Skip to contentSkip to conversation

Revert "net/mlx5: Update cyclecounter shift value to improve ptp free running mode precision"

PATCH In review Vadim Fedorenko 4 participants 7 messages Aug 15, 2023 – Aug 21, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Revert "net/mlx5: Update cyclecounter shift value to improve ptp free running mode precision"

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

---
 drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c +1 -1
index 377372f0578a..3e504e7d24ce 100644
@@ -39,7 +39,7 @@
 #include "clock.h"
 
 enum {
-	MLX5_CYCLES_SHIFT	= 31
+	MLX5_CYCLES_SHIFT	= 23
 };
 
 enum {
Signature · 1 line
2.39.3
Reviewed-by on v1 1/1
#1 lore raw
Quoted 4 lines · On Tue, 15 Aug, 2023 08:15:07 -0700 Vadim Fedorenko <[email protected]> wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

Quoted 8 lines

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

---

I agree with this revert. This change was made with the assumption that all mlx5 compatible devices were running a 1Ghz internal timer. Will sync with folks internally about how we can support higher precision free running mode while accounting for the different device timer clock speeds.

#2 lore raw
Quoted 11 lines · On Tue, 15 Aug, 2023 09:53:27 -0700 Rahul Rameshbabu <[email protected]>…
Quoted 10 lines · On Tue, 15 Aug, 2023 08:15:07 -0700 Vadim Fedorenko <[email protected]> wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

Quoted 8 lines
Quoted 2 lines

---

I agree with this revert. This change was made with the assumption that all mlx5 compatible devices were running a 1Ghz internal timer. Will sync with folks internally about how we can support higher precision free running mode while accounting for the different device timer clock speeds.

I have prepared a patch that resolves the issue of selecting a valid shift constant based on the device frequency. The patch includes logic for picking the optimal shift constant for PTP frequency adjustment based on the device frequency. I have tested this against a variety of mlx5 devices and believe this solution would be ideal. I expect this patch to make its way to the mailing list by the end of the week.

--
Thanks,

Rahul Rameshbabu

maintainer
#3 lore raw
In reply to Vadim Fedorenko
Quoted 9 lines · On Tue, 15 Aug 2023 08:15:07 -0700 Vadim Fedorenko wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

Hi Saeed, any preference here? Given we're past -rc6 and the small size of the revert it seems like a tempting solution?

#4 lore raw
Quoted 5 lines · Jakub Kicinski
Quoted 4 lines · On Tue, 15 Aug 2023 08:15:07 -0700 Vadim Fedorenko wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

Quoted 8 lines
Quoted 5 lines

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

Hi Saeed, any preference here? Given we're past -rc6 and the small size of the revert it seems like a tempting solution?

Rahul's solution is also very small and already passed review, we will be conducting some tests, share the patch with Vadim for testing and I will be submitting it early next week.

Quoted 1 line
#5 lore raw

Hi!

I have tested the diff, looks like it works on both CX4 and CX6 cards I have right now. Wasn't able to find CX5 to test it, but I think it's ok to publish it and wait for some other users of CX5 to test.

Best,
Vadim

________________________________________
From: Saeed Mahameed <[email protected]>
Sent: 17 August 2023 17:37
To: Jakub Kicinski
Cc: Saeed Mahameed; Vadim Fedorenko; Rahul Rameshbabu; Gal Pressman; Bar Shapira; Vadim Fedorenko; Richard Cochran; [email protected]
Subject: Re: [PATCH net] Revert "net/mlx5: Update cyclecounter shift value to improve ptp free running mode precision"

Quoted 5 lines · Jakub Kicinski
Quoted 4 lines · On Tue, 15 Aug 2023 08:15:07 -0700 Vadim Fedorenko wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

Quoted 8 lines
Quoted 5 lines

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

Hi Saeed, any preference here? Given we're past -rc6 and the small size of the revert it seems like a tempting solution?

Rahul's solution is also very small and already passed review, we will be conducting some tests, share the patch with Vadim for testing and I will be submitting it early next week.

Quoted 1 line
#6 lore raw
Quoted 5 lines · On Mon, 21 Aug, 2023 21:37:51 +0000 Vadim Fedorenko <[email protected]> wrote:

Hi!

I have tested the diff, looks like it works on both CX4 and CX6 cards I have right now. Wasn't able to find CX5 to test it, but I think it's ok to publish it and wait for some other users of CX5 to test.

Thanks Vadim. I have tested on CX5 myself, but a CX5 user reached out to me directly since he was also impacted by the original change. If we can, getting his feedback would be good.

Quoted 25 lines

Best,
Vadim

________________________________________
From: Saeed Mahameed <[email protected]>
Sent: 17 August 2023 17:37
To: Jakub Kicinski
Cc: Saeed Mahameed; Vadim Fedorenko; Rahul Rameshbabu; Gal Pressman; Bar Shapira; Vadim Fedorenko; Richard Cochran; [email protected]
Subject: Re: [PATCH net] Revert "net/mlx5: Update cyclecounter shift value to improve ptp free running mode precision"

Quoted 11 lines · Jakub Kicinski
Quoted 9 lines · On Tue, 15 Aug 2023 08:15:07 -0700 Vadim Fedorenko wrote:

From: Vadim Fedorenko <[email protected]>

This reverts commit 6a40109275626267ebf413ceda81c64719b5c431.

There was an assumption in the original commit that all the devices supported by mlx5 advertise 1GHz as an internal timer frequency. Apparently at least ConnectX-4 Lx (MCX4431N-GCAN) provides 156.250Mhz as an internal frequency and the original commit breaks PTP synchronization on these cards.

Quoted 7 lines
Quoted 2 lines

Hi Saeed, any preference here? Given we're past -rc6 and the small size of the revert it seems like a tempting solution?

Rahul's solution is also very small and already passed review, we will be conducting some tests, share the patch with Vadim for testing and I will be submitting it early next week.

Hi Saeed, I think we can send the patch out whenever you want to. I will share the patch directly with the individual CX5 user who reached out to me as well.

--
Thanks,

Rahul Rameshbabu