[U-Boot] [PATCH 0/3] dm: video: enhancements for Shell.efi

Split out of earlier efi_loader patchset, and updated with some comments, use of simple_strtol[1], and added kconfig option to enable ANSI escape sequence support.
[1] not a hard dependency, but to ensure correct behavior in all cases, depends on "lib: strto: fix incorrect handling of specified base"
Rob Clark (3): dm: video: Fix cache flushes dm: video: Add basic ANSI escape sequence support dm: video: Add color ANSI escape sequence support
drivers/video/Kconfig | 8 ++ drivers/video/vidconsole-uclass.c | 206 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 ++ include/video_console.h | 11 ++ 5 files changed, 233 insertions(+), 3 deletions(-)

Content can come to screen via putc() and we cannot always rely on updates ending with a puts(). This is the case with efi_console output to vidconsole. Fixes corruption with Shell.efi.
Signed-off-by: Rob Clark robdclark@gmail.com --- drivers/video/vidconsole-uclass.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/video/vidconsole-uclass.c b/drivers/video/vidconsole-uclass.c index b5afd72227..e081d5a0ee 100644 --- a/drivers/video/vidconsole-uclass.c +++ b/drivers/video/vidconsole-uclass.c @@ -163,6 +163,7 @@ static void vidconsole_putc(struct stdio_dev *sdev, const char ch) struct udevice *dev = sdev->priv;
vidconsole_put_char(dev, ch); + video_sync(dev->parent); }
static void vidconsole_puts(struct stdio_dev *sdev, const char *s) @@ -260,6 +261,8 @@ static int do_video_puts(cmd_tbl_t *cmdtp, int flag, int argc, for (s = argv[1]; *s; s++) vidconsole_put_char(dev, *s);
+ video_sync(dev->parent); + return 0; }

Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com --- drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig index 082cc4a528..add156cb70 100644 --- a/drivers/video/Kconfig +++ b/drivers/video/Kconfig @@ -44,6 +44,14 @@ config VIDEO_BPP32 this option, such displays will not be supported and console output will be empty.
+config VIDEO_ANSI + bool "Support ANSI escape sequences in video console" + depends on DM_VIDEO + default y if DM_VIDEO + help + Enable ANSI escape sequence decoding for a more fully functional + console. + config CONSOLE_NORMAL bool "Support a simple text console" depends on DM_VIDEO diff --git a/drivers/video/vidconsole-uclass.c b/drivers/video/vidconsole-uclass.c index e081d5a0ee..0a88cc0a42 100644 --- a/drivers/video/vidconsole-uclass.c +++ b/drivers/video/vidconsole-uclass.c @@ -9,6 +9,7 @@ */
#include <common.h> +#include <linux/ctype.h> #include <dm.h> #include <video.h> #include <video_console.h> @@ -107,12 +108,120 @@ static void vidconsole_newline(struct udevice *dev) video_sync(dev->parent); }
+static char *parsenum(char *s, int *num) +{ + char *end; + *num = simple_strtol(s, &end, 10); + return end; +} + +/* + * Process a character while accumulating an escape string. Chars are + * accumulated into escape_buf until the end of escape sequence is + * found, at which point the sequence is parsed and processed. + */ +static void vidconsole_escape_char(struct udevice *dev, char ch) +{ + struct vidconsole_priv *priv = dev_get_uclass_priv(dev); + + if (!IS_ENABLED(CONFIG_VIDEO_ANSI)) + goto error; + + /* Sanity checking for bogus ESC sequences: */ + if (priv->escape_len >= sizeof(priv->escape_buf)) + goto error; + if (priv->escape_len == 0 && ch != '[') + goto error; + + priv->escape_buf[priv->escape_len++] = ch; + + /* + * Escape sequences are terminated by a letter, so keep + * accumulating until we get one: + */ + if (!isalpha(ch)) + return; + + /* + * clear escape mode first, otherwise things will get highly + * surprising if you hit any debug prints that come back to + * this console. + */ + priv->escape = 0; + + switch (ch) { + case 'H': + case 'f': { + int row, col; + char *s = priv->escape_buf; + + /* + * Set cursor position: [%d;%df or [%d;%dH + */ + s++; /* [ */ + s = parsenum(s, &row); + s++; /* ; */ + s = parsenum(s, &col); + + priv->ycur = row * priv->y_charsize; + priv->xcur_frac = priv->xstart_frac + + VID_TO_POS(col * priv->x_charsize); + + break; + } + case 'J': { + int mode; + + /* + * Clear part/all screen: + * [J or [0J - clear screen from cursor down + * [1J - clear screen from cursor up + * [2J - clear entire screen + * + * TODO we really only handle entire-screen case, others + * probably require some additions to video-uclass (and + * are not really needed yet by efi_console) + */ + parsenum(priv->escape_buf + 1, &mode); + + if (mode == 2) { + video_clear(dev->parent); + video_sync(dev->parent); + priv->ycur = 0; + priv->xcur_frac = priv->xstart_frac; + } else { + debug("unsupported clear mode: %d\n", mode); + } + break; + } + default: + debug("unrecognized escape sequence: %*s\n", + priv->escape_len, priv->escape_buf); + } + + return; + +error: + /* something went wrong, just revert to normal mode: */ + priv->escape = 0; + return; +} + int vidconsole_put_char(struct udevice *dev, char ch) { struct vidconsole_priv *priv = dev_get_uclass_priv(dev); int ret;
+ if (priv->escape) { + vidconsole_escape_char(dev, ch); + return 0; + } + switch (ch) { + case '\x1b': + priv->escape_len = 0; + priv->escape = 1; + break; case '\a': /* beep */ break; diff --git a/drivers/video/video-uclass.c b/drivers/video/video-uclass.c index 3036e3a1f2..0163039821 100644 --- a/drivers/video/video-uclass.c +++ b/drivers/video/video-uclass.c @@ -87,7 +87,7 @@ int video_reserve(ulong *addrp) return 0; }
-static int video_clear(struct udevice *dev) +void video_clear(struct udevice *dev) { struct video_priv *priv = dev_get_uclass_priv(dev);
@@ -100,8 +100,6 @@ static int video_clear(struct udevice *dev) } else { memset(priv->fb, priv->colour_bg, priv->fb_size); } - - return 0; }
/* Flush video activity to the caches */ diff --git a/include/video.h b/include/video.h index 5b4e78b182..61ff653121 100644 --- a/include/video.h +++ b/include/video.h @@ -115,6 +115,13 @@ struct video_ops { int video_reserve(ulong *addrp);
/** + * video_clear() - Clear a device's frame buffer to background color. + * + * @dev: Device to clear + */ +void video_clear(struct udevice *dev); + +/** * video_sync() - Sync a device's frame buffer with its hardware * * Some frame buffers are cached or have a secondary frame buffer. This diff --git a/include/video_console.h b/include/video_console.h index 26047934da..9dce234bd9 100644 --- a/include/video_console.h +++ b/include/video_console.h @@ -29,6 +29,9 @@ * @xsize_frac: Width of the display in fractional units * @xstart_frac: Left margin for the text console in fractional units * @last_ch: Last character written to the text console on this line + * @escape: TRUE if currently accumulating an ANSI escape sequence + * @escape_len: Length of accumulated escape sequence so far + * @escape_buf: Buffer to accumulate escape sequence */ struct vidconsole_priv { struct stdio_dev sdev; @@ -42,6 +45,14 @@ struct vidconsole_priv { int xsize_frac; int xstart_frac; int last_ch; + /* + * ANSI escape sequences are accumulated character by character, + * starting after the ESC char (0x1b) until the entire sequence + * is consumed at which point it is acted upon. + */ + int escape; + int escape_len; + char escape_buf[32]; };
/**

On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com
drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
I don't see the test though - is that in another patch?

On Sun, Sep 17, 2017 at 1:55 PM, Simon Glass sjg@chromium.org wrote:
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com
drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
I don't see the test though - is that in another patch?
well, at this point test is load/bootefi Shell.efi and does it look messed up on screen. If you have better ideas, let me know.
BR, -R

Hi Rob,
On 17 September 2017 at 13:26, Rob Clark robdclark@gmail.com wrote:
On Sun, Sep 17, 2017 at 1:55 PM, Simon Glass sjg@chromium.org wrote:
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com
drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
I don't see the test though - is that in another patch?
well, at this point test is load/bootefi Shell.efi and does it look messed up on screen. If you have better ideas, let me know.
It should be easy enough to update test/dm/video.o to use the new feature in a new test. This code has nothing to do with EFI really.
The tests are fairly slow in that they gzip the display to check that it is correctly, but they work.
Regards, Simon

On Sun, Sep 17, 2017 at 3:30 PM, Simon Glass sjg@chromium.org wrote:
Hi Rob,
On 17 September 2017 at 13:26, Rob Clark robdclark@gmail.com wrote:
On Sun, Sep 17, 2017 at 1:55 PM, Simon Glass sjg@chromium.org wrote:
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com
drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
I don't see the test though - is that in another patch?
well, at this point test is load/bootefi Shell.efi and does it look messed up on screen. If you have better ideas, let me know.
It should be easy enough to update test/dm/video.o to use the new feature in a new test. This code has nothing to do with EFI really.
The tests are fairly slow in that they gzip the display to check that it is correctly, but they work.
ok, I guess snapshotting fb and comparing to a reference is a way.. where do I look to figure out how to build/run these tests (and presumable update reference screenshots?)
I'd be inclined to add any test as a patch on top of the following patch to cover color escape sequences at the same time..
BR, -R

Hi Rob,
On 17 September 2017 at 13:39, Rob Clark robdclark@gmail.com wrote:
On Sun, Sep 17, 2017 at 3:30 PM, Simon Glass sjg@chromium.org wrote:
Hi Rob,
On 17 September 2017 at 13:26, Rob Clark robdclark@gmail.com wrote:
On Sun, Sep 17, 2017 at 1:55 PM, Simon Glass sjg@chromium.org wrote:
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Really just the subset that is needed by efi_console. Perhaps more will be added later, for example color support would be useful to implement efi_cout_set_attribute().
Signed-off-by: Rob Clark robdclark@gmail.com
drivers/video/Kconfig | 8 +++ drivers/video/vidconsole-uclass.c | 109 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 +++ include/video_console.h | 11 ++++ 5 files changed, 136 insertions(+), 3 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
I don't see the test though - is that in another patch?
well, at this point test is load/bootefi Shell.efi and does it look messed up on screen. If you have better ideas, let me know.
It should be easy enough to update test/dm/video.o to use the new feature in a new test. This code has nothing to do with EFI really.
The tests are fairly slow in that they gzip the display to check that it is correctly, but they work.
ok, I guess snapshotting fb and comparing to a reference is a way.. where do I look to figure out how to build/run these tests (and presumable update reference screenshots?)
make tests
will run the tests.
To run the video tests, something like:
./test/py/test.py -k video --build-dir sandbox/
BTW there are no real reference screenshots. Just run the test once to get the value, check that the display looks good (run sandbox with -l) and then update your test with that value.
I'd be inclined to add any test as a patch on top of the following patch to cover color escape sequences at the same time..
Yes that sounds good.
- Simon

Note that this doesn't differentiate (due to lack of information in video_priv) between different possible component orders for 32bpp. But the main user at this point is efi_loader, and GOP expects xBGR so any video drivers that this is incorrect for already have problems. (Also, conveniently, this matches what simple-framebuffer bindings expect for kernels that use the simple-framebuffer DT binding to take over the bootloader display.)
Signed-off-by: Rob Clark robdclark@gmail.com --- drivers/video/vidconsole-uclass.c | 94 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 94 insertions(+)
diff --git a/drivers/video/vidconsole-uclass.c b/drivers/video/vidconsole-uclass.c index 0a88cc0a42..820540b1bf 100644 --- a/drivers/video/vidconsole-uclass.c +++ b/drivers/video/vidconsole-uclass.c @@ -108,6 +108,41 @@ static void vidconsole_newline(struct udevice *dev) video_sync(dev->parent); }
+static const struct { + unsigned r; + unsigned g; + unsigned b; +} colors[] = { + { 0x00, 0x00, 0x00 }, /* black */ + { 0xff, 0x00, 0x00 }, /* red */ + { 0x00, 0xff, 0x00 }, /* green */ + { 0xff, 0xff, 0x00 }, /* yellow */ + { 0x00, 0x00, 0xff }, /* blue */ + { 0xff, 0x00, 0xff }, /* magenta */ + { 0x00, 0xff, 0xff }, /* cyan */ + { 0xff, 0xff, 0xff }, /* white */ +}; + +static void set_color(struct video_priv *priv, unsigned idx, unsigned *c) +{ + switch (priv->bpix) { + case VIDEO_BPP16: + *c = ((colors[idx].r >> 3) << 0) | + ((colors[idx].g >> 2) << 5) | + ((colors[idx].b >> 3) << 11); + break; + case VIDEO_BPP32: + *c = 0xff000000 | + (colors[idx].r << 0) | + (colors[idx].g << 8) | + (colors[idx].b << 16); + break; + default: + /* unsupported, leave current color in place */ + break; + } +} + static char *parsenum(char *s, int *num) { char *end; @@ -194,6 +229,65 @@ static void vidconsole_escape_char(struct udevice *dev, char ch) } break; } + case 'm': { + struct video_priv *vid_priv = dev_get_uclass_priv(dev->parent); + char *s = priv->escape_buf; + char *end = &priv->escape_buf[priv->escape_len]; + + /* + * Set graphics mode: [%d;...;%dm + * + * Currently only supports the color attributes: + * + * Foreground Colors: + * + * 30 Black + * 31 Red + * 32 Green + * 33 Yellow + * 34 Blue + * 35 Magenta + * 36 Cyan + * 37 White + * + * Background Colors: + * + * 40 Black + * 41 Red + * 42 Green + * 43 Yellow + * 44 Blue + * 45 Magenta + * 46 Cyan + * 47 White + */ + + s++; /* [ */ + while (s < end) { + int val; + + s = parsenum(s, &val); + s++; + + switch (val) { + case 30 ... 37: + /* fg color */ + set_color(vid_priv, val - 30, + (unsigned *)&vid_priv->colour_fg); + break; + case 40 ... 47: + /* bg color */ + set_color(vid_priv, val - 40, + (unsigned *)&vid_priv->colour_bg); + break; + default: + /* unknown/unsupported */ + break; + } + } + + break; + } default: debug("unrecognized escape sequence: %*s\n", priv->escape_len, priv->escape_buf);

Hi Rob,
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Split out of earlier efi_loader patchset, and updated with some comments, use of simple_strtol[1], and added kconfig option to enable ANSI escape sequence support.
[1] not a hard dependency, but to ensure correct behavior in all cases, depends on "lib: strto: fix incorrect handling of specified base"
Rob Clark (3): dm: video: Fix cache flushes dm: video: Add basic ANSI escape sequence support dm: video: Add color ANSI escape sequence support
drivers/video/Kconfig | 8 ++ drivers/video/vidconsole-uclass.c | 206 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 ++ include/video_console.h | 11 ++ 5 files changed, 233 insertions(+), 3 deletions(-)
I'd like to apply this series but am waiting for the test update. Did you get my email about this?
Regards, Simon

On Thu, Sep 28, 2017 at 8:17 AM, Simon Glass sjg@chromium.org wrote:
Hi Rob,
On 13 September 2017 at 16:12, Rob Clark robdclark@gmail.com wrote:
Split out of earlier efi_loader patchset, and updated with some comments, use of simple_strtol[1], and added kconfig option to enable ANSI escape sequence support.
[1] not a hard dependency, but to ensure correct behavior in all cases, depends on "lib: strto: fix incorrect handling of specified base"
Rob Clark (3): dm: video: Fix cache flushes dm: video: Add basic ANSI escape sequence support dm: video: Add color ANSI escape sequence support
drivers/video/Kconfig | 8 ++ drivers/video/vidconsole-uclass.c | 206 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 ++ include/video_console.h | 11 ++ 5 files changed, 233 insertions(+), 3 deletions(-)
I'd like to apply this series but am waiting for the test update. Did you get my email about this?
yes, I guess you missed: https://patchwork.ozlabs.org/project/uboot/list/?series=5009
BR, -R

Hi Rob,
On Wed, 13 Sep 2017 18:12:19 -0400 Rob Clark robdclark@gmail.com wrote:
Split out of earlier efi_loader patchset, and updated with some comments, use of simple_strtol[1], and added kconfig option to enable ANSI escape sequence support.
[1] not a hard dependency, but to ensure correct behavior in all cases, depends on "lib: strto: fix incorrect handling of specified base"
Rob Clark (3): dm: video: Fix cache flushes dm: video: Add basic ANSI escape sequence support dm: video: Add color ANSI escape sequence support
drivers/video/Kconfig | 8 ++ drivers/video/vidconsole-uclass.c | 206 ++++++++++++++++++++++++++++++++++++++ drivers/video/video-uclass.c | 4 +- include/video.h | 7 ++ include/video_console.h | 11 ++ 5 files changed, 233 insertions(+), 3 deletions(-)
Series applied to u-boot-video/master, thanks!
-- Anatolij
participants (3)
-
Anatolij Gustschin
-
Rob Clark
-
Simon Glass