search

notro / fbtft

Linux Framebuffer drivers for small TFT LCD display modules. Development has moved to https://git.kernel.org/cgit/linux/kernel/git/gregkh/staging.git/tree/drivers/staging/fbtft?h=staging-testing
1.86k stars 494 forks source link

Correct padding for 9 bit emulation #478

Closed StefanBruens closed 3 years ago

StefanBruens commented 7 years ago

fbtft_write_spi_emulate_9 pads write up to a multiple of 4 words, although the lowest common multiple of 8 (buswidth) and 9 (device wordsize) is 72 bits, i.e. 8 device words.

Also, the padding/NOP commands may be different from 0x00, e.g.: SSD1325: 0x00 undefined, NOP = 0xE3 SSD1306: 0x00 - Lower column start address=0, NOP = 0xE3 SSD1305: same as SSD1306 SSD1351: 0x00 undefined, NOP = 0xE3 or NOP = 0xD1

For SSD1306/1305, the 0x00 command is unproblematic as long as it followed by the correct address command, but this is not given in general.

notro commented 7 years ago

Have you seen displays with those controllers that use 9-bit?

The 9-bit emulation was written for MIPI DBI compatible controllers. fbtft isn't the best it could be because I had only 2 displays to start with, which I built a library around. And this library has turned out to be quite inflexible.

tinydrm is the fbtft successor: https://github.com/notro/tinydrm/wiki

StefanBruens commented 7 years ago

The padding is correct in the new in-kernel tinydrm code, thanks for upstreaming this.

The 1306 is quite common for small 128x64 monochrome OLED displays, the SSD1351 starts to appear in 128x128 full color OLEDs.

For the 0xE3 NOP maybe an display dependent alternative to MIPI_DCS_NOP would do.

notro commented 7 years ago

I mean, have you seen displays where you can actually use the 9-bit mode? AFAIK all these Adafruit displays use D/C with SPI.

StefanBruens commented 7 years ago

I have one of these: http://xdec.de/msp430-oled-display-ssd1306-128x64/ which I have set to 3-Wire mode

There is also this one: http://www.dx.com/p/waveshare-0-96-oled-b-ssd1306-display-screen-module-deep-blue-340467 and: http://www.hotmcu.com/ssd1306-096-128%C3%9764-oled-display-%E2%80%93-i2cspi-interface-p-144.html

For the SSD1351, there is http://www.newhavendisplay.com/nhd15128128ugc3-p-5602.html, which has BS0/1 (interface select) on Pins 19/20.

notro commented 7 years ago

Any particular reason why you choose spi over i2c for the ssd1306? i2c is even less wires and there's a fbdev version: drivers/video/fbdev/ssd1307fb.c

StefanBruens commented 7 years ago

I use i2c for reading out sensors, and any transfer >= 50 bytes is killing bus access latency in my case. As far as I can see, the i2c fbdev driver always does full frame updates, i.e. 1kByte, thus the bus would be busy for 20/80 ms (@400kBit/s resp 100kBit/s).

github-actions[bot] commented 3 years ago

This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 7 days.