Merge branch 'topic/midi20' into for-next

This is a (largish) patch set for adding the support of MIDI 2.0
functionality, mainly targeted for USB devices.  MIDI 2.0 is a
complete overhaul of the 40-years old MIDI 1.0.  Unlike MIDI 1.0 byte
stream, MIDI 2.0 uses packets in 32bit words for Universal MIDI Packet
(UMP) protocol.  It supports both MIDI 1.0 commands for compatibility
and the extended MIDI 2.0 commands for higher resolutions and more
functions.

For supporting the UMP, the patch set extends the existing ALSA
rawmidi and sequencer interfaces, and adds the USB MIDI 2.0 support to
the standard USB-audio driver.

The rawmidi for UMP has a different device name (/dev/snd/umpC*D*) and
it reads/writes UMP packet data in 32bit CPU-native endianness.  For
the old MIDI 1.0 applications, the legacy rawmidi interface is
provided, too.

As default, USB-audio driver will take the alternate setting for MIDI
2.0 interface, and the compatibility with MIDI 1.0 is provided via the
rawmidi common layer.  However, user may let the driver falling back
to the old MIDI 1.0 interface by a module option, too.

A UMP-capable rawmidi device can create the corresponding ALSA
sequencer client(s) to support the UMP Endpoint and UMP Group
connections.  As a nature of ALSA sequencer, arbitrary connections
between clients/ports are allowed, and the ALSA sequencer core
performs the automatic conversions for the connections between a new
UMP sequencer client and a legacy MIDI 1.0 sequencer client.  It
allows the existing application to use MIDI 2.0 devices without
changes.

The MIDI-CI, which is another major extension in MIDI 2.0, isn't
covered by this patch set.  It would be implemented rather in
user-space.

Roughly speaking, the first half of this patch set is for extending
the rawmidi and USB-audio, and the second half is for extending the
ALSA sequencer interface.

The patch set is based on 6.4-rc2 kernel, but all patches can be
cleanly applicable on 6.2 and 6.3 kernels, too (while 6.1 and older
kernels would need minor adjustment for uapi header changes).

The updates for alsa-lib and alsa-utils will follow shortly later.

The author thanks members of MIDI Association OS/API Working Group,
especially Andrew Mee, for great helps for the initial design and
debugging / testing the drivers.

Link: https://lore.kernel.org/r/20230523075358.9672-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>

1 files changed, 4 insertions, 4 deletions

diff --git a/drivers/video/fbdev/hpfb.c b/drivers/video/fbdev/hpfb.c
index cdd44e5deafe..77fbff47b1a8 100644
--- a/drivers/video/fbdev/hpfb.c
+++ b/drivers/video/fbdev/hpfb.c
@@ -92,7 +92,7 @@ static int hpfb_setcolreg(unsigned regno, unsigned red, unsigned green,
 
 	if (regno >= info->cmap.len)
 		return 1;
-	
+
 	while (in_be16(fb_regs + 0x6002) & 0x4) udelay(1);
 
 	out_be16(fb_regs + 0x60ba, 0xff);
@@ -143,7 +143,7 @@ static void topcat_blit(int x0, int y0, int x1, int y1, int w, int h, int rr)
 	out_8(fb_regs + WMOVE, fb_bitmask);
 }
 
-static void hpfb_copyarea(struct fb_info *info, const struct fb_copyarea *area) 
+static void hpfb_copyarea(struct fb_info *info, const struct fb_copyarea *area)
 {
 	topcat_blit(area->sx, area->sy, area->dx, area->dy, area->width, area->height, RR_COPY);
 }
@@ -315,7 +315,7 @@ unmap_screen_base:
 	return ret;
 }
 
-/* 
+/*
  * Check that the secondary ID indicates that we have some hope of working with this
  * framebuffer.  The catseye boards are pretty much like topcats and we can muddle through.
  */
@@ -323,7 +323,7 @@ unmap_screen_base:
 #define topcat_sid_ok(x)  (((x) == DIO_ID2_LRCATSEYE) || ((x) == DIO_ID2_HRCCATSEYE)    \
 			   || ((x) == DIO_ID2_HRMCATSEYE) || ((x) == DIO_ID2_TOPCAT))
 
-/* 
+/*
  * Initialise the framebuffer
  */
 static int hpfb_dio_probe(struct dio_dev *d, const struct dio_device_id *ent)