Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <b79c83af-e9fc-9fa0-dff7-f3a8a39887ff@acm.org>
Date:   Tue, 23 Aug 2022 17:09:12 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.11.0
From:   Bart Van Assche <bvanassche@acm.org>
Subject: Re: [for-linus][PATCH 01/10] tracing: Suppress sparse warnings
 triggered by is_signed_type()
To:     Linus Torvalds <torvalds@linux-foundation.org>
Cc:     Rasmus Villemoes <linux@rasmusvillemoes.dk>,
        Steven Rostedt <rostedt@goodmis.org>,
        linux-kernel@vger.kernel.org, Ingo Molnar <mingo@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Christoph Hellwig <hch@lst.de>,
        Luc Van Oostenryck <luc.vanoostenryck@gmail.com>,
        Jens Axboe <axboe@kernel.dk>
References: <20220821000737.328590235@goodmis.org>
 <20220821000844.510643400@goodmis.org>
 <CAHk-=wjsxu782N0P+oMu35N7rJAOeh3buQFWJaZHZTNmVSB=3Q@mail.gmail.com>
 <5700ac75-f6a9-877e-4011-9b314f12b5ab@acm.org>
 <CAHk-=wjqkWEr0MRO5hWuBoTDgNUj4qQK8V_Y36et=61mdPztJw@mail.gmail.com>
 <02daa3d6-2847-d7e0-e23e-411076c6d4db@rasmusvillemoes.dk>
 <0163b361-14bf-7b4c-751a-14f1a004b1a9@acm.org>
 <CAHk-=wjMLb30d0WT_RyKBCX+JBkg3QQU6pCYkrV8f58Ya4Rgzw@mail.gmail.com>
Content-Language: en-US
In-Reply-To: <CAHk-=wjMLb30d0WT_RyKBCX+JBkg3QQU6pCYkrV8f58Ya4Rgzw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 8/23/22 16:18, Linus Torvalds wrote:
> On Tue, Aug 23, 2022 at 3:05 PM Bart Van Assche <bvanassche@acm.org> wrote:
>>
>> Thank you Rasmus for having shared this information. Since sparse will
>> have to be modified anyway, how about extending it such that the bitwise
>> attribute can be removed from a type, e.g. via a new no_bitwise
>> attribute?
> 
> I think it's actually easier to just make sparse happy.
> 
> Can you try the sparse version at
> 
>     git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/sparse.git
> 
> which I just set up temporarily with some patches of mine. It also
> makes that '__cond_acquires' thing work that refcount_dec_and_lock()
> uses.
> 
> It does require that kernel change to make
> 
>    #define is_signed_type(type)   (((type)(-1)) <= (type)0)
> 
> in both places, since only "no bits set" and "all bits set" are
> special values for bitwise types.
> 
> Those patches of mine are fairly hacky, and I think Luc would probably
> do it differently, but apart from the very last one, they aren't
> actively disgusting.

Hi Linus,

I'm probably doing something wrong but even with sparse commit 658ee8e0f631
("unrestricted values are unrestricted even after a cast") I see warnings
being triggered by users of the is_signed_type() macro, warnings that
disappear if I change the definition of the is_signed_type() macro into 0:

$ make C=2 fs/f2fs/ </dev/null |& grep blk_opf_t
./include/trace/events/f2fs.h:1027:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1027:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1027:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1027:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1086:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1086:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1086:1: warning: restricted blk_opf_t degrades to integer
./include/trace/events/f2fs.h:1086:1: warning: restricted blk_opf_t degrades to integer

This is the kernel patch that I applied:

diff --git a/include/linux/overflow.h b/include/linux/overflow.h
index f1221d11f8e5..10c55f97e02b 100644
--- a/include/linux/overflow.h
+++ b/include/linux/overflow.h
@@ -30,7 +30,7 @@
   * https://mail-index.netbsd.org/tech-misc/2007/02/05/0000.html -
   * credit to Christian Biere.
   */
-#define is_signed_type(type)       (((type)(-1)) < (type)1)
+#define is_signed_type(type)       (((__force type)(-1)) <= (__force type)0)
  #define __type_half_max(type) ((type)1 << (8*sizeof(type) - 1 - is_signed_type(type)))
  #define type_max(T) ((T)((__type_half_max(T) - 1) + __type_half_max(T)))
  #define type_min(T) ((T)((T)-type_max(T)-(T)1))
diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
index b18759a673c6..c74cfa657025 100644
--- a/include/linux/trace_events.h
+++ b/include/linux/trace_events.h
@@ -9,6 +9,7 @@
  #include <linux/hardirq.h>
  #include <linux/perf_event.h>
  #include <linux/tracepoint.h>
+#include <linux/overflow.h>

  struct trace_array;
  struct array_buffer;
@@ -814,8 +815,6 @@ extern int trace_add_event_call(struct trace_event_call *call);
  extern int trace_remove_event_call(struct trace_event_call *call);
  extern int trace_event_get_offsets(struct trace_event_call *call);

-#define is_signed_type(type)   (((type)(-1)) < (type)1)
-
  int ftrace_set_clr_event(struct trace_array *tr, char *buf, int set);
  int trace_set_clr_event(const char *system, const char *event, int set);
  int trace_array_set_clr_event(struct trace_array *tr, const char *system,

Thanks,

Bart.