linux-kernel - Re: [PATCH v3 4/6] drm/tinydrm: add support for LEGO MINDSTORMS EV3 LCD

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] [day] [month] [year] [list]

Date:   Mon, 7 Aug 2017 11:03:36 -0500
From:   David Lechner <david@...hnology.com>
To:     Noralf Trønnes <noralf@...nnes.org>,
        dri-devel@...ts.freedesktop.org, devicetree@...r.kernel.org
Cc:     Daniel Vetter <daniel@...ll.ch>, David Airlie <airlied@...ux.ie>,
        Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Sekhar Nori <nsekhar@...com>,
        Kevin Hilman <khilman@...nel.org>,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/6] drm/tinydrm: add support for LEGO MINDSTORMS EV3
 LCD

On 08/05/2017 01:19 PM, Noralf Trønnes wrote:
> 
> Den 04.08.2017 00.33, skrev David Lechner:
>> +
>> +    buf = kmalloc(len, GFP_KERNEL);
>> +    if (!buf)
>> +        return;
>> +
>> +    tinydrm_xrgb8888_to_gray8(buf, vaddr, fb, clip);
>> +    src = buf;
>> +
>> +    for (y = clip->y1; y < clip->y2; y++) {
>> +        for (x = clip->x1; x < clip->x2; x += 3) {
>> +            val = *src++ & 0xc0;
>> +            if (val & 0xc0)
>> +                val |= 0x20;
>> +            val |= (*src++ & 0xc0) >> 3;
>> +            if (val & 0x18)
>> +                val |= 0x04;
>> +            val |= *src++ >> 6;
>> +            *dst++ = ~val;
> 
> I don't understand how this pixel packing matches the one described in
> the datasheet. Why do you flip the bits at the end?
> 

I a trying to be too clever. :-)

Here is the comment I will add to the next revision:

/*
  * The ST7586 controller has an unusual pixel format where 2bpp 
grayscale is
  * packed 3 pixels per byte with the first two pixels using 3 bits and 
the 3rd
  * pixel using only 2 bits.
  *
  * |  D7  |  D6  |  D5  ||  D1  |  D0  || 2bpp |
  * | (D4) | (D3) | (D2) ||      |      || GRAY |
  * +------+------+------++------+------++------+
  * |  1   |  1   |  1   ||  1   |  1   || 0  0 | black
  * |  1   |  0   |  0   ||  1   |  0   || 0  1 | dark gray
  * |  0   |  1   |  0   ||  0   |  1   || 1  0 | light gray
  * |  0   |  0   |  0   ||  0   |  0   || 1  1 | white
  */


As you can see, in the controller DRAM 1's are black and 0's are white, 
but in the kernel, it is the opposite. Also, if you look at the truth 
table, you can see that the extra 3rd bit has the pattern if D7 == 0 or 
D6 == 0 then D5 is zero.

I suppose it could be better to do this with a lookup table:

static const u8 st7586_lookup[] = { 0x7, 0x4, 0x2, 0x0 };

...

	for (y = clip->y1; y < clip->y2; y++) {
		for (x = clip->x1; x < clip->x2; x += 3) {
			val = st7586_lookup[*src++ >> 6] << 5;
			val |= st7586_lookup[*src++ >> 6] << 2;
			val |= st7586_lookup[*src++ >> 6] >> 1;
			*dst++ = val;
		}
	}