[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240913191428.1539653-1-javierm@redhat.com>
Date: Fri, 13 Sep 2024 21:13:59 +0200
From: Javier Martinez Canillas <javierm@...hat.com>
To: linux-kernel@...r.kernel.org
Cc: chrome-platform@...ts.linux.dev,
intel-gfx@...ts.freedesktop.org,
dri-devel@...ts.freedesktop.org,
Hugues Bruant <hugues.bruant@...il.com>,
Julius Werner <jwerner@...omium.org>,
Brian Norris <briannorris@...omium.org>,
Thomas Zimmermann <tzimmermann@...e.de>,
Borislav Petkov <bp@...en8.de>,
Javier Martinez Canillas <javierm@...hat.com>,
Tzung-Bi Shih <tzungbi@...nel.org>
Subject: [PATCH v2] firmware: coreboot: Don't register a pdev if screen_info data is present
On Coreboot platforms, a system framebuffer may be provided to the Linux
kernel by filling a LB_TAG_FRAMEBUFFER entry in the Coreboot table. But
a Coreboot payload (e.g: SeaBIOS) could also provide its own framebuffer
information to the Linux kernel.
If that's the case, arch x86 boot code will fill the global screen_info
data and this used by the Generic System Framebuffers (sysfb) framework,
to register a platform device with pdata about the system's framebuffer.
But later, the framebuffer_coreboot driver will try to do the same and
attempt to register a "simple-framebuffer" platform device (using the
information from the Coreboot table), which will lead to an error due a
device with the same name already being registered:
sysfs: cannot create duplicate filename '/bus/platform/devices/simple-framebuffer.0'
...
coreboot: could not register framebuffer
framebuffer coreboot8: probe with driver framebuffer failed with error -17
To prevent this issue, make the framebuffer_core driver to not register
a platform device if the global struct screen_info data has been filled.
Reported-by: Brian Norris <briannorris@...omium.org>
Closes: https://lore.kernel.org/all/ZuCG-DggNThuF4pj@b20ea791c01f/T/#ma7fb65acbc1a56042258adac910992bb225a20d2
Suggested-by: Julius Werner <jwerner@...omium.org>
Signed-off-by: Javier Martinez Canillas <javierm@...hat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@...e.de>
Reviewed-by: Brian Norris <briannorris@...omium.org>
---
Changes in v2:
- Declare the struct screen_info as constant variable (Thomas Zimmermann).
- Use screen_info_video_type() instead of checking the screen_info video
types directly (Thomas Zimmermann).
- Fix missing "device" word in a comment (Brian Norris).
- Fix some mispellings in a comment (Brian Norris).
- Change error code returned from -EINVAL to -ENODEV (Brian Norris).
drivers/firmware/google/framebuffer-coreboot.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/drivers/firmware/google/framebuffer-coreboot.c b/drivers/firmware/google/framebuffer-coreboot.c
index daadd71d8ddd..49eeb13814e4 100644
--- a/drivers/firmware/google/framebuffer-coreboot.c
+++ b/drivers/firmware/google/framebuffer-coreboot.c
@@ -15,6 +15,7 @@
#include <linux/module.h>
#include <linux/platform_data/simplefb.h>
#include <linux/platform_device.h>
+#include <linux/screen_info.h>
#include "coreboot_table.h"
@@ -27,8 +28,10 @@ static int framebuffer_probe(struct coreboot_device *dev)
int i;
u32 length;
struct lb_framebuffer *fb = &dev->framebuffer;
+ const struct screen_info *si = &screen_info;
struct platform_device *pdev;
struct resource res;
+ unsigned int type;
struct simplefb_platform_data pdata = {
.width = fb->x_resolution,
.height = fb->y_resolution,
@@ -36,6 +39,20 @@ static int framebuffer_probe(struct coreboot_device *dev)
.format = NULL,
};
+ /*
+ * On Coreboot systems, the advertised LB_TAG_FRAMEBUFFER entry
+ * in the Coreboot table should only be used if the payload did
+ * not pass a framebuffer information to the Linux kernel.
+ *
+ * If the global screen_info data has been filled, the Generic
+ * System Framebuffers (sysfb) will already register a platform
+ * device and pass that screen_info as platform_data to a driver
+ * that can scan-out using the system provided framebuffer.
+ */
+ type = screen_info_video_type(si);
+ if (type)
+ return -ENODEV;
+
if (!fb->physical_address)
return -ENODEV;
--
2.46.0
Powered by blists - more mailing lists