Re: [PATCH 6 of 6] tpm: move infineon driver off pci_dev

[Matthieu CASTET <castet.matthieu@free.fr>]

Hi,

nice to see that the tpm framework was cleanned.

Le Tue, 25 Oct 2005 10:22:18 -0500, Kylene Jo Hall a écrit :

> ---
> 
> --- linux-2.6.14-rc4/drivers/char/tpm/tpm_infineon.c	2005-10-19 17:03:52.000000000 -0500
> +++ linux-2.6.13-tpm/drivers/char/tpm/tpm_infineon.c	2005-10-21 13:07:24.000000000 -0500
> @@ -14,6 +14,7 @@
>   * License.
>   */
>  
> +#include <acpi/acpi_bus.h>
Why is it needed ?



> -/* These values will be filled after PnP-call */
> +/* These values will be filled after ACPI-call */
Why not keep PnP ?


> -MODULE_DEVICE_TABLE(pnp, tpm_pnp_tbl);
>  
Why removing that ?
This will allow hotplug and udev to auto load the module.


> +
> +	/* read IO-ports from ACPI */
> +	TPM_INF_ADDR = (pnp_port_start(dev, 0) & 0xff);
> +	TPM_INF_DATA = ((TPM_INF_ADDR + 1) & 0xff);
No need to set mask, this is already done by pnp_port_start.
And I'll keep PNP instead of ACPI.


> +	tpm_inf.base = pnp_port_start(dev, 1);
Can't you be coherent ?
why not using tpm_inf.addr
or TPM_INF_BASE ?



> +	dev_info(&dev->dev, "Found %s with ID %s\n",
> +		 dev->name, dev_id->id);
> +	if (!((tpm_inf.base >> 8) & 0xff))
> +		tpm_inf.base = 0;
>  
>  	/* Make sure, we have received valid config ports */
You should also do :
pnp_port_flags(device, 0) & IORESOURCE_DISABLED) in order to check the
resources.


> +	.probe = tpm_inf_acpi_probe,
> +	.remove = tpm_inf_remove,
Not coherent : acpi vs nothing.
Again prefer pnp instead of acpi.


Matthieu