[linux-next:master 4911/5794] include/linux/bitfield.h:119:3: error: call to '__bad_mask' declared with attribute error: bad bitfield mask
6:58 a.m.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head: 92edc4aef86780a8ad01b092c6d6630bb3cb423d
commit: 4a2d447bea5bdb175e0defaefcba74fc1517dff0 [4911/5794] pwm: Add PWM driver for Intel
Keem Bay
config: arm64-randconfig-p002-20201116 (attached as .config)
compiler: aarch64-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O
~/bin/make.cross
chmod +x ~/bin/make.cross
#
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commi...
git remote add linux-next
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
git fetch --no-tags linux-next master
git checkout 4a2d447bea5bdb175e0defaefcba74fc1517dff0
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=arm64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp(a)intel.com>
All errors (new ones prefixed by >>):
In file included from drivers/pwm/pwm-keembay.c:16:
In function 'field_multiplier',
inlined from 'keembay_pwm_update_bits.isra.0' at
include/linux/bitfield.h:124:17:
> include/linux/bitfield.h:119:3: error: call to
'__bad_mask' declared with attribute error: bad bitfield mask
119 |
__bad_mask();
| ^~~~~~~~~~~~
In function 'field_multiplier',
inlined from 'keembay_pwm_update_bits.isra.0' at
include/linux/bitfield.h:154:1:
> include/linux/bitfield.h:119:3: error: call to
'__bad_mask' declared with attribute error: bad bitfield mask
119 |
__bad_mask();
| ^~~~~~~~~~~~
vim +/__bad_mask +119 include/linux/bitfield.h
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 43
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 44 #define __BF_FIELD_CHECK(_mask, _reg,
_val, _pfx) \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 45 ({ \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 46
BUILD_BUG_ON_MSG(!__builtin_constant_p(_mask), \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 47 _pfx "mask is not
constant"); \
e36488c83b6d871 Arnd Bergmann 2018-08-17 48 BUILD_BUG_ON_MSG((_mask) == 0, _pfx
"mask is zero"); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 49
BUILD_BUG_ON_MSG(__builtin_constant_p(_val) ? \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 50 ~((_mask) >> __bf_shf(_mask))
& (_val) : 0, \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 51 _pfx "value too large for the
field"); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 52 BUILD_BUG_ON_MSG((_mask) >
(typeof(_reg))~0ull, \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 53 _pfx "type of reg too small for
mask"); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 54 __BUILD_BUG_ON_NOT_POWER_OF_2((_mask)
+ \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 55 (1ULL <<
__bf_shf(_mask))); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 56 })
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 57
e31a50162feb352 Alex Elder 2020-03-12 58 /**
e31a50162feb352 Alex Elder 2020-03-12 59 * FIELD_MAX() - produce the maximum
value representable by a field
e31a50162feb352 Alex Elder 2020-03-12 60 * @_mask: shifted mask defining the
field's length and position
e31a50162feb352 Alex Elder 2020-03-12 61 *
e31a50162feb352 Alex Elder 2020-03-12 62 * FIELD_MAX() returns the maximum value
that can be held in the field
e31a50162feb352 Alex Elder 2020-03-12 63 * specified by @_mask.
e31a50162feb352 Alex Elder 2020-03-12 64 */
e31a50162feb352 Alex Elder 2020-03-12 65 #define FIELD_MAX(_mask) \
e31a50162feb352 Alex Elder 2020-03-12 66 ({ \
e31a50162feb352 Alex Elder 2020-03-12 67 __BF_FIELD_CHECK(_mask, 0ULL, 0ULL,
"FIELD_MAX: "); \
e31a50162feb352 Alex Elder 2020-03-12 68 (typeof(_mask))((_mask) >>
__bf_shf(_mask)); \
e31a50162feb352 Alex Elder 2020-03-12 69 })
e31a50162feb352 Alex Elder 2020-03-12 70
1697599ee301a52 Jakub Kicinski 2017-02-09 71 /**
1697599ee301a52 Jakub Kicinski 2017-02-09 72 * FIELD_FIT() - check if value fits in
the field
1697599ee301a52 Jakub Kicinski 2017-02-09 73 * @_mask: shifted mask defining the
field's length and position
1697599ee301a52 Jakub Kicinski 2017-02-09 74 * @_val: value to test against the
field
1697599ee301a52 Jakub Kicinski 2017-02-09 75 *
1697599ee301a52 Jakub Kicinski 2017-02-09 76 * Return: true if @_val can fit inside
@_mask, false if @_val is too big.
1697599ee301a52 Jakub Kicinski 2017-02-09 77 */
1697599ee301a52 Jakub Kicinski 2017-02-09 78 #define FIELD_FIT(_mask, _val) \
1697599ee301a52 Jakub Kicinski 2017-02-09 79 ({ \
444da3f52407d74 Jakub Kicinski 2020-08-10 80 __BF_FIELD_CHECK(_mask, 0ULL, 0ULL,
"FIELD_FIT: "); \
1697599ee301a52 Jakub Kicinski 2017-02-09 81 !((((typeof(_mask))_val) <<
__bf_shf(_mask)) & ~(_mask)); \
1697599ee301a52 Jakub Kicinski 2017-02-09 82 })
1697599ee301a52 Jakub Kicinski 2017-02-09 83
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 84 /**
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 85 * FIELD_PREP() - prepare a bitfield
element
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 86 * @_mask: shifted mask defining the
field's length and position
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 87 * @_val: value to put in the field
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 88 *
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 89 * FIELD_PREP() masks and shifts up the
value. The result should
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 90 * be combined with other fields of the
bitfield using logical OR.
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 91 */
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 92 #define FIELD_PREP(_mask, _val) \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 93 ({ \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 94 __BF_FIELD_CHECK(_mask, 0ULL, _val,
"FIELD_PREP: "); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 95 ((typeof(_mask))(_val) <<
__bf_shf(_mask)) & (_mask); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 96 })
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 97
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 98 /**
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 99 * FIELD_GET() - extract a bitfield
element
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 100 * @_mask: shifted mask defining the
field's length and position
7240767450d6d83 Masahiro Yamada 2017-10-03 101 * @_reg: value of entire bitfield
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 102 *
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 103 * FIELD_GET() extracts the field
specified by @_mask from the
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 104 * bitfield passed in as @_reg by masking
and shifting it down.
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 105 */
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 106 #define FIELD_GET(_mask, _reg) \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 107 ({ \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 108 __BF_FIELD_CHECK(_mask, _reg, 0U,
"FIELD_GET: "); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 109 (typeof(_mask))(((_reg) & (_mask))
>> __bf_shf(_mask)); \
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 110 })
3e9b3112ec74f19 Jakub Kicinski 2016-08-31 111
e7d4a95da86e0b0 Johannes Berg 2018-06-20 112 extern void
__compiletime_error("value doesn't fit into mask")
00b0c9b82663ac4 Al Viro 2017-12-14 113 __field_overflow(void);
00b0c9b82663ac4 Al Viro 2017-12-14 114 extern void __compiletime_error("bad
bitfield mask")
00b0c9b82663ac4 Al Viro 2017-12-14 115 __bad_mask(void);
00b0c9b82663ac4 Al Viro 2017-12-14 116 static __always_inline u64
field_multiplier(u64 field)
00b0c9b82663ac4 Al Viro 2017-12-14 117 {
00b0c9b82663ac4 Al Viro 2017-12-14 118 if ((field | (field - 1)) & ((field
| (field - 1)) + 1))
00b0c9b82663ac4 Al Viro 2017-12-14 @119 __bad_mask();
00b0c9b82663ac4 Al Viro 2017-12-14 120 return field & -field;
00b0c9b82663ac4 Al Viro 2017-12-14 121 }
00b0c9b82663ac4 Al Viro 2017-12-14 122 static __always_inline u64 field_mask(u64
field)
00b0c9b82663ac4 Al Viro 2017-12-14 123 {
00b0c9b82663ac4 Al Viro 2017-12-14 124 return field / field_multiplier(field);
00b0c9b82663ac4 Al Viro 2017-12-14 125 }
e31a50162feb352 Alex Elder 2020-03-12 126 #define
field_max(field) ((typeof(field))field_mask(field))
00b0c9b82663ac4 Al Viro 2017-12-14 127 #define
____MAKE_OP(type,base,to,from) \
00b0c9b82663ac4 Al Viro 2017-12-14 128 static __always_inline __##type
type##_encode_bits(base v, base field) \
00b0c9b82663ac4 Al Viro 2017-12-14 129 { \
e7d4a95da86e0b0 Johannes Berg 2018-06-20 130 if (__builtin_constant_p(v) &&
(v & ~field_mask(field))) \
00b0c9b82663ac4 Al Viro 2017-12-14 131 __field_overflow(); \
00b0c9b82663ac4 Al Viro 2017-12-14 132 return to((v & field_mask(field)) *
field_multiplier(field)); \
00b0c9b82663ac4 Al Viro 2017-12-14 133 } \
00b0c9b82663ac4 Al Viro 2017-12-14 134 static __always_inline __##type
type##_replace_bits(__##type old, \
00b0c9b82663ac4 Al Viro 2017-12-14 135 base val, base field) \
00b0c9b82663ac4 Al Viro 2017-12-14 136 { \
00b0c9b82663ac4 Al Viro 2017-12-14 137 return (old & ~to(field)) |
type##_encode_bits(val, field); \
00b0c9b82663ac4 Al Viro 2017-12-14 138 } \
00b0c9b82663ac4 Al Viro 2017-12-14 139 static __always_inline void
type##p_replace_bits(__##type *p, \
00b0c9b82663ac4 Al Viro 2017-12-14 140 base val, base field) \
00b0c9b82663ac4 Al Viro 2017-12-14 141 { \
00b0c9b82663ac4 Al Viro 2017-12-14 142 *p = (*p & ~to(field)) |
type##_encode_bits(val, field); \
00b0c9b82663ac4 Al Viro 2017-12-14 143 } \
00b0c9b82663ac4 Al Viro 2017-12-14 144 static __always_inline base
type##_get_bits(__##type v, base field) \
00b0c9b82663ac4 Al Viro 2017-12-14 145 { \
00b0c9b82663ac4 Al Viro 2017-12-14 146 return (from(v) &
field)/field_multiplier(field); \
00b0c9b82663ac4 Al Viro 2017-12-14 147 }
00b0c9b82663ac4 Al Viro 2017-12-14 148 #define __MAKE_OP(size) \
00b0c9b82663ac4 Al Viro 2017-12-14 149
____MAKE_OP(le##size,u##size,cpu_to_le##size,le##size##_to_cpu) \
00b0c9b82663ac4 Al Viro 2017-12-14 150
____MAKE_OP(be##size,u##size,cpu_to_be##size,be##size##_to_cpu) \
00b0c9b82663ac4 Al Viro 2017-12-14 151 ____MAKE_OP(u##size,u##size,,)
37a3862e1238262 Johannes Berg 2018-06-20 152 ____MAKE_OP(u8,u8,,)
00b0c9b82663ac4 Al Viro 2017-12-14 153 __MAKE_OP(16)
00b0c9b82663ac4 Al Viro 2017-12-14 154 __MAKE_OP(32)
00b0c9b82663ac4 Al Viro 2017-12-14 155 __MAKE_OP(64)
00b0c9b82663ac4 Al Viro 2017-12-14 156 #undef __MAKE_OP
00b0c9b82663ac4 Al Viro 2017-12-14 157 #undef ____MAKE_OP
00b0c9b82663ac4 Al Viro 2017-12-14 158
:::::: The code at line 119 was first introduced by commit
:::::: 00b0c9b82663ac42e5a09f58ce960f81f29d64ee Add primitives for manipulating bitfields
both in host- and fixed-endian.
:::::: TO: Al Viro <viro(a)zeniv.linux.org.uk>
:::::: CC: Al Viro <viro(a)zeniv.linux.org.uk>
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
8:08 p.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
The driver used this construct:
#define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
static inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
buff = u32_replace_bits(buff, val, mask);
writel(buff, priv->base + offset);
}
...
keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
KMB_PWM_LEADIN_OFFSET(pwm->hwpwm));
With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
triggers:
In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-keembay.c:16:
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’ declared
with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’ declared
with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
The compiler doesn't seem to be able to notice that with field being
0x3ffffff the expression
if ((field | (field - 1)) & ((field | (field - 1)) + 1))
__bad_mask();
can be optimized away.
So use __always_inline and document the problem in a comment to fix
this.
Reported-by: kernel test robot <lkp(a)intel.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
---
Hello,
I'm not sure this is the right fix. Maybe the bitfield stuff can be
changed somehow to make this problem go away, too?
Best regards
Uwe
drivers/pwm/pwm-keembay.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/pwm/pwm-keembay.c b/drivers/pwm/pwm-keembay.c
index 2b6dd070daa4..cdfdef66ff8e 100644
--- a/drivers/pwm/pwm-keembay.c
+++ b/drivers/pwm/pwm-keembay.c
@@ -63,7 +63,12 @@ static int keembay_clk_enable(struct device *dev, struct clk *clk)
return devm_add_action_or_reset(dev, keembay_clk_unprepare, clk);
}
-static inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32 mask,
+/*
+ * With gcc 10, CONFIG_CC_OPTIMIZE_FOR_SIZE and only "inline" instead of
+ * "__always_inline" this fails to compile because the compiler doesn't
notice
+ * for all valid masks (e.g. KMB_PWM_LEADIN_MASK) that they are ok.
+ */
+static __always_inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
--
2.28.0
4:29 a.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
Hi Uwe,
From: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
Sent: Monday, 16 November, 2020 5:08 PM
Subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
The driver used this construct:
#define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
static inline void keembay_pwm_update_bits(struct keembay_pwm
*priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
buff = u32_replace_bits(buff, val, mask);
writel(buff, priv->base + offset);
}
...
keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
KMB_PWM_LEADIN_OFFSET(pwm-
>hwpwm));
With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
triggers:
In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-
keembay.c:16:
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
‘__bad_mask’ declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
‘__bad_mask’ declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
The compiler doesn't seem to be able to notice that with field being
0x3ffffff the expression
if ((field | (field - 1)) & ((field | (field - 1)) + 1))
__bad_mask();
can be optimized away.
So use __always_inline and document the problem in a comment to fix
this.
Reported-by: kernel test robot <lkp(a)intel.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
Thank you for spending time in resolving this build failure.
I shall prepare and share the next version of patch with your approach.
---
Hello,
I'm not sure this is the right fix. Maybe the bitfield stuff can be
changed somehow to make this problem go away, too?
Best regards
Uwe
drivers/pwm/pwm-keembay.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/pwm/pwm-keembay.c b/drivers/pwm/pwm-keembay.c
index 2b6dd070daa4..cdfdef66ff8e 100644
--- a/drivers/pwm/pwm-keembay.c
+++ b/drivers/pwm/pwm-keembay.c
@@ -63,7 +63,12 @@ static int keembay_clk_enable(struct device *dev,
struct clk *clk)
return devm_add_action_or_reset(dev, keembay_clk_unprepare, clk);
}
-static inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32
mask,
+/*
+ * With gcc 10, CONFIG_CC_OPTIMIZE_FOR_SIZE and only "inline" instead of
+ * "__always_inline" this fails to compile because the compiler doesn't
notice
+ * for all valid masks (e.g. KMB_PWM_LEADIN_MASK) that they are ok.
+ */
+static __always_inline void keembay_pwm_update_bits(struct
keembay_pwm *priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
--
2.28.0
Thanks,
Vijay
8:48 p.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
On Tue, Nov 17, 2020 at 05:29:01PM +0000, Ayyathurai, Vijayakannan wrote:
Hi Uwe,
> From: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
> Sent: Monday, 16 November, 2020 5:08 PM
> Subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
>
> The driver used this construct:
>
> #define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
>
> static inline void keembay_pwm_update_bits(struct keembay_pwm
> *priv, u32 mask,
> u32 val, u32 offset)
> {
> u32 buff = readl(priv->base + offset);
>
> buff = u32_replace_bits(buff, val, mask);
> writel(buff, priv->base + offset);
> }
>
> ...
> keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
> KMB_PWM_LEADIN_OFFSET(pwm-
> >hwpwm));
>
> With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
> triggers:
>
> In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-
> keembay.c:16:
> In function ‘field_multiplier’,
> inlined from ‘keembay_pwm_update_bits’ at
> /home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
> /home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
> ‘__bad_mask’ declared with attribute error: bad bitfield mask
> 119 | __bad_mask();
> | ^~~~~~~~~~~~
> In function ‘field_multiplier’,
> inlined from ‘keembay_pwm_update_bits’ at
> /home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
> /home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
> ‘__bad_mask’ declared with attribute error: bad bitfield mask
> 119 | __bad_mask();
> | ^~~~~~~~~~~~
>
> The compiler doesn't seem to be able to notice that with field being
> 0x3ffffff the expression
>
> if ((field | (field - 1)) & ((field | (field - 1)) + 1))
> __bad_mask();
>
> can be optimized away.
>
> So use __always_inline and document the problem in a comment to fix
> this.
>
> Reported-by: kernel test robot <lkp(a)intel.com>
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
Thank you for spending time in resolving this build failure.
I shall prepare and share the next version of patch with your approach.
I don't understand this last sentence. IMHO there is currently nothing
you have to do for this problem. You can send an Ack however if you want
to.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
9:06 p.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
[Cc: += linux-pwm which I forgot for the initial submission]
Hello,
On Mon, Nov 16, 2020 at 10:08:04AM +0100, Uwe Kleine-König wrote:
The driver used this construct:
#define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
static inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
buff = u32_replace_bits(buff, val, mask);
writel(buff, priv->base + offset);
}
...
keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
KMB_PWM_LEADIN_OFFSET(pwm->hwpwm));
With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
triggers:
In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-keembay.c:16:
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’
declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’
declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
The compiler doesn't seem to be able to notice that with field being
0x3ffffff the expression
if ((field | (field - 1)) & ((field | (field - 1)) + 1))
__bad_mask();
can be optimized away.
So use __always_inline and document the problem in a comment to fix
this.
Reported-by: kernel test robot <lkp(a)intel.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
---
Hello,
I'm not sure this is the right fix. Maybe the bitfield stuff can be
changed somehow to make this problem go away, too?
Note, this patch
Fixes: cdbea243f419 ("pwm: Add PWM driver for Intel Keem Bay")
so this isn't critical for v5.10.
@thierry: If this is ok for you and Vijayakannan, you can squash this
into the original commit.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
4:41 a.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
Hi Thierry,
From: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
Subject: Re: [PATCH RFC] pwm: keembay: Fix build failure with -Os
[Cc: += linux-pwm which I forgot for the initial submission]
Hello,
On Mon, Nov 16, 2020 at 10:08:04AM +0100, Uwe Kleine-König wrote:
> The driver used this construct:
>
> #define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
>
> static inline void keembay_pwm_update_bits(struct keembay_pwm
*priv, u32 mask,
> u32 val, u32 offset)
> {
> u32 buff = readl(priv->base + offset);
>
> buff = u32_replace_bits(buff, val, mask);
> writel(buff, priv->base + offset);
> }
>
> ...
> keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
> KMB_PWM_LEADIN_OFFSET(pwm-
>hwpwm));
>
> With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
> triggers:
>
> In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-
keembay.c:16:
> In function ‘field_multiplier’,
> inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
> /home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
‘__bad_mask’ declared with attribute error: bad bitfield mask
> 119 | __bad_mask();
> | ^~~~~~~~~~~~
> In function ‘field_multiplier’,
> inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
> /home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to
‘__bad_mask’ declared with attribute error: bad bitfield mask
> 119 | __bad_mask();
> | ^~~~~~~~~~~~
>
> The compiler doesn't seem to be able to notice that with field being
> 0x3ffffff the expression
>
> if ((field | (field - 1)) & ((field | (field - 1)) + 1))
> __bad_mask();
>
> can be optimized away.
>
> So use __always_inline and document the problem in a comment to fix
> this.
>
> Reported-by: kernel test robot <lkp(a)intel.com>
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
> ---
> Hello,
>
> I'm not sure this is the right fix. Maybe the bitfield stuff can be
> changed somehow to make this problem go away, too?
Note, this patch
Fixes: cdbea243f419 ("pwm: Add PWM driver for Intel Keem Bay")
so this isn't critical for v5.10.
@thierry: If this is ok for you and Vijayakannan, you can squash this
into the original commit.
I am ok with Uwe approach.
I have compiled the change and tested in Keembay board as well.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
Thanks,
Vijay
5 a.m.
New subject: [PATCH RFC] pwm: keembay: Fix build failure with -Os
On Mon, Nov 16, 2020 at 10:08:04AM +0100, Uwe Kleine-König wrote:
The driver used this construct:
#define KMB_PWM_LEADIN_MASK GENMASK(30, 0)
static inline void keembay_pwm_update_bits(struct keembay_pwm *priv, u32 mask,
u32 val, u32 offset)
{
u32 buff = readl(priv->base + offset);
buff = u32_replace_bits(buff, val, mask);
writel(buff, priv->base + offset);
}
...
keembay_pwm_update_bits(priv, KMB_PWM_LEADIN_MASK, 0,
KMB_PWM_LEADIN_OFFSET(pwm->hwpwm));
With CONFIG_CC_OPTIMIZE_FOR_SIZE the compiler (here: gcc 10.2.0) this
triggers:
In file included from /home/uwe/gsrc/linux/drivers/pwm/pwm-keembay.c:16:
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:124:17:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’
declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
In function ‘field_multiplier’,
inlined from ‘keembay_pwm_update_bits’ at
/home/uwe/gsrc/linux/include/linux/bitfield.h:154:1:
/home/uwe/gsrc/linux/include/linux/bitfield.h:119:3: error: call to ‘__bad_mask’
declared with attribute error: bad bitfield mask
119 | __bad_mask();
| ^~~~~~~~~~~~
The compiler doesn't seem to be able to notice that with field being
0x3ffffff the expression
if ((field | (field - 1)) & ((field | (field - 1)) + 1))
__bad_mask();
can be optimized away.
So use __always_inline and document the problem in a comment to fix
this.
Reported-by: kernel test robot <lkp(a)intel.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig(a)pengutronix.de>
---
Hello,
I'm not sure this is the right fix. Maybe the bitfield stuff can be
changed somehow to make this problem go away, too?
Best regards
Uwe
drivers/pwm/pwm-keembay.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
Applied, thanks.
Thierry
671
days inactive
674
days old
7 comments
4 participants
participants (4)
-
Ayyathurai, Vijayakannan -
kernel test robot -
Thierry Reding -
Uwe Kleine-König