[RFC,v3,6/7] virtio_rtc: Add Arm Generic Timer cross-timestamping

Commit Message

Peter Hilber Dec. 18, 2023, 7:38 a.m. UTC

Add PTP_SYS_OFFSET_PRECISE2 support on platforms using the Arm Generic
Timer.

Always report the CP15 virtual counter as the HW counter in use by
arm_arch_timer, since the Linux kernel's usage of the Arm Generic Timer
should always be compatible with this.

Signed-off-by: Peter Hilber <[e-mail]>
---

Notes:
    v2:
    
    - Depend on prerequisite patch series "treewide: Use clocksource id for
      get_device_system_crosststamp()".
    
    - Return clocksource id instead of calling dropped arm_arch_timer helpers.
    
    - Always report the CP15 virtual counter to be in use by arm_arch_timer,
      since distinction of Arm physical and virtual counter appears unneeded
      after discussion with Marc Zyngier.

 drivers/virtio/Kconfig          | 13 +++++++++++++
 drivers/virtio/Makefile         |  1 +
 drivers/virtio/virtio_rtc_arm.c | 22 ++++++++++++++++++++++
 3 files changed, 36 insertions(+)
 create mode 100644 drivers/virtio/virtio_rtc_arm.c

Comments

David Woodhouse June 15, 2024, 7:50 a.m. UTC | #1

On Mon, 2023-12-18 at 08:38 +0100, Peter Hilber wrote:
> 
> +int viortc_hw_xtstamp_params(u16 *hw_counter, enum clocksource_ids *cs_id)
> +{
> +       *hw_counter = VIRTIO_RTC_COUNTER_ARM_VIRT;

Hm, but what if it isn't? I think you need to put this in
drivers/clocksource/arm_arch_timer.c where it can do something like
kvm_arch_ptp_get_crosststamp() does to decide:

        if (arch_timer_uses_ppi == ARCH_TIMER_VIRT_PPI)
                ptp_counter = KVM_PTP_VIRT_COUNTER;
        else
                ptp_counter = KVM_PTP_PHYS_COUNTER;


> +       *cs_id = CSID_ARM_ARCH_COUNTER;
> +
> +       return 0;
> +}

Peter Hilber June 20, 2024, 12:06 p.m. UTC | #2

Changing virtio-dev address to the new one.

On 15.06.24 09:50, David Woodhouse wrote:
> On Mon, 2023-12-18 at 08:38 +0100, Peter Hilber wrote:
>>
>> +int viortc_hw_xtstamp_params(u16 *hw_counter, enum clocksource_ids *cs_id)
>> +{
>> +       *hw_counter = VIRTIO_RTC_COUNTER_ARM_VIRT;
> 
> Hm, but what if it isn't? I think you need to put this in
> drivers/clocksource/arm_arch_timer.c where it can do something like
> kvm_arch_ptp_get_crosststamp() does to decide:
> 
>          if (arch_timer_uses_ppi == ARCH_TIMER_VIRT_PPI)
>                  ptp_counter = KVM_PTP_VIRT_COUNTER;
>          else
>                  ptp_counter = KVM_PTP_PHYS_COUNTER;
> 
> 
>> +       *cs_id = CSID_ARM_ARCH_COUNTER;
>> +
>> +       return 0;
>> +}
> 

I had such a check in v1, but Marc Zyngier didn't like the distinction [1,
2] - maybe also exposing the timer through a generic helper was not
desired. Quoting from [2]:

>> This was the rationale to come up with the physical/virtual counter
>> distinction for the Virtio RTC device. Looking at extensions such as
>> FEAT_ECV, where the CNTPCT_EL0 value can depend on the EL, or FEAT_NV*,
>> it might be a bit simplistic.
> 
> Not just simplistic. It doesn't make sense. For this to work, you'd
> need to know the global offset that KVM applies to the global counter,
> plus the *virtualised* CNTPOFF/CNTVOFF values that the guest can
> change at any time on a *per-CPU* basis. None of that is available
> outside of KVM, nor would it make any sense anyway.
> 
>> Does this physical/virtual counter distinction sound like a good idea?
>> Otherwise I would drop the arch_timer_counter_get_type() in the next
>> iteration.
> 
> My take on this is that only the global counter value makes any sense.
> That value is already available from the host as the virtual counter,
> because we guarantee that CNTVOFF is 0 when outside of the guest
> (well, technically, outside of the vcpu_load/vcpu_put section).

So I put the assumption that the virtual counter will always be the right
choice for current Linux kernels, from the Virtio device POV.

Thanks for the comment,

Peter

[1] https://lore.kernel.org/lkml/[e-mail]/T/#ma8d596de1cbc8f4a78a18b2aa995db18423494a7
[2] https://lore.kernel.org/lkml/[e-mail]/T/#m65fa1d715933360498c4e33d7225e4220215a9d6

Patch

diff --git a/drivers/virtio/Kconfig b/drivers/virtio/Kconfig
index 8542b2f20201..d35c728778d2 100644
--- a/drivers/virtio/Kconfig
+++ b/drivers/virtio/Kconfig
@@ -205,6 +205,19 @@  config VIRTIO_RTC_PTP
 
 	 If unsure, say Y.
 
+config VIRTIO_RTC_ARM
+	bool "Virtio RTC cross-timestamping using Arm Generic Timer"
+	default y
+	depends on VIRTIO_RTC_PTP && ARM_ARCH_TIMER
+	help
+	 This enables Virtio RTC cross-timestamping using the Arm Generic Timer.
+	 It only has an effect if the Virtio RTC device also supports this. The
+	 cross-timestamp is available through the PTP clock driver precise
+	 cross-timestamp ioctl (PTP_SYS_OFFSET_PRECISE2 or
+	 PTP_SYS_OFFSET_PRECISE).
+
+	 If unsure, say Y.
+
 endif # VIRTIO_RTC
 
 endif # VIRTIO_MENU
diff --git a/drivers/virtio/Makefile b/drivers/virtio/Makefile
index 4d48cbcae6bb..781dff9f8822 100644
--- a/drivers/virtio/Makefile
+++ b/drivers/virtio/Makefile
@@ -15,3 +15,4 @@  obj-$(CONFIG_VIRTIO_DMA_SHARED_BUFFER) += virtio_dma_buf.o
 obj-$(CONFIG_VIRTIO_RTC) += virtio_rtc.o
 virtio_rtc-y := virtio_rtc_driver.o
 virtio_rtc-$(CONFIG_VIRTIO_RTC_PTP) += virtio_rtc_ptp.o
+virtio_rtc-$(CONFIG_VIRTIO_RTC_ARM) += virtio_rtc_arm.o
diff --git a/drivers/virtio/virtio_rtc_arm.c b/drivers/virtio/virtio_rtc_arm.c
new file mode 100644
index 000000000000..5185b130b3f1
--- /dev/null
+++ b/drivers/virtio/virtio_rtc_arm.c
@@ -0,0 +1,22 @@ 
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Provides cross-timestamp params for Arm.
+ *
+ * Copyright (C) 2022-2023 OpenSynergy GmbH
+ */
+
+#include <linux/clocksource_ids.h>
+
+#include <uapi/linux/virtio_rtc.h>
+
+#include "virtio_rtc_internal.h"
+
+/* see header for doc */
+
+int viortc_hw_xtstamp_params(u16 *hw_counter, enum clocksource_ids *cs_id)
+{
+	*hw_counter = VIRTIO_RTC_COUNTER_ARM_VIRT;
+	*cs_id = CSID_ARM_ARCH_COUNTER;
+
+	return 0;
+}