Re: [PATCH] fbdev: suppress warning when assigning vga-save/restore base
From: Tomi Valkeinen <hidden>
Date: 2014-02-10 11:23:58
Also in:
linux-fbdev
Attachments
- signature.asc [application/pgp-signature] 901 bytes
From: Tomi Valkeinen <hidden>
Date: 2014-02-10 11:23:58
Also in:
linux-fbdev
Hi David, On 04/08/13 20:25, David Herrmann wrote:
If drivers use "struct resource" objects to retrieve the vga-base, they
must correctly cast the integer to pointer. With x86+PAE we have 32bit
pointers but 64bit resource_size_t. Hence, cast it to "unsigned long"
before casting to "void*" to suppress warnings due to size differences.
As IO addresses are always low addresses, we can safely drop the higher
part of the address. This is what these drivers did before, anyway.
Signed-off-by: David Herrmann <redacted>
Reported-by: H. Peter Anvin <hpa@zytor.com>
---
Hi
hpa reported build-warnings on i386+PAE:
/home/hpa/kernel/distwork/drivers/video/arkfb.c:1019:23: warning: cast to
pointer from integer of different size [-Wint-to-pointer-cast]
par->state.vgabase = (void __iomem *) vga_res.start;
^
/home/hpa/kernel/distwork/drivers/video/s3fb.c: In function ‘s3_pci_probe’:
/home/hpa/kernel/distwork/drivers/video/s3fb.c:1186:23: warning: cast to pointer
from integer of different size [-Wint-to-pointer-cast]
par->state.vgabase = (void __iomem *) vga_res.start;
^
This is due to resource_size_t being 64bit but "void*" 32bit. This patch tries
to suppress these warnings but I am not really comfortable fixing this. I have
no idea whether my assumption (IO address are 32bit) is right. Please verify.
@David: The following 3 commits of yours presumably introduced the warnings. I
would be glad if you could review this:What was the conclusion on this patch? Should I apply for 3.14 fixes? Tomi