Re: [PATCH] ACPI: don't attempt to use hest_tab unless !acpi_pci_disabled

[Andres Salomon <dilinger@queued.net>]

On Sat, 15 Jan 2011 21:48:43 +0100
"Rafael J. Wysocki" <rjw@sisk.pl> wrote:

> On Saturday, January 15, 2011, Andres Salomon wrote:
> > On Sat, 15 Jan 2011 21:38:10 +0100
> > "Rafael J. Wysocki" <rjw@sisk.pl> wrote:
> > 
> > > On Saturday, January 15, 2011, Andres Salomon wrote:
> > > > On Sat, 15 Jan 2011 14:01:18 +0100
> > > > "Rafael J. Wysocki" <rjw@sisk.pl> wrote:
> > > > 
> > > > > On Saturday, January 15, 2011, Andres Salomon wrote:
> > > > > > On Fri, 14 Jan 2011 23:24:27 -0500 (EST)
> > > > > > Len Brown <lenb@kernel.org> wrote:
> > > > > > 
> > > > > > [...]
> > > > > > > > This patch causes apei_hest_parse to check both
> > > > > > > > hest_disabled and acpi_pci_disabled before continuing.
> > > > > > > > With it, the XO-1 boots properly.
> > > > > > > 
> > > > > > > The X0-1 has no ACPI support, and thus you are running
> > > > > > > with acpi_disabled=1, yes?
> > > > > > 
> > > > > > The XO-1 has no ACPI support, but ACPI support is enabled
> > > > > > in the kernel (as the XO-1.5 does have ACPI support, and
> > > > > > the same kernel is used between both) and I'm not passing
> > > > > > any arguments to the kernel regarding it. The kernel's ACPI
> > > > > > code detects that it's not there and disables it (the
> > > > > > kernel message that's seen is "ACPI: Interpreter disabled.")
> > > > > > 
> > > > > > This is what sets apci_pci_disabled to 1; I'm not doing it
> > > > > > manually.
> > > > > 
> > > > > Hmm.  Does the appended patch help instead?
> > > > 
> > > > Nope.  acpi_hest_init returns immediately if acpi_disabled, and
> > > > thus never sets hest_disable.  You either need to set
> > > > hest_disable, or ensure apei_hest_parse checks more than just
> > > > hest_disable.
> > > 
> > > OK, so the new one below should help, right?
> > 
> > Yes, that looks better
> 
> OK
> 
> > (though you probably just want to use 'goto err;' instead).
> 
> Indeed.
> 
> > I can test it later today.
> 
> Cool, thanks.  Please try the version below with the 'goto'.
> 
> Rafael


Yep, just tested it out; it works fine.


> 
> ---
>  drivers/acpi/apei/hest.c |    6 +++---
>  drivers/acpi/pci_root.c  |    4 ++--
>  2 files changed, 5 insertions(+), 5 deletions(-)
> 
> Index: linux-2.6/drivers/acpi/pci_root.c
> ===================================================================
> --- linux-2.6.orig/drivers/acpi/pci_root.c
> +++ linux-2.6/drivers/acpi/pci_root.c
> @@ -631,11 +631,11 @@ static int acpi_pci_root_remove(struct a
>  
>  static int __init acpi_pci_root_init(void)
>  {
> +	acpi_hest_init();
> +
>  	if (acpi_pci_disabled)
>  		return 0;
>  
> -	acpi_hest_init();
> -
>  	pci_acpi_crs_quirks();
>  	if (acpi_bus_register_driver(&acpi_pci_root_driver) < 0)
>  		return -ENODEV;
> Index: linux-2.6/drivers/acpi/apei/hest.c
> ===================================================================
> --- linux-2.6.orig/drivers/acpi/apei/hest.c
> +++ linux-2.6/drivers/acpi/apei/hest.c
> @@ -201,14 +201,14 @@ void __init acpi_hest_init(void)
>  	int rc = -ENODEV;
>  	unsigned int ghes_count = 0;
>  
> -	if (acpi_disabled)
> -		return;
> -
>  	if (hest_disable) {
>  		pr_info(HEST_PFX "Table parsing disabled.\n");
>  		return;
>  	}
>  
> +	if (acpi_disabled)
> +		goto err;
> +
>  	status = acpi_get_table(ACPI_SIG_HEST, 0,
>  				(struct acpi_table_header
> **)&hest_tab); if (status == AE_NOT_FOUND) {