[PATCH] spi/acpi: avoid spurious matches during slave enumeration

[PATCH] spi/acpi: avoid spurious matches during slave enumeration

[Ard Biesheuvel]

In the new SPI ACPI slave enumeration code, we use the value of
lookup.max_speed_khz as a flag to decide whether a match occurred.
However, doing so only makes sense if we initialize its value to
zero beforehand, or otherwise, random junk from the stack will
cause spurious matches.

So zero initialize the lookup struct fully, and only set the non-zero
members explicitly.

Fixes: 4c3c59544f33 ("spi/acpi: enumerate all SPI slaves in the namespace")
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: andy.shevchenko@gmail.com
Cc: masahisa.kojima@linaro.org
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Cc: linux-acpi@vger.kernel.org
Cc: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 drivers/spi/spi.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index a31e1e291335..4057f256ef76 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -1953,7 +1953,7 @@ static acpi_status acpi_register_spi_device(struct spi_controller *ctlr,
 {
 	acpi_handle parent_handle = NULL;
 	struct list_head resource_list;
-	struct acpi_spi_lookup lookup;
+	struct acpi_spi_lookup lookup = {};
 	struct spi_device *spi;
 	int ret;
 
@@ -1962,8 +1962,6 @@ static acpi_status acpi_register_spi_device(struct spi_controller *ctlr,
 		return AE_OK;
 
 	lookup.ctlr		= ctlr;
-	lookup.mode		= 0;
-	lookup.bits_per_word	= 0;
 	lookup.irq		= -1;
 
 	INIT_LIST_HEAD(&resource_list);
-- 
2.20.1

Re: [PATCH] spi/acpi: avoid spurious matches during slave enumeration

[Jarkko Nikula]

On 6/20/19 3:36 PM, Ard Biesheuvel wrote:
> In the new SPI ACPI slave enumeration code, we use the value of
> lookup.max_speed_khz as a flag to decide whether a match occurred.
> However, doing so only makes sense if we initialize its value to
> zero beforehand, or otherwise, random junk from the stack will
> cause spurious matches.
> 
> So zero initialize the lookup struct fully, and only set the non-zero
> members explicitly.
> 
> Fixes: 4c3c59544f33 ("spi/acpi: enumerate all SPI slaves in the namespace")
> Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
> Cc: andy.shevchenko@gmail.com
> Cc: masahisa.kojima@linaro.org
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: Jarkko Nikula <jarkko.nikula@linux.intel.com>
> Cc: linux-acpi@vger.kernel.org
> Cc: Lukas Wunner <lukas@wunner.de>
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> ---
>   drivers/spi/spi.c | 4 +---
>   1 file changed, 1 insertion(+), 3 deletions(-)
> 
Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>

Re: [PATCH] spi/acpi: avoid spurious matches during slave enumeration

[Mika Westerberg]

On Thu, Jun 20, 2019 at 02:36:49PM +0200, Ard Biesheuvel wrote:
> In the new SPI ACPI slave enumeration code, we use the value of
> lookup.max_speed_khz as a flag to decide whether a match occurred.
> However, doing so only makes sense if we initialize its value to
> zero beforehand, or otherwise, random junk from the stack will
> cause spurious matches.
> 
> So zero initialize the lookup struct fully, and only set the non-zero
> members explicitly.
> 
> Fixes: 4c3c59544f33 ("spi/acpi: enumerate all SPI slaves in the namespace")
> Cc: Mika Westerberg <mika.westerberg@linux.intel.com>

Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>