Re: [PATCH] serial: tegra: Fixes lower tolerance baud rate limit for older tegra chips introduced by d781ec21bae6

[Greg KH]

On Mon, Nov 22, 2021 at 01:44:26PM +0100, Patrik John wrote:
> The current implementation uses 0 as lower limit for the baud rate tolerance which contradicts the initial commit description (https://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git/commit/drivers/tty/serial/serial-tegra.c?h=for-next&id=d781ec21bae6ff8f9e07682e8947a654484611f5) of +4/-4% tolerance for older tegra chips other than Tegra186 and Tegra194.
> This causes issues on UART initilization as soon as the actual baud rate clock is slightly lower than required which we have seen on the Tegra124-based Toradex Apalis TK1 which also uses tegra30-hsuart as compatible in the DT serial node (for reference line 1540ff https://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git/tree/arch/arm/boot/dts/tegra124-apalis-v1.2.dtsi?h=for-next)

All of these links will break in a few days.

And a line number is not "1540ff" :(

> 
> The standard baud rate tolerance limits are also stated in the tegra20-hsuart driver description (https://www.kernel.org/doc/Documentation/devicetree/bindings/serial/nvidia%2Ctegra20-hsuart.txt).

You can just reference a file in the kernel source tree directly, no
need to go back to kernel.org

> 
> The previously introduced check_rate_in_range() always fails due to the lower limit set to 0 even if the actual baud rate is within the required -4% tolerance.

Can you please wrap your changelog text at 72 columns like git asked you
to when you committed the change to your local tree?

> 
> static int tegra_check_rate_in_range(struct tegra_uart_port *tup)
> {
>     long diff;
>     diff = ((long)(tup->configured_rate - tup->required_rate) * 10000)
>         / tup->required_rate;
>     if (diff < (tup->cdata->error_tolerance_low_range * 100) ||
>         diff > (tup->cdata->error_tolerance_high_range * 100)) {
>         dev_err(tup->uport.dev,
>             "configured baud rate is out of range by %ld", diff);
>         return -EIO;
>     }
>     return 0;
> }

I do not understand, why is this code in the changelog?

> 
> Changing the lower tolerance limit to the actual -4% resolved the issues for the Tegra124 and should resolve potential issues for other Tegra20/Tegra30 based platforms as well.
> 
> Signed-off-by: Patrik John <patrik.john@xxxxxxxxxx>

What commit does this fix?  Should it have a "Fixes:" tag in it?

And should it go to stable kernel(s)?

Also, this is a v2 patch, please include below the --- line what changed
from the previous version when you resend v3.

thanks,

greg k-h