Re: [PATCH 3/5] init: use do_mount() instead of ksys_mount()

[Dominik Brodowski]

On Mon, Dec 16, 2019 at 10:45:56AM +0100, Borislav Petkov wrote:
> On Thu, Dec 12, 2019 at 07:14:20PM +0100, Dominik Brodowski wrote:
> > diff --git a/init/do_mounts.c b/init/do_mounts.c
> > index 43f6d098c880..f55cbd9cb818 100644
> > --- a/init/do_mounts.c
> > +++ b/init/do_mounts.c
> > @@ -387,12 +387,25 @@ static void __init get_fs_names(char *page)
> >  	*s = '\0';
> >  }
> >  
> > -static int __init do_mount_root(char *name, char *fs, int flags, void *data)
> > +static int __init do_mount_root(const char *name, const char *fs,
> > +				 const int flags, const void *data)
> >  {
> >  	struct super_block *s;
> > -	int err = ksys_mount(name, "/root", fs, flags, data);
> > -	if (err)
> > -		return err;
> > +	char *data_page;
> > +	struct page *p;
> > +	int ret;
> > +
> > +	/* do_mount() requires a full page as fifth argument */
> > +	p = alloc_page(GFP_KERNEL);
> > +	if (!p)
> > +		return -ENOMEM;
> > +
> > +	data_page = page_address(p);
> 	^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> That doesn't work in my guest as it gives a funny address:
> 
> [    3.155314] mount_block_root: entry
> [    3.155868] mount_block_root: fs_name: [ext3]
> [    3.156512] do_mount_root: will copy data page: 0x00000000adf0ddb8
> 
> leading to the splat below.

Does

https://lore.kernel.org/lkml/CAHk-=wh8VLe3AEKhz=1bzSO=1fv4EM71EhufxuC=Gp=+bLhXoA@xxxxxxxxxxxxxx/

solve the issue?

Thanks,
	Dominik