Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-17 Thread Mor Bar-Gabay 

On 09/07/2025 20:28, Markus Blöchl wrote:

get_time_fn() callback implementations are expected to fill out the
entire system_counterval_t struct as it may be initially uninitialized.

This broke with the removal of convert_art_to_tsc() helper functions
which left use_nsecs uninitialized.

Assign the entire struct again.

Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
Cc: [email protected]
Signed-off-by: Markus Blöchl 
Reviewed-by: Aleksandr Loktionov 
Reviewed-by: Simon Horman 
Reviewed-by: Aleksandr Loktionov 
Reviewed-by: Vitaly Lifshits 
---
Notes:
 Related-To: 


Changes in v2:
- Add Lakshmi in Cc:
- Add Signed-off-by: trailer which was lost in b4 workflow
- Link to v1: 
https://lore.kernel.org/r/[email protected]
---
  drivers/net/ethernet/intel/e1000e/ptp.c | 7 +--
  1 file changed, 5 insertions(+), 2 deletions(-)


Tested-by: Mor Bar-Gabay

Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-13 Thread Lifshits, Vitaly 

On 7/10/2025 1:33 PM, Loktionov, Aleksandr wrote:




-Original Message-
From: Intel-wired-lan  On Behalf
Of Simon Horman
Sent: Thursday, July 10, 2025 12:25 PM
To: Markus Blöchl 
Cc: Nguyen, Anthony L ; Kitszel,
Przemyslaw ; Richard Cochran
; Thomas Gleixner ;
Lakshmi Sowjanya D ; Andrew Lunn
; David S. Miller ; Eric
Dumazet ; Jakub Kicinski ; Paolo
Abeni ; Bloechl, Markus
; John Stultz ;
[email protected]; [email protected]; linux-
[email protected]
Subject: Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire
system_counterval_t in get_time_fn() callback

On Wed, Jul 09, 2025 at 07:28:07PM +0200, Markus Blöchl wrote:

get_time_fn() callback implementations are expected to fill out the
entire system_counterval_t struct as it may be initially

uninitialized.


This broke with the removal of convert_art_to_tsc() helper functions
which left use_nsecs uninitialized.

Assign the entire struct again.

Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
Cc: [email protected]
Signed-off-by: Markus Blöchl 
---
Notes:
 Related-To:


<https://lore.kernel.org/lkml/txyrr26hxe3xpq3ebqb5ewkgvhvp7xalotaouwlu

djtjifnah2@7tmgczln4aoo/>

Changes in v2:
- Add Lakshmi in Cc:
- Add Signed-off-by: trailer which was lost in b4 workflow
- Link to v1:
https://lore.kernel.org/r/20250709-e1000e_crossts-v1-1-

f8a80c792e4f@bl

ochl.de


Reviewed-by: Simon Horman 

Reviewed-by: Aleksandr Loktionov 


Reviewed-by: Vitaly Lifshits

Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-10 Thread Loktionov, Aleksandr 


> -Original Message-
> From: Intel-wired-lan  On Behalf
> Of Simon Horman
> Sent: Thursday, July 10, 2025 12:25 PM
> To: Markus Blöchl 
> Cc: Nguyen, Anthony L ; Kitszel,
> Przemyslaw ; Richard Cochran
> ; Thomas Gleixner ;
> Lakshmi Sowjanya D ; Andrew Lunn
> ; David S. Miller ; Eric
> Dumazet ; Jakub Kicinski ; Paolo
> Abeni ; Bloechl, Markus
> ; John Stultz ;
> [email protected]; [email protected]; linux-
> [email protected]
> Subject: Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire
> system_counterval_t in get_time_fn() callback
> 
> On Wed, Jul 09, 2025 at 07:28:07PM +0200, Markus Blöchl wrote:
> > get_time_fn() callback implementations are expected to fill out the
> > entire system_counterval_t struct as it may be initially
> uninitialized.
> >
> > This broke with the removal of convert_art_to_tsc() helper functions
> > which left use_nsecs uninitialized.
> >
> > Assign the entire struct again.
> >
> > Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
> > Cc: [email protected]
> > Signed-off-by: Markus Blöchl 
> > ---
> > Notes:
> > Related-To:
> >
> <https://lore.kernel.org/lkml/txyrr26hxe3xpq3ebqb5ewkgvhvp7xalotaouwlu
> > djtjifnah2@7tmgczln4aoo/>
> >
> > Changes in v2:
> > - Add Lakshmi in Cc:
> > - Add Signed-off-by: trailer which was lost in b4 workflow
> > - Link to v1:
> > https://lore.kernel.org/r/20250709-e1000e_crossts-v1-1-
> f8a80c792e4f@bl
> > ochl.de
> 
> Reviewed-by: Simon Horman 
Reviewed-by: Aleksandr Loktionov

Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-10 Thread Simon Horman 
On Wed, Jul 09, 2025 at 07:28:07PM +0200, Markus Blöchl wrote:
> get_time_fn() callback implementations are expected to fill out the
> entire system_counterval_t struct as it may be initially uninitialized.
> 
> This broke with the removal of convert_art_to_tsc() helper functions
> which left use_nsecs uninitialized.
> 
> Assign the entire struct again.
> 
> Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
> Cc: [email protected]
> Signed-off-by: Markus Blöchl 
> ---
> Notes:
> Related-To: 
> 
> 
> Changes in v2:
> - Add Lakshmi in Cc:
> - Add Signed-off-by: trailer which was lost in b4 workflow
> - Link to v1: 
> https://lore.kernel.org/r/[email protected]

Reviewed-by: Simon Horman

Re: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-09 Thread Loktionov, Aleksandr 


> -Original Message-
> From: Intel-wired-lan  On Behalf
> Of Markus Blöchl
> Sent: Wednesday, July 9, 2025 7:28 PM
> To: Nguyen, Anthony L ; Kitszel,
> Przemyslaw ; Richard Cochran
> ; Thomas Gleixner 
> Cc: Lakshmi Sowjanya D ; Andrew Lunn
> ; David S. Miller ; Eric
> Dumazet ; Jakub Kicinski ; Paolo
> Abeni ; Bloechl, Markus
> ; John Stultz ;
> [email protected]; [email protected]; linux-
> [email protected]; Markus Blöchl 
> Subject: [Intel-wired-lan] [PATCH v2] e1000e: Populate entire
> system_counterval_t in get_time_fn() callback
> 
> get_time_fn() callback implementations are expected to fill out the
> entire system_counterval_t struct as it may be initially
> uninitialized.
> 
> This broke with the removal of convert_art_to_tsc() helper functions
> which left use_nsecs uninitialized.
> 
> Assign the entire struct again.
> 
> Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
> Cc: [email protected]
> Signed-off-by: Markus Blöchl 
Reviewed-by: Aleksandr Loktionov 

> ---
> Notes:
> Related-To:
> <https://lore.kernel.org/lkml/txyrr26hxe3xpq3ebqb5ewkgvhvp7xalotaouwlu
> djtjifnah2@7tmgczln4aoo/>
> 
> Changes in v2:
> - Add Lakshmi in Cc:
> - Add Signed-off-by: trailer which was lost in b4 workflow
> - Link to v1: https://lore.kernel.org/r/20250709-e1000e_crossts-v1-1-
> [email protected]
> ---
>  drivers/net/ethernet/intel/e1000e/ptp.c | 7 +--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/e1000e/ptp.c
> b/drivers/net/ethernet/intel/e1000e/ptp.c
> index
> ea3c3eb2ef2020d513d49c1368679f27d17edb04..f01506504ee3a11822930115e9ed
> 07661d81532c 100644
> --- a/drivers/net/ethernet/intel/e1000e/ptp.c
> +++ b/drivers/net/ethernet/intel/e1000e/ptp.c
> @@ -124,8 +124,11 @@ static int e1000e_phc_get_syncdevicetime(ktime_t
> *device,
>   sys_cycles = er32(PLTSTMPH);
>   sys_cycles <<= 32;
>   sys_cycles |= er32(PLTSTMPL);
> - system->cycles = sys_cycles;
> - system->cs_id = CSID_X86_ART;
> + *system = (struct system_counterval_t) {
> + .cycles = sys_cycles,
> + .cs_id = CSID_X86_ART,
> + .use_nsecs = false,
> + };
> 
>   return 0;
>  }
> 
> ---
> base-commit: 733923397fd95405a48f165c9b1fbc8c4b0a4681
> change-id: 20250709-e1000e_crossts-7745674f682a
> 
> Best regards,
> --
> Markus Blöchl

[Intel-wired-lan] [PATCH v2] e1000e: Populate entire system_counterval_t in get_time_fn() callback

2025-07-09 Thread Markus Blöchl 
get_time_fn() callback implementations are expected to fill out the
entire system_counterval_t struct as it may be initially uninitialized.

This broke with the removal of convert_art_to_tsc() helper functions
which left use_nsecs uninitialized.

Assign the entire struct again.

Fixes: bd48b50be50a ("e1000e: Replace convert_art_to_tsc()")
Cc: [email protected]
Signed-off-by: Markus Blöchl 
---
Notes:
Related-To: 


Changes in v2:
- Add Lakshmi in Cc:
- Add Signed-off-by: trailer which was lost in b4 workflow
- Link to v1: 
https://lore.kernel.org/r/[email protected]
---
 drivers/net/ethernet/intel/e1000e/ptp.c | 7 +--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/ptp.c 
b/drivers/net/ethernet/intel/e1000e/ptp.c
index 
ea3c3eb2ef2020d513d49c1368679f27d17edb04..f01506504ee3a11822930115e9ed07661d81532c
 100644
--- a/drivers/net/ethernet/intel/e1000e/ptp.c
+++ b/drivers/net/ethernet/intel/e1000e/ptp.c
@@ -124,8 +124,11 @@ static int e1000e_phc_get_syncdevicetime(ktime_t *device,
sys_cycles = er32(PLTSTMPH);
sys_cycles <<= 32;
sys_cycles |= er32(PLTSTMPL);
-   system->cycles = sys_cycles;
-   system->cs_id = CSID_X86_ART;
+   *system = (struct system_counterval_t) {
+   .cycles = sys_cycles,
+   .cs_id = CSID_X86_ART,
+   .use_nsecs = false,
+   };
 
return 0;
 }

---
base-commit: 733923397fd95405a48f165c9b1fbc8c4b0a4681
change-id: 20250709-e1000e_crossts-7745674f682a

Best regards,
-- 
Markus Blöchl