[PATCH 0/3] ASoC: Clean-up W=1 build warnings - part3
This is a much smaller set of cleanups, all related to warnings thrown by the use of GENMASK() with an unsigned variable. I just made the warning go away but I wonder if there's a better fix in the definition of GENMASK() itself?
Pierre-Louis Bossart (3): ASoC: Intel: Skylake: remove comparison always false warning ASoC: meson: axg-pdm: remove comparison always false warning ASoC: meson: axg-spdifin: remove comparison always false warning
sound/soc/intel/skylake/skl-sst-dsp.c | 2 +- sound/soc/meson/axg-pdm.c | 2 +- sound/soc/meson/axg-spdifin.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-)
Fix W=1 warnings:
skl-sst-dsp.c: In function ‘skl_dsp_get_enabled_cores’: include/linux/bits.h:26:28: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
cast the core number to a long to avoid checking if an unsigned value is lower than zero.
Signed-off-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com --- sound/soc/intel/skylake/skl-sst-dsp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/intel/skylake/skl-sst-dsp.c b/sound/soc/intel/skylake/skl-sst-dsp.c index 225706d148d8..cb83b395f210 100644 --- a/sound/soc/intel/skylake/skl-sst-dsp.c +++ b/sound/soc/intel/skylake/skl-sst-dsp.c @@ -52,7 +52,7 @@ unsigned int skl_dsp_get_enabled_cores(struct sst_dsp *ctx) unsigned int core_mask, en_cores_mask; u32 val;
- core_mask = SKL_DSP_CORES_MASK(skl->cores.count); + core_mask = SKL_DSP_CORES_MASK((long)skl->cores.count);
val = sst_dsp_shim_read_unlocked(ctx, SKL_ADSP_REG_ADSPCS);
Fix W=1 warning.
sound/soc/meson/axg-pdm.c: In function ‘axg_pdm_set_channel_mask’: include/linux/bits.h:26:28: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
cast the channel number to an int to avoid checking if an unsigned value is lower than zero.
Signed-off-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com --- sound/soc/meson/axg-pdm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/meson/axg-pdm.c b/sound/soc/meson/axg-pdm.c index bfd37d49a73e..8fe5b2563619 100644 --- a/sound/soc/meson/axg-pdm.c +++ b/sound/soc/meson/axg-pdm.c @@ -206,7 +206,7 @@ static int axg_pdm_set_sample_pointer(struct axg_pdm *priv) static void axg_pdm_set_channel_mask(struct axg_pdm *priv, unsigned int channels) { - unsigned int mask = GENMASK(channels - 1, 0); + unsigned int mask = GENMASK((int)channels - 1, 0);
/* Put all channel in reset */ regmap_update_bits(priv->map, PDM_CTRL,
Fix W=1 warning:
sound/soc/meson/axg-spdifin.c: In function ‘axg_spdifin_write_mode_param’: include/linux/bits.h:26:28: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits]
Cast width to an int to avoid checking if an unsigned value is lower than zero.
Signed-off-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com --- sound/soc/meson/axg-spdifin.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/meson/axg-spdifin.c b/sound/soc/meson/axg-spdifin.c index d0d09f945b48..5fa411d4e985 100644 --- a/sound/soc/meson/axg-spdifin.c +++ b/sound/soc/meson/axg-spdifin.c @@ -154,7 +154,7 @@ static void axg_spdifin_write_mode_param(struct regmap *map, int mode, reg = offset * regmap_get_reg_stride(map) + base_reg; shift = width * (num_per_reg - 1 - rem);
- regmap_update_bits(map, reg, GENMASK(width - 1, 0) << shift, + regmap_update_bits(map, reg, GENMASK((int)width - 1, 0) << shift, val << shift); }
On Tue 07 Jul 2020 at 21:23, Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com wrote:
This is a much smaller set of cleanups, all related to warnings thrown by the use of GENMASK() with an unsigned variable. I just made the warning go away but I wonder if there's a better fix in the definition of GENMASK() itself?
Looking at the patch I was going to ask the same thing. It does not make much sense to me to force GENMASK arguments to be integer (instead of unsigned integer) to then check there are positive ...
Pierre-Louis Bossart (3): ASoC: Intel: Skylake: remove comparison always false warning ASoC: meson: axg-pdm: remove comparison always false warning ASoC: meson: axg-spdifin: remove comparison always false warning
sound/soc/intel/skylake/skl-sst-dsp.c | 2 +- sound/soc/meson/axg-pdm.c | 2 +- sound/soc/meson/axg-spdifin.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-)
On 7/7/20 2:39 PM, Jerome Brunet wrote:
On Tue 07 Jul 2020 at 21:23, Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com wrote:
This is a much smaller set of cleanups, all related to warnings thrown by the use of GENMASK() with an unsigned variable. I just made the warning go away but I wonder if there's a better fix in the definition of GENMASK() itself?
Looking at the patch I was going to ask the same thing. It does not make much sense to me to force GENMASK arguments to be integer (instead of unsigned integer) to then check there are positive ...
Agree, it's just that the following macro isn't exactly simple to change:
#define GENMASK_INPUT_CHECK(h, l) \ (BUILD_BUG_ON_ZERO(__builtin_choose_expr( \ __builtin_constant_p((l) > (h)), (l) > (h), 0)))
I couldn't find a means to avoid the comparison.
I just realized this is a fairly recent addition in 295bcca84916 ('linux/bits.h: add compile time sanity check of GENMASK inputs'), adding the author Rikard Falkeborn in CC:
include/linux/bits.h:26:28: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits] 26 | __builtin_constant_p((l) > (h)), (l) > (h), 0)))
On Tue, 07 Jul 2020, Pierre-Louis Bossart wrote:
On 7/7/20 2:39 PM, Jerome Brunet wrote:
On Tue 07 Jul 2020 at 21:23, Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com wrote:
This is a much smaller set of cleanups, all related to warnings thrown by the use of GENMASK() with an unsigned variable. I just made the warning go away but I wonder if there's a better fix in the definition of GENMASK() itself?
Looking at the patch I was going to ask the same thing. It does not make much sense to me to force GENMASK arguments to be integer (instead of unsigned integer) to then check there are positive ...
Agree, it's just that the following macro isn't exactly simple to change:
#define GENMASK_INPUT_CHECK(h, l) \ (BUILD_BUG_ON_ZERO(__builtin_choose_expr( \ __builtin_constant_p((l) > (h)), (l) > (h), 0)))
I couldn't find a means to avoid the comparison.
I just realized this is a fairly recent addition in 295bcca84916 ('linux/bits.h: add compile time sanity check of GENMASK inputs'), adding the author Rikard Falkeborn in CC:
include/linux/bits.h:26:28: warning: comparison of unsigned expression < 0 is always false [-Wtype-limits] 26 | __builtin_constant_p((l) > (h)), (l) > (h), 0)))
Linus recently complained about the type-limits warning, saying that it was invalid. He preferred the warning to be bumped from W=1 to W=2, although I haven't seen a patch doing this yet.
Rikard also tried to fix GENMASK directly; however, Linus did not approve of this either.
On Wed, Jul 08, 2020 at 07:31:00AM +0100, Lee Jones wrote:
Linus recently complained about the type-limits warning, saying that it was invalid. He preferred the warning to be bumped from W=1 to W=2, although I haven't seen a patch doing this yet.
Rikard also tried to fix GENMASK directly; however, Linus did not approve of this either.
Let's just leave this series for now then.
participants (4)
-
Jerome Brunet -
Lee Jones -
Mark Brown -
Pierre-Louis Bossart