[PATCH v3 2/4] fbdev/efifb: Use screen_info pointer from device

[Thomas Zimmermann]

Use the screen_info instance from the device instead of dereferencing
the global screen_info state. Decouples the driver from per-architecture
code. Duplicated the screen_info data, so that efifb can modify it at
will.

v2:
        * comment on devm_kmemdup() usage (Javier)

Signed-off-by: Thomas Zimmermann <tzimmerm...@suse.de>
Tested-by: Sui Jingfeng <sui.jingf...@linux.dev>
Reviewed-by: Javier Martinez Canillas <javi...@redhat.com>
---
 drivers/video/fbdev/efifb.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/video/fbdev/efifb.c b/drivers/video/fbdev/efifb.c
index 6b3b5cca1d41..10fc14ad5d12 100644
--- a/drivers/video/fbdev/efifb.c
+++ b/drivers/video/fbdev/efifb.c
@@ -359,7 +359,7 @@ static u64 bar_offset;
 
 static int efifb_probe(struct platform_device *dev)
 {
-       struct screen_info *si = &screen_info;
+       struct screen_info *si;
        struct fb_info *info;
        struct efifb_par *par;
        int err, orientation;
@@ -369,6 +369,18 @@ static int efifb_probe(struct platform_device *dev)
        char *option = NULL;
        efi_memory_desc_t md;
 
+       /*
+        * If we fail probing the device, the kernel might try a different
+        * driver. We get a copy of the attached screen_info, so that we can
+        * modify its values without affecting later drivers.
+        */
+       si = dev_get_platdata(&dev->dev);
+       if (!si)
+               return -ENODEV;
+       si = devm_kmemdup(&dev->dev, si, sizeof(*si), GFP_KERNEL);
+       if (!si)
+               return -ENOMEM;
+
        if (si->orig_video_isVGA != VIDEO_TYPE_EFI || pci_dev_disabled)
                return -ENODEV;
 
-- 
2.43.0