diff mbox

[resend,v4] dell_rbu: make firmware payload memory uncachable

Message ID 5318cc53-31b4-5c83-abb7-c3ff5d892af2@gmail.com (mailing list archive)
State Superseded, archived
Delegated to: Andy Shevchenko
Headers show

Commit Message

stuart hayes July 6, 2018, 4:18 p.m. UTC
The dell_rbu driver takes firmware update payloads and puts them in memory so
the system BIOS can find them after a reboot.  This sometimes fails (though
rarely), because the memory containing the payload is in the CPU cache but
never gets written back to main memory before the system is rebooted (CPU
cache contents are lost on reboot).

With this patch, the payload memory will be changed to uncachable to ensure
that the payload is actually in main memory before the system is rebooted.

Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Reviewed-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
v2 Added include, removed extra parentheses
v3 Corrected formatting and include line
v4 Moved set_memory_uc() outside the while loop so that the memory is
   definitely allocated before it is set to uncachable

This driver has no maintainer.

Comments

Limonciello, Mario July 6, 2018, 4:28 p.m. UTC | #1
> -----Original Message-----

> From: platform-driver-x86-owner@vger.kernel.org [mailto:platform-driver-x86-

> owner@vger.kernel.org] On Behalf Of Stuart Hayes

> Sent: Friday, July 6, 2018 11:19 AM

> To: linux-kernel@vger.kernel.org

> Cc: Takashi Iwai; Andy Shevchenko; dvhart@infradead.org; Platform Driver

> Subject: [PATCH resend v4] dell_rbu: make firmware payload memory uncachable

> 

> The dell_rbu driver takes firmware update payloads and puts them in memory so

> the system BIOS can find them after a reboot.  This sometimes fails (though

> rarely), because the memory containing the payload is in the CPU cache but

> never gets written back to main memory before the system is rebooted (CPU

> cache contents are lost on reboot).

> 

> With this patch, the payload memory will be changed to uncachable to ensure

> that the payload is actually in main memory before the system is rebooted.

> 

> Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>

> Reviewed-by: Takashi Iwai <tiwai@suse.de>

> Signed-off-by: Takashi Iwai <tiwai@suse.de>


Reviewed-by: Mario Limonciello <mario.limonciello@dell.com>


> ---

> v2 Added include, removed extra parentheses

> v3 Corrected formatting and include line

> v4 Moved set_memory_uc() outside the while loop so that the memory is

>    definitely allocated before it is set to uncachable

> 

> This driver has no maintainer.


Stuart,

Outside of this patch, I think it would make sense to send some follow up patch
series that do the following:

1) Let you take over as maintainer (Update MAINTAINERS)
2) Move this driver to platform-x86
3) Update dell_rbu.txt with the correct URLs and mention that it's only for enterprise
hardware.    Client hardware has instead opted to support UEFI capsules for updates
from Linux.

Since it's really a platform feature and specific to only Dell enterprise hardware
I think it makes sense to move over to an area with subsystem maintainers that
can help to review patches as needed.

> 

> 

> diff --git a/drivers/firmware/dell_rbu.c b/drivers/firmware/dell_rbu.c

> index 2f452f1f7c8a..53f27a6e2d76 100644

> --- a/drivers/firmware/dell_rbu.c

> +++ b/drivers/firmware/dell_rbu.c

> @@ -45,6 +45,7 @@

>  #include <linux/moduleparam.h>

>  #include <linux/firmware.h>

>  #include <linux/dma-mapping.h>

> +#include <asm/set_memory.h>

> 

>  MODULE_AUTHOR("Abhay Salunke <abhay_salunke@dell.com>");

>  MODULE_DESCRIPTION("Driver for updating BIOS image on DELL systems");

> @@ -181,6 +182,11 @@ static int create_packet(void *data, size_t length)

>  			packet_data_temp_buf = NULL;

>  		}

>  	}

> +	/*

> +	 * set to uncachable or it may never get written back before reboot

> +	 */

> +	set_memory_uc((unsigned long)packet_data_temp_buf, 1 << ordernum);

> +

>  	spin_lock(&rbu_data.lock);

> 

>  	newpacket->data = packet_data_temp_buf;

> @@ -349,6 +355,8 @@ static void packet_empty_list(void)

>  		 * to make sure there are no stale RBU packets left in memory

>  		 */

>  		memset(newpacket->data, 0, rbu_data.packetsize);

> +		set_memory_wb((unsigned long)newpacket->data,

> +			1 << newpacket->ordernum);

>  		free_pages((unsigned long) newpacket->data,

>  			newpacket->ordernum);

>  		kfree(newpacket);
Andy Shevchenko July 13, 2018, 2:26 p.m. UTC | #2
On Fri, Jul 6, 2018 at 7:28 PM,  <Mario.Limonciello@dell.com> wrote:

> Reviewed-by: Mario Limonciello <mario.limonciello@dell.com>

Pushed to my review and testing queue, thanks!

>> This driver has no maintainer.
>
> Stuart,
>
> Outside of this patch, I think it would make sense to send some follow up patch
> series that do the following:
>
> 1) Let you take over as maintainer (Update MAINTAINERS)

> 2) Move this driver to platform-x86

Darren, what do you think about this step?

If we would go this way, first move, then take maintainership to reduce a churn.

> 3) Update dell_rbu.txt with the correct URLs and mention that it's only for enterprise
> hardware.    Client hardware has instead opted to support UEFI capsules for updates
> from Linux.
>
> Since it's really a platform feature and specific to only Dell enterprise hardware
> I think it makes sense to move over to an area with subsystem maintainers that
> can help to review patches as needed.
diff mbox

Patch

diff --git a/drivers/firmware/dell_rbu.c b/drivers/firmware/dell_rbu.c
index 2f452f1f7c8a..53f27a6e2d76 100644
--- a/drivers/firmware/dell_rbu.c
+++ b/drivers/firmware/dell_rbu.c
@@ -45,6 +45,7 @@ 
 #include <linux/moduleparam.h>
 #include <linux/firmware.h>
 #include <linux/dma-mapping.h>
+#include <asm/set_memory.h>
 
 MODULE_AUTHOR("Abhay Salunke <abhay_salunke@dell.com>");
 MODULE_DESCRIPTION("Driver for updating BIOS image on DELL systems");
@@ -181,6 +182,11 @@  static int create_packet(void *data, size_t length)
 			packet_data_temp_buf = NULL;
 		}
 	}
+	/*
+	 * set to uncachable or it may never get written back before reboot
+	 */
+	set_memory_uc((unsigned long)packet_data_temp_buf, 1 << ordernum);
+
 	spin_lock(&rbu_data.lock);
 
 	newpacket->data = packet_data_temp_buf;
@@ -349,6 +355,8 @@  static void packet_empty_list(void)
 		 * to make sure there are no stale RBU packets left in memory
 		 */
 		memset(newpacket->data, 0, rbu_data.packetsize);
+		set_memory_wb((unsigned long)newpacket->data,
+			1 << newpacket->ordernum);
 		free_pages((unsigned long) newpacket->data,
 			newpacket->ordernum);
 		kfree(newpacket);