Hi Pekka,
On 4/27/22 04:55, Pekka Paalanen wrote:
On Tue, 26 Apr 2022 21:53:19 -0300
Igor Torrente <igormtorrente@xxxxxxxxx> wrote:
Hi Pekka,
On 4/21/22 07:58, Pekka Paalanen wrote:
On Mon, 4 Apr 2022 17:45:15 -0300
Igor Torrente <igormtorrente@xxxxxxxxx> wrote:
Adds this common format to vkms.
This commit also adds new helper macros to deal with fixed-point
arithmetic.
It was done to improve the precision of the conversion to ARGB16161616
since the "conversion ratio" is not an integer.
V3: Adapt the handlers to the new format introduced in patch 7 V3.
V5: Minor improvements
Signed-off-by: Igor Torrente <igormtorrente@xxxxxxxxx>
---
drivers/gpu/drm/vkms/vkms_formats.c | 70 +++++++++++++++++++++++++++
drivers/gpu/drm/vkms/vkms_plane.c | 6 ++-
drivers/gpu/drm/vkms/vkms_writeback.c | 3 +-
3 files changed, 76 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/vkms/vkms_formats.c b/drivers/gpu/drm/vkms/vkms_formats.c
index 8d913fa7dbde..4af8b295f31e 100644
--- a/drivers/gpu/drm/vkms/vkms_formats.c
+++ b/drivers/gpu/drm/vkms/vkms_formats.c
@@ -5,6 +5,23 @@
#include "vkms_formats.h"
+/* The following macros help doing fixed point arithmetic. */
+/*
+ * With Fixed-Point scale 15 we have 17 and 15 bits of integer and fractional
+ * parts respectively.
+ * | 0000 0000 0000 0000 0.000 0000 0000 0000 |
+ * 31 0
+ */
+#define FIXED_SCALE 15
I think this would usually be called a "shift" since it's used in
bit-shifts.
Ok, I will rename this.
+
+#define INT_TO_FIXED(a) ((a) << FIXED_SCALE)
+#define FIXED_MUL(a, b) ((s32)(((s64)(a) * (b)) >> FIXED_SCALE))
+#define FIXED_DIV(a, b) ((s32)(((s64)(a) << FIXED_SCALE) / (b)))
A truncating div, ok.
+/* This macro converts a fixed point number to int, and round half up it */
+#define FIXED_TO_INT_ROUND(a) (((a) + (1 << (FIXED_SCALE - 1))) >> FIXED_SCALE)
Yes.
+/* Convert divisor and dividend to Fixed-Point and performs the division */
+#define INT_TO_FIXED_DIV(a, b) (FIXED_DIV(INT_TO_FIXED(a), INT_TO_FIXED(b)))
Ok, this is obvious to read, even though it's the same as FIXED_DIV()
alone. Not sure the compiler would optimize that extra bit-shift away...
If one wanted to, it would be possible to write type-safe functions for
these so that fixed and integer could not be mixed up.
Ok, I will move to a function.
That's not all.
If you want it type-safe, then you need something like
struct vkms_fixed_point {
s32 value;
};
And use `struct vkms_fixed_point` (by value) everywhere where you pass
a fixed point value, and never as a plain s32 type. Then it will be
impossible to do incorrect arithmetic or conversions by accident on
fixed point values.
Is it worth it? I don't know, since it's limited into this one file.
A simple 'typedef s32 vkms_fixed_point' does not work, because it does
not prevent computing with vkms_fixed_point as if it was just a normal
s32. Using a struct prevents that.
ohhh. Got it!
I wonder if the kernel doesn't already have something like this
available in general...
After some time searching I found `include/drm/drm_fixed.h`[1].
It seems fine. There are minor things to consider:
1. It doesn't have a `FIXED_TO_INT_ROUND` equivalent.
2. We can use `fixed20_12` for rgb565 but We have to use s64 for YUV
formats or add a `sfixed20_12` with s32.
In terms of consistency, do you think worth using this "library" given
that we may need to use two distinct ways to represent the fixed point
soonish? Or it's better to implement `sfixed20_12`? Or just continue
with the
current macros?
[1] - https://elixir.bootlin.com/linux/latest/source/include/drm/drm_fixed.h
+ u16 r = FIXED_TO_INT_ROUND(FIXED_DIV(fp_r, fp_rb_ratio));
+ u16 g = FIXED_TO_INT_ROUND(FIXED_DIV(fp_g, fp_g_ratio));
+ u16 b = FIXED_TO_INT_ROUND(FIXED_DIV(fp_b, fp_rb_ratio));
+
+ *dst_pixels = cpu_to_le16(r << 11 | g << 5 | b);
Looks good.
You are using signed variables (int, s64, s32) when negative values
should never occur. It doesn't seem wrong, just unexpected.
I left the signal so I can reuse them in the YUV formats.
Good point.
The use of int in code vs. s32 in the macros is a bit inconsistent as
well.
Right. I think I will stick with s32 and s64 then.
...
diff --git a/drivers/gpu/drm/vkms/vkms_writeback.c b/drivers/gpu/drm/vkms/vkms_writeback.c
index cb63a5da9af1..98da7bee0f4b 100644
--- a/drivers/gpu/drm/vkms/vkms_writeback.c
+++ b/drivers/gpu/drm/vkms/vkms_writeback.c
@@ -16,7 +16,8 @@
static const u32 vkms_wb_formats[] = {
DRM_FORMAT_XRGB8888,
DRM_FORMAT_XRGB16161616,
- DRM_FORMAT_ARGB16161616
+ DRM_FORMAT_ARGB16161616,
+ DRM_FORMAT_RGB565
};
static const struct drm_connector_funcs vkms_wb_connector_funcs = {
I wonder, would it be possible to add a unit test to make sure that
get_plane_fmt_transform_function() or get_wb_fmt_transform_function()
does not return NULL for any of the listed formats, respectively?
Or is that too paranoid?
I'm not opposed to it. But I also don't think it needs to be in this
series of patches either.
A new todo maybe?
If it's a good thing, then it belongs in this series, because those
function getters are introduced in this series, opening the door for
the mistakes that the tests would prevent. I don't mean IGT tests but
kernel internal tests. I think there is a unit test framework?
Yeah, kernel have kunit and kselftest. Idk what are the differences
between them or how to use them, but I know they exist.
You really should get a kernel maintainer's opinion on these questions,
as I am not a kernel developer.
Ok.
Thanks,
pq
