Re: [PATCH] [BUGFIX] media: vb2: Fix videobuf2 to map correct area

[Marek Szyprowski]

Hi Masami,

On 2018-02-05 03:30, Masami Hiramatsu wrote:
> Fixes vb2_vmalloc_get_userptr() to ioremap correct area.
> Since the current code does ioremap the page address, if the offset > 0,
> it does not do ioremap the last page and results in kernel panic.
>
> This fixes to pass the page address + offset to ioremap so that ioremap
> can map correct area. Also, this uses __pfn_to_phys() to get the physical
> address of given PFN.
>
> Signed-off-by: Masami Hiramatsu <mhiramat@xxxxxxxxxx>
> Reported-by: Takao Orito <orito.takao@xxxxxxxxxxxxx>
> ---
>   drivers/media/v4l2-core/videobuf2-vmalloc.c |    2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/media/v4l2-core/videobuf2-vmalloc.c b/drivers/media/v4l2-core/videobuf2-vmalloc.c
> index 3a7c80cd1a17..896f2f378b40 100644
> --- a/drivers/media/v4l2-core/videobuf2-vmalloc.c
> +++ b/drivers/media/v4l2-core/videobuf2-vmalloc.c
> @@ -106,7 +106,7 @@ static void *vb2_vmalloc_get_userptr(struct device *dev, unsigned long vaddr,
>   			if (nums[i-1] + 1 != nums[i])
>   				goto fail_map;
>   		buf->vaddr = (__force void *)
> -				ioremap_nocache(nums[0] << PAGE_SHIFT, size);
> +			ioremap_nocache(__pfn_to_phys(nums[0]) + offset, size);

Thanks for reporting this issue. However the above line doesn't look like
a proper fix. Please note that at the end of that function there is already
"buf->vaddr += offset;".

IMHO the proper fix is to create a larger mapping, which would include the
in-page start offset:

ioremap_nocache(__pfn_to_phys(nums[0]), offset + size);

BTW, thanks for updating "<< PAGE_SHIFT" to better __pfn_to_phys() macro!

>   	} else {
>   		buf->vaddr = vm_map_ram(frame_vector_pages(vec), n_pages, -1,
>   					PAGE_KERNEL);

Best regards
--
Marek Szyprowski, PhD
Samsung R&D Institute Poland