| Message ID | 20190830134827.12237-1-geert+renesas@glider.be (mailing list archive) |
|---|---|
| State | Mainlined |
| Commit | 672c29b91670ab34de288bb7015e9a12c63d007e |
| Delegated to: | Kieran Bingham |
| Headers | show |
| Series | media: i2c: Use div64_ul() for u64-by-unsigned-long divide | expand |
On Fri, Aug 30, 2019 at 03:48:27PM +0200, Geert Uytterhoeven wrote: > div_u64() does a 64-by-32 division, while the divisor max2175.xtal_freq > is unsigned long, thus 64-bit on 64-bit platforms. Hence the proper > function to call is div64_ul(). > > Note that this change does not have any functional impact, as the > crystal frequency must be much lower than the 32-bit limit anyway. > On 32-bit platforms, the generated code is the same. But at least on > arm64, this saves an AND-instruction to truncate xtal_freq to 32-bit. > > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Reviewed-by: Simon Horman <horms+renesas@verge.net.au> > --- > drivers/media/i2c/max2175.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/media/i2c/max2175.c b/drivers/media/i2c/max2175.c > index 19a3ceea3bc20f0f..506a30e69ced45c4 100644 > --- a/drivers/media/i2c/max2175.c > +++ b/drivers/media/i2c/max2175.c > @@ -591,8 +591,8 @@ static int max2175_set_lo_freq(struct max2175 *ctx, u32 lo_freq) > lo_freq *= lo_mult; > > int_desired = lo_freq / ctx->xtal_freq; > - frac_desired = div_u64((u64)(lo_freq % ctx->xtal_freq) << 20, > - ctx->xtal_freq); > + frac_desired = div64_ul((u64)(lo_freq % ctx->xtal_freq) << 20, > + ctx->xtal_freq); > > /* Check CSM is not busy */ > ret = max2175_poll_csm_ready(ctx); > -- > 2.17.1 >
diff --git a/drivers/media/i2c/max2175.c b/drivers/media/i2c/max2175.c index 19a3ceea3bc20f0f..506a30e69ced45c4 100644 --- a/drivers/media/i2c/max2175.c +++ b/drivers/media/i2c/max2175.c @@ -591,8 +591,8 @@ static int max2175_set_lo_freq(struct max2175 *ctx, u32 lo_freq) lo_freq *= lo_mult; int_desired = lo_freq / ctx->xtal_freq; - frac_desired = div_u64((u64)(lo_freq % ctx->xtal_freq) << 20, - ctx->xtal_freq); + frac_desired = div64_ul((u64)(lo_freq % ctx->xtal_freq) << 20, + ctx->xtal_freq); /* Check CSM is not busy */ ret = max2175_poll_csm_ready(ctx);
div_u64() does a 64-by-32 division, while the divisor max2175.xtal_freq is unsigned long, thus 64-bit on 64-bit platforms. Hence the proper function to call is div64_ul(). Note that this change does not have any functional impact, as the crystal frequency must be much lower than the 32-bit limit anyway. On 32-bit platforms, the generated code is the same. But at least on arm64, this saves an AND-instruction to truncate xtal_freq to 32-bit. Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> --- drivers/media/i2c/max2175.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)