Fw: [Bug 220168] New: flexcan-core.c

Fw: [Bug 220168] New: flexcan-core.c

[Stephen Hemminger]

Begin forwarded message:

Date: Wed, 28 May 2025 15:19:21 +0000
From: bugzilla-daemon@kernel.org
To: stephen@networkplumber.org
Subject: [Bug 220168] New: flexcan-core.c


https://bugzilla.kernel.org/show_bug.cgi?id=220168

            Bug ID: 220168
           Summary: flexcan-core.c
           Product: Networking
           Version: 2.5
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: Other
          Assignee: stephen@networkplumber.org
          Reporter: robbinvandamme@gmail.com
        Regression: No

Bit15 (TDCEN) should not be enabled according to the spec if dbrp > 2 and in
the driver it gets enabled when i set dbrp = 10. Resulting in RX Error Frames.


spec:
ISO11898-1:2015(E), section 11.3.3 (Transmtiter delay compensation) (page 52):
It shall be programmable whether the mechanism is used at all. When this
mechanism is used, the value of the prescaler for the data time quantum m(D)
shall be one or two.

The "prescaler for the data time quantim m(D)" is commonly referred to as
"dbrp" by implementations.


Kind regards, 

Robbin

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.

Re: Fw: [Bug 220168] New: flexcan-core.c

[Vincent Mailhol]

+To: Robbin

Hi Stephen and Robbin,

Thanks for forwarding us this bug report.

On Thu. 29 May 2025 à 06:48, Stephen Hemminger
<stephen@networkplumber.org> wrote:
> Begin forwarded message:
>
> Date: Wed, 28 May 2025 15:19:21 +0000
> From: bugzilla-daemon@kernel.org
> To: stephen@networkplumber.org
> Subject: [Bug 220168] New: flexcan-core.c
>
>
> https://bugzilla.kernel.org/show_bug.cgi?id=220168
>
>             Bug ID: 220168
>            Summary: flexcan-core.c
>            Product: Networking
>            Version: 2.5
>           Hardware: All
>                 OS: Linux
>             Status: NEW
>           Severity: normal
>           Priority: P3
>          Component: Other
>           Assignee: stephen@networkplumber.org
>           Reporter: robbinvandamme@gmail.com
>         Regression: No
>
> Bit15 (TDCEN) should not be enabled according to the spec if dbrp > 2 and in
> the driver it gets enabled when i set dbrp = 10. Resulting in RX Error Frames.
>
>
> spec:
> ISO11898-1:2015(E), section 11.3.3 (Transmtiter delay compensation) (page 52):
> It shall be programmable whether the mechanism is used at all. When this
> mechanism is used, the value of the prescaler for the data time quantum m(D)
> shall be one or two.
>
> The "prescaler for the data time quantim m(D)" is commonly referred to as
> "dbrp" by implementations.

The issue which you describe here as well as the issue in bug 220170
[1] share roughly the same root cause: the driver is setting the TDC
parameters manually without doing any proper checking. The reason is
that, at the time of writing, the CAN subsystem did not have a
framework to handle the TDC logic. This has changed since.

We had a very similar issue on the mcp251xfd driver recently which was
fixed in [2]. Other relevant background information can be found in
[3]. The flexcan needs to be modified in a similar fashion.

I do not have the hardware so it would be hard to fix on my side.

@Robbin: I assume you have access to the hardware. Would you like to
try to write the patch for this? I can provide guidance.

[1] https://bugzilla.kernel.org/show_bug.cgi?id=220170

[2] commit 5e1663810e11 ("can: mcp251xfd: fix TDC setting for low data
bit rates")
Link: https://git.kernel.org/torvalds/c/5e1663810e11

[3] Re: mcp251xfd: forced TDC
Link: https://lore.kernel.org/all/CAMZ6RqKdyFPRwMbCZY5HwzLTeT9jcbxFiOfcuHbo+VgFF3ZViw@mail.gmail.com/

Re: Fw: [Bug 220168] New: flexcan-core.c

[Vincent Mailhol]

+To: Robbin

Hi Stephen and Robbin,

Thanks for forwarding us this bug report.

On 29/05/2025 at 06:48, Stephen Hemminger wrote:
> Begin forwarded message:
> 
> Date: Wed, 28 May 2025 15:19:21 +0000
> From: bugzilla-daemon@kernel.org
> To: stephen@networkplumber.org
> Subject: [Bug 220168] New: flexcan-core.c
> 
> 
> https://bugzilla.kernel.org/show_bug.cgi?id=220168
> 
>             Bug ID: 220168
>            Summary: flexcan-core.c
>            Product: Networking
>            Version: 2.5
>           Hardware: All
>                 OS: Linux
>             Status: NEW
>           Severity: normal
>           Priority: P3
>          Component: Other
>           Assignee: stephen@networkplumber.org
>           Reporter: robbinvandamme@gmail.com
>         Regression: No
> 
> Bit15 (TDCEN) should not be enabled according to the spec if dbrp > 2 and in
> the driver it gets enabled when i set dbrp = 10. Resulting in RX Error Frames.
> 
> 
> spec:
> ISO11898-1:2015(E), section 11.3.3 (Transmtiter delay compensation) (page 52):
> It shall be programmable whether the mechanism is used at all. When this
> mechanism is used, the value of the prescaler for the data time quantum m(D)
> shall be one or two.
> 
> The "prescaler for the data time quantim m(D)" is commonly referred to as
> "dbrp" by implementations.

The issue which you describe here as well as the issue in bug 220170
[1] share roughly the same root cause: the driver is setting the TDC
parameters manually without doing any proper checking. The reason is
that, at the time of writing, the CAN subsystem did not have a
framework to handle the TDC logic. This has changed since.

We had a very similar issue on the mcp251xfd driver recently which was
fixed in [2]. Other relevant background information can be found in
[3]. The flexcan needs to be modified in a similar fashion.

I do not have the hardware so it would be hard to fix on my side.

@Robbin: I assume you have access to the hardware. Would you like to
try to write the patch for this? I can provide guidance.

[1] https://bugzilla.kernel.org/show_bug.cgi?id=220170

[2] commit 5e1663810e11 ("can: mcp251xfd: fix TDC setting for low data
bit rates")
Link: https://git.kernel.org/torvalds/c/5e1663810e11

[3] Re: mcp251xfd: forced TDC
Link:
https://lore.kernel.org/all/CAMZ6RqKdyFPRwMbCZY5HwzLTeT9jcbxFiOfcuHbo+VgFF3ZViw@mail.gmail.com/


Yours sincerely,
Vincent Mailhol