Re: linux-next: Tree for Oct 18 (objtool)

From: Christian Brauner <christian.brauner@ubuntu.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Randy Dunlap <rdunlap@infradead.org>,
	Stephen Rothwell <sfr@canb.auug.org.au>,
	Linux Next Mailing List <linux-next@vger.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	linux@rasmusvillemoes.dk, cyphar@cyphar.com,
	keescook@chromium.org
Subject: Re: linux-next: Tree for Oct 18 (objtool)
Date: Mon, 21 Oct 2019 15:19:48 +0200	[thread overview]
Message-ID: <20191021131947.yih3wqjbfroaj4dz@wittgenstein> (raw)
In-Reply-To: <20191021131149.GA19358@hirez.programming.kicks-ass.net>

On Mon, Oct 21, 2019 at 03:11:49PM +0200, Peter Zijlstra wrote:
> On Mon, Oct 21, 2019 at 02:35:49PM +0200, Peter Zijlstra wrote:
> > On Fri, Oct 18, 2019 at 08:33:11AM -0700, Randy Dunlap wrote:
> > > On 10/18/19 12:03 AM, Stephen Rothwell wrote:
> > > > Hi all,
> > > > 
> > > > Changes since 20191017:
> > > > 
> > > 
> > > on x86_64:
> > > lib/usercopy.o: warning: objtool: check_zeroed_user()+0x35f: call to __ubsan_handle_shift_out_of_bounds() with UACCESS enabled
> > 
> > Blergh... I suppose the below will fix that. I'm a bit conflicted on it
> > though, the alternative is annotating more ubsan crud.
> 
> By popular request; here's that alternative. Completely untested :-)

Am I not getting some mails? :)
I prefer this one as it allows us to avoid working around this in
usercopy.c. Should especially make if this potentially helps in other
cases as well?

> 
> ---
>  lib/ubsan.c           | 5 ++++-
>  tools/objtool/check.c | 1 +
>  2 files changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/lib/ubsan.c b/lib/ubsan.c
> index 39d5952c4273..0dce3ff45b5b 100644
> --- a/lib/ubsan.c
> +++ b/lib/ubsan.c
> @@ -359,9 +359,10 @@ void __ubsan_handle_shift_out_of_bounds(struct shift_out_of_bounds_data *data,
>  	struct type_descriptor *lhs_type = data->lhs_type;
>  	char rhs_str[VALUE_LENGTH];
>  	char lhs_str[VALUE_LENGTH];
> +	unsigned long flags = user_access_save();
>  
>  	if (suppress_report(&data->location))
> -		return;
> +		goto out;
>  
>  	ubsan_prologue(&data->location);
>  
> @@ -387,6 +388,8 @@ void __ubsan_handle_shift_out_of_bounds(struct shift_out_of_bounds_data *data,
>  			lhs_type->type_name);
>  
>  	ubsan_epilogue();
> +out:
> +	user_access_restore(flags);
>  }
>  EXPORT_SYMBOL(__ubsan_handle_shift_out_of_bounds);
>  
> diff --git a/tools/objtool/check.c b/tools/objtool/check.c
> index 543c068096b1..4768d91c6d68 100644
> --- a/tools/objtool/check.c
> +++ b/tools/objtool/check.c
> @@ -482,6 +482,7 @@ static const char *uaccess_safe_builtin[] = {
>  	"ubsan_type_mismatch_common",
>  	"__ubsan_handle_type_mismatch",
>  	"__ubsan_handle_type_mismatch_v1",
> +	"__ubsan_handle_shift_out_of_bounds",
>  	/* misc */
>  	"csum_partial_copy_generic",
>  	"__memcpy_mcsafe",