This is tricky: if we just make "mode" global, the whole timing may
fall apart, with the DMA locking up. So let's take small bisectable
steps to get there ...
- ubb-vga.c (struct mode), ubb-vga.h: moved mode entry declaration to
ubb-vga.h
- ubb-vga.c (mode_db): moved up
- tstimg.c (printc): added diagonal to '0' (zero) to distinguish it from
'O' (oh)
- tstimg.c (printc): added "nose" to '1' to better distinguish it from
'I'
- tstimg.c (printc): completed the alphabet
- tstimg.c (printc): added '.' character
- ubb-vga.c (setup_noirq, line): moved DMA channel command setting to
per-session setup
- ubb-vga.c (setup_noirq, line): moved DMA request type setting to
per-session setup
- ubb-vga.c (line): added commend that the target address needs to be
set for each transfer, even if it doesn't change
Note that DMA is very sensitive to the video timing. At the moment,
only 800x600 mode works without locking up the DMA controller.
- ubb-vga.c (DMA, setup_noirq, cleanup_noirq, line): replaced MMC PIO
with DMA
- ubb-vga.c (frame): pass physical addresses to "line"
- ubb-vga.c (frame): detect if the DMA controller is stuck
- ubb-vga.c (ben_setup): use REG_WINDOW instead of calculating the
window size
- ubb-vga.c (PAGE_SIZS): we no longer need it - removed
- ubb-vga.c (line): shift bus width setting in MSC_CMDAT by the field
position, not the bit position
- ubb-vga.c (line): moved comment detailing the search for the
adjustment value to the mode database, so that we don't clutter
this function
- regs4740.h (REG_PADDR): translate a register access to the
corresponding physical address
- regs4740.h (REG_WINDOW): added register window size
- regs4740.h (CGU, INT, TCU, GPIO, MSC): prefix register group macros
with an underscore, to avoid name collision with DMAC
- regs4740.h: added DMAC registers
- regs4740.h (MSC_NOP): should be MSC_NOB
- physmem.c (calloc_phys_vec): malloc the vector instead of sbrk'ing it
- ubb-vga.h (xlat_virt), physmem.c (xlat_one, xlat_virt): translate a
vector of virtual addresses to physical addresses
- physmem.c (align_sbrk): align the end of the data segment
- physmem.c (calloc_phys_vec): align the line pointer vector to a machine
word, just in case
- physmem.c (calloc_phys_vec): crudely align the line buffers within
pages, to avoid page-crossing within a line (yes, this does reduce
noise)
The indirection introduced with the non-contiguous frame buffer caused
significant horizontal distortions. We can compensate for them by
starting the timer before loading the line pointer and calling "line".
- ubb-vga.h (calloc_phys_vec), physmem.c: non-contiguous memory allocator
(for now, without really considering any mapping to physical memory)
- Makefile (OBJS): added physmem.o
- ubb-vga.h (grabfb), grabfb.c (grabfb): API change for non-contiguous
frame buffer
- ubb-vga.h (ppmimg), ppmimg.c (convert, ppmimg): API change for
non-contiguous frame buffer
- ubb-vga.h (tstimg), tstimg.c (pixel, color_bars, grill, grid, sides,
dot, line45, arc, printc, text, ctext, tstimg): API change for
non-contiguous frame buffer
- tstimg.c (tstimg): we no longer need to clear the frame buffer
- ubb-vga.c (frame, session, main): use a vector of line buffers instead
of a contiguous frame buffer
- ubb-vga2.c (XREs, YRES, res, res_db, setup, line, hdelay, frame,
session): replaced hard-coded display characteristics with a mode
database
- ubb-vga2.c (usage, main): new option -m to select the display mode
- ubb-vga2.c (session): allocate frame buffer dynamically
- ubb-vga2.c (session): since we no longer prefetch, we don't need to
allocate space beyond the end of the frame buffer
- ubb-vga2.c (line): resurrected "quick load" and extended it to fill the
FIFO completely
- regs4740.h: macros to access selected Jz4740 registers
- ubb-vga2.c: replaced all pointers to registers with access macros
- ubb-vga2.c (line_cycles): we now seem to be 0.1 us slower
- ubb-vga2.c (line): write the first word to the TX FIFO after starting
the command or it is lost
- ubb-vga2.c (line): fetch the first word while waiting for the back
porch, making sure we have plenty of time for cache effects
- ubb-vga2.c (line): we no longer need to prefetch the first word of
the next line
- ubb-vga.h (ccube_init, ccube_map), ccube.c: color mapper based on
proximity in color cube
- grabfb.c (pattern, grabfb), ppmimg.c (pattern, convert): use the color
cube mapper instead of inferios threshold-based mapping
- ubb-vga2.c (session): initialize the color cube
- ubb-vga.h (thres), grabfb.c (thres), ubb-vga2.c (usage, main): removed
the threshold along with the option (-l) to set it
- Makefile (OBJS): added ccube.o
- Makefile (OBJS): added ppm.o and ppmimg.o
- ppm.h (load_ppm), ppm.c: PPM file loader, adapted from
eda-tools/schhist/ppmdiff/ppmdiff.c
- ubb-vga.h (img_name, ppmimg), ppmimg.c: PPM image to frame buffer
converter
- ubb-vga2.c (usage, main): the threshold is now set with the option -l
- ubb-vga2.c (usage, main): if a second argument is given, treat it as
a PPM file
- ubb-vga2.c (usage): also documented option -t
- tstimg.c: test image generator
- Makefile (OBJS): added tstimg.o
- Makefile (LDFLAGS): tstimg.c needs libm
- ubb-vga2.c (session, main): pass image generator as a function pointer
- ubb-vga2.c (main): new option -t to select the test image generator
- ubb-vga2.c (session, grab): made the frame buffer "void *" and pass
frame buffer resolution
- ubb-vga.h: color encoding and items shared between compilation units
- ubb-vga2.c (thres, pattern, grab), grabfb.c: move to grabfb.c
- ubb-vga2.c (map): make global, for use by grabfb.c
- ubb-vga2.c (pick, pattern, tricolor, grid): removed obsolete test
image generators
- Makefile (OBJS, ubb-vga2): added grabfb.o
- Makefile (clean): remove only .o files
- Makefile (spotless): remove only executables
- ubb-vga2.c (frame, grab, session): keep all 480 lines in the frame
buffer instead of duplicating them on output
- ubb-vga2.c (bad): make "static"
- ubb-vga2.c (session, main): removed the "single" argument
- ubb-vga2.c (main): complain about timeouts only if any occurred
- ubb-vga.sch: use resistive divider instead of diodes
- ubb-vga.sch, ubb-vga.c (VSYNC): moved VSYNC from DAT3 to CLK to make
all MMC data lines available for pixel data
- ubb-vga.sch, ubb-vga.c (R): moved red from DAT2 to DAT3 for convenient
access to the 4th channel
- ubb-vga.sch: added luminance (Y) channel
- ubb-vga.sch: added pull-down to HSYNC/CMD, for later use with the MMC
controller
- ubb-vga.c (pdfuns, pdfunc, ben_setup, setup): make UBB signals GPIOs,
in case someone left them in a different configuration
- README: added compatibility of double mode
- ubb-vga.c (usage, main): new option -d to double the number of set/clear
pairs (making lines longer and the apparent pixel clock slower)
- ubb-vga.c (grab, session, main): have a complete set/clear pair for each
pixel in double mode
- ubb-vga.c (line_pairs, line, frame, tricolor, grid, session): line
length is now kept in a variable, instead of hard-coding its value
and the values derived from it
- ubb-vga.c (line_cycles, line, hdelay): the total line duration is now
kept in a variable, instead of hard-coding it all over the place
- ubb-vga.c: added more section titles
- ubb-vga.c (pick, pattern): grouped pixel-generating functions with
image generation
- ubb-vga.c (usage, main): properly parse command-line options ("getopt"
is a bit of overkill at the moment but will be needed later)
The history of the split prefetch code:
Once up on a time, I thought that it was important to stay very close
to the "official" pixel clock. We could do this when copying 32 bit
words from cache to the GPIO set/clear registers, but not when
shifting 8 bit values as we do now.
One problem with 32 bit values is that transferring that amount of
data exceeds the time available in the hsync pulse. Hence the idea
to split the prefetch between hsync pulse and horizontal front porch.
Note that we also split prefetches between the two VGA lines of each
QVGA line. This commit does not change that mechanism.
- ubb-vga.c (line, frame): removed code fragments for prefetch split
between hsync pulse and front porch
- ubb-vga.c (line, frame): remove code fragments for measuring time
relative to the rising edge of hsync (and letting the prefetch time
control the hsync length), not the falling edge
- nanonote-nxuart.patch (avrdude.conf): typo, CMD is signal 6, not 5 (by
David Reyes Samblas Martinez)
- series, nanonote-icsp.patch (avrdude.conf): NanoNote ICSP 8:10 cable,
by David Reyes Samblas Martinez
- series, low-addr.patch: record the lowest address touched (only for
ihex, all others default to 0)
- series, skip-low-addr.patch: skip untouched addresses in avr_read
and avr_write
