linux-kernel - Re: [PATCH] tools/nolibc: ensure fast64 integer types have 64 bits

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZHx8rshTueq8V5fC@1wt.eu>
Date:   Sun, 4 Jun 2023 13:59:42 +0200
From:   Willy Tarreau <w@....eu>
To:     Thomas Weißschuh <linux@...ssschuh.net>
Cc:     Shuah Khan <shuah@...nel.org>,
        "Paul E. McKenney" <paulmck@...nel.org>,
        Vincent Dagonneau <v@....io>, linux-kernel@...r.kernel.org,
        linux-kselftest@...r.kernel.org
Subject: Re: [PATCH] tools/nolibc: ensure fast64 integer types have 64 bits

On Sun, Jun 04, 2023 at 12:50:03PM +0200, Willy Tarreau wrote:
> On Tue, May 30, 2023 at 11:18:00AM +0200, Thomas Weißschuh wrote:
> > On 32bit platforms size_t is not enough to represent [u]int_fast64_t.
> > 
> > Fixes: 3e9fd4e9a1d5 ("tools/nolibc: add integer types and integer limit macros")
> > Signed-off-by: Thomas Weißschuh <linux@...ssschuh.net>
> > ---
> > Cc: Vincent Dagonneau <v@....io>
> > 
> > Note: We could also fall back to compiler-provided data like:
> > 
> > __UINT_FAST{8,16,32,64}_{TYPE,MIN,MAX}__
> 
> I'm fine with the way you did it. I'm wondering how we managed to miss
> this one given the tests in place!

BTW, it failed on 32-bit platforms:

4407 tests passed, 84 skipped, 63 failed
$ grep '^linux_arch\|FAIL' test14.out
linux_arch=i386 qemu_arch=i386 gcc_arch=i386
52 limit_int_fast64_min = -2147483648                           [FAIL]
53 limit_int_fast64_max = 2147483647                            [FAIL]
54 limit_uint_fast64_max = 4294967295                           [FAIL]

The reason is that the constants also have to be adjusted. With the fix
below everything works right:

--- a/tools/include/nolibc/stdint.h
+++ b/tools/include/nolibc/stdint.h
@@ -84,17 +84,17 @@ typedef uint64_t          uintmax_t;
 #define  INT_FAST8_MIN   INT8_MIN
 #define INT_FAST16_MIN   INTPTR_MIN
 #define INT_FAST32_MIN   INTPTR_MIN
-#define INT_FAST64_MIN   INTPTR_MIN
+#define INT_FAST64_MIN   INT64_MIN
 
 #define  INT_FAST8_MAX   INT8_MAX
 #define INT_FAST16_MAX   INTPTR_MAX
 #define INT_FAST32_MAX   INTPTR_MAX
-#define INT_FAST64_MAX   INTPTR_MAX
+#define INT_FAST64_MAX   INT64_MAX
 
 #define  UINT_FAST8_MAX  UINT8_MAX
 #define UINT_FAST16_MAX  SIZE_MAX
 #define UINT_FAST32_MAX  SIZE_MAX
-#define UINT_FAST64_MAX  SIZE_MAX
+#define UINT_FAST64_MAX  UINT64_MAX
 
 #ifndef INT_MIN
 #define INT_MIN          (-__INT_MAX__ - 1)


4470 tests passed, 84 skipped, 0 failed
$ grep '^linux_arch\|fast64' test15.out 
linux_arch=i386 qemu_arch=i386 gcc_arch=i386
52 limit_int_fast64_min = -9223372036854775808                   [OK]
53 limit_int_fast64_max = 9223372036854775807                    [OK]
54 limit_uint_fast64_max = -1                                    [OK]

If you're fine with it, I'll squash it into your patch.

Thanks,
Willy