (permalink / raw)
To: Stanislav Kinsburskii
Cc: skinsburskii, Stanislav Kinsburskiy, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Wei Liu, Dexuan Cui,
linux-hyperv, linux-kernel
On Wed, Sep 21, 2022 at 06:39:05PM +0000, Stanislav Kinsburskii wrote:
> From: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
>
> Waiting to 5 seconds in case of missing VMBUS ACPI device is redundant as the
> device is either present already or won't be available at all.
> This patch enforces synchronous probing to make sure the bus traversal,
> happening upon driver registering will either find the device (if present) or
> not spend any additional time if device is absent.
>
> Signed-off-by: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> CC: "K. Y. Srinivasan" <kys@microsoft.com>
> CC: Haiyang Zhang <haiyangz@microsoft.com>
> CC: Stephen Hemminger <sthemmin@microsoft.com>
> CC: Wei Liu <wei.liu@kernel.org>
> CC: Dexuan Cui <decui@microsoft.com>
> CC: linux-hyperv@vger.kernel.org
> CC: linux-kernel@vger.kernel.org
Applied to hyperv-next. I changed vmbus to VMBus in the first paragraph
and added a line break.
^ permalink raw reply [flat|nested] 5+ messages in thread