* [PATCH 0/2] ALSA: seq: ump: Convert MIDI 2.0 RPN and NRPN for MIDI 1.0 clients
@ 2026-11-22 10:36 HyeongJun An
2026-11-22 10:36 ` [PATCH 1/2] ALSA: seq: ump: Don't drop " HyeongJun An
2026-11-22 10:36 ` [PATCH 2/2] selftests/alsa: Test the sequencer UMP MIDI 2.0 RPN conversion HyeongJun An
0 siblings, 2 replies; 3+ messages in thread
From: HyeongJun An @ 2026-11-22 10:36 UTC (permalink / raw)
To: Takashi Iwai
Cc: Jaroslav Kysela, Mark Brown, Shuah Khan, linux-sound,
linux-kselftest, linux-kernel, HyeongJun An
A UMP MIDI 1.0 client gets nothing when a UMP MIDI 2.0 client sends an
RPN or NRPN. Patch 1 converts them to CCs the way the core converter
does, and patch 2 adds a test.
A related question. In the other direction, CC 101, 100, 6 and 38 from
a UMP MIDI 1.0 client reach a MIDI 2.0 client as four MIDI 2.0 CC
messages. The same CCs from a legacy client are combined into one RPN
message by cc_ev_to_ump_midi2(). Should cvt_ump_midi1_to_midi2() do
that too? I can send it if you want it.
Tested in QEMU (x86_64) on sound/for-next 1afd6b7547b2. seq-ump-test
passes 28 of 29 without patch 1, where the new test gets no packets,
and 29 of 29 with it.
Prepared with Claude Code (Claude Opus 5.5).
HyeongJun An (2):
ALSA: seq: ump: Don't drop MIDI 2.0 RPN and NRPN for MIDI 1.0 clients
selftests/alsa: Test the sequencer UMP MIDI 2.0 RPN conversion
sound/core/seq/seq_ump_convert.c | 36 +++++++++++++++++++++
tools/testing/selftests/alsa/seq-ump-test.c | 20 ++++++++++++
2 files changed, 56 insertions(+)
--
2.43.0
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 1/2] ALSA: seq: ump: Don't drop MIDI 2.0 RPN and NRPN for MIDI 1.0 clients
2026-11-22 10:36 [PATCH 0/2] ALSA: seq: ump: Convert MIDI 2.0 RPN and NRPN for MIDI 1.0 clients HyeongJun An
@ 2026-11-22 10:36 ` HyeongJun An
2026-11-22 10:36 ` [PATCH 2/2] selftests/alsa: Test the sequencer UMP MIDI 2.0 RPN conversion HyeongJun An
1 sibling, 0 replies; 3+ messages in thread
From: HyeongJun An @ 2026-11-22 10:36 UTC (permalink / raw)
To: Takashi Iwai
Cc: Jaroslav Kysela, Mark Brown, Shuah Khan, linux-sound,
linux-kselftest, linux-kernel, HyeongJun An, stable
The cvt_ump_midi2_to_midi1() converts MIDI 2.0 channel voice messages
for a UMP MIDI 1.0 client. It has no case for the RPN and NRPN
messages, so they fall into the default case and are dropped. A pitch
bend range or a tuning set by a MIDI 2.0 client never reaches a MIDI 1.0
client.
The core converter, cvt_ump_midi2_to_legacy(), turns the same messages
into CC 101 and 100 (99 and 98 for NRPN) followed by CC 6 and 38. The
program change case here already sends the bank select CCs ahead of the
program in the same way.
To fix this, deliver the four CCs that cvt_ump_midi2_to_legacy()
produces.
Fixes: e9e02819a98a ("ALSA: seq: Automatic conversion of UMP events")
Cc: stable@vger.kernel.org
Signed-off-by: HyeongJun An <sammiee5311@gmail.com>
---
sound/core/seq/seq_ump_convert.c | 36 ++++++++++++++++++++++++++++++++
1 file changed, 36 insertions(+)
diff --git a/sound/core/seq/seq_ump_convert.c b/sound/core/seq/seq_ump_convert.c
index b42a7ffbf31e..e76fdb54df17 100644
--- a/sound/core/seq/seq_ump_convert.c
+++ b/sound/core/seq/seq_ump_convert.c
@@ -432,6 +432,38 @@ static int cvt_ump_midi1_to_midi2(struct snd_seq_client *dest,
atomic, hop);
}
+/* send a MIDI 2.0 RPN/NRPN as the four MIDI 1.0 CCs that select and set it */
+static int cvt_ump_midi2_rpn_to_midi1(struct snd_seq_client *dest,
+ struct snd_seq_client_port *dest_port,
+ struct snd_seq_ump_event *ev_cvt,
+ const union snd_ump_midi2_msg *midi2,
+ int atomic, int hop)
+{
+ union snd_ump_midi1_msg *midi1 = (union snd_ump_midi1_msg *)ev_cvt->ump;
+ bool rpn = midi2->rpn.status == UMP_MSG_STATUS_RPN;
+ u16 v = downscale_32_to_14bit(midi2->rpn.data);
+ const unsigned char cc[4][2] = {
+ { rpn ? UMP_CC_RPN_MSB : UMP_CC_NRPN_MSB, midi2->rpn.bank },
+ { rpn ? UMP_CC_RPN_LSB : UMP_CC_NRPN_LSB, midi2->rpn.index },
+ { UMP_CC_DATA, v >> 7 },
+ { UMP_CC_DATA_LSB, v & 0x7f },
+ };
+ int i, err = 0;
+
+ midi1->cc.status = UMP_MSG_STATUS_CC;
+ for (i = 0; i < ARRAY_SIZE(cc); i++) {
+ midi1->cc.index = cc[i][0];
+ midi1->cc.data = cc[i][1];
+ err = __snd_seq_deliver_single_event(dest, dest_port,
+ (struct snd_seq_event *)ev_cvt,
+ atomic, hop);
+ if (err < 0)
+ return err;
+ }
+
+ return err;
+}
+
/* convert UMP packet from MIDI 2.0 to MIDI 1.0 and deliver it */
static int cvt_ump_midi2_to_midi1(struct snd_seq_client *dest,
struct snd_seq_client_port *dest_port,
@@ -499,6 +531,10 @@ static int cvt_ump_midi2_to_midi1(struct snd_seq_client *dest,
midi1->pb.data_msb = v >> 7;
midi1->pb.data_lsb = v & 0x7f;
break;
+ case UMP_MSG_STATUS_RPN:
+ case UMP_MSG_STATUS_NRPN:
+ return cvt_ump_midi2_rpn_to_midi1(dest, dest_port, &ev_cvt,
+ midi2, atomic, hop);
default:
return 0;
}
--
2.43.0
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 2/2] selftests/alsa: Test the sequencer UMP MIDI 2.0 RPN conversion
2026-11-22 10:36 [PATCH 0/2] ALSA: seq: ump: Convert MIDI 2.0 RPN and NRPN for MIDI 1.0 clients HyeongJun An
2026-11-22 10:36 ` [PATCH 1/2] ALSA: seq: ump: Don't drop " HyeongJun An
@ 2026-11-22 10:36 ` HyeongJun An
1 sibling, 0 replies; 3+ messages in thread
From: HyeongJun An @ 2026-11-22 10:36 UTC (permalink / raw)
To: Takashi Iwai
Cc: Jaroslav Kysela, Mark Brown, Shuah Khan, linux-sound,
linux-kselftest, linux-kernel, HyeongJun An
Check that an RPN and an NRPN from a UMP MIDI 2.0 client reach a UMP
MIDI 1.0 client as the four CCs that select and set them. The NRPN data
has low bits set, so CC 38 is checked as well.
Signed-off-by: HyeongJun An <sammiee5311@gmail.com>
---
tools/testing/selftests/alsa/seq-ump-test.c | 20 ++++++++++++++++++++
1 file changed, 20 insertions(+)
diff --git a/tools/testing/selftests/alsa/seq-ump-test.c b/tools/testing/selftests/alsa/seq-ump-test.c
index 55bb7ea5b02f..7030d061dfae 100644
--- a/tools/testing/selftests/alsa/seq-ump-test.c
+++ b/tools/testing/selftests/alsa/seq-ump-test.c
@@ -620,4 +620,24 @@ TEST_F(seq_pair, ump_midi2_to_midi1_note)
{ 0x20913c01, 0 }, { 0x20813c00, 0 });
}
+TEST_F(seq_pair, ump_midi2_to_midi1_rpn)
+{
+ int skip = 0;
+ int err;
+
+ err = open_pair(&self->dst, SNDRV_SEQ_CLIENT_UMP_MIDI_1_0,
+ &self->src, SNDRV_SEQ_CLIENT_UMP_MIDI_2_0, &skip);
+ if (skip)
+ SKIP(return, "UMP sequencer client unavailable: %s",
+ strerror(-err));
+ ASSERT_EQ(0, err);
+
+ SEND_UMP(0x40210001, 0x80000000);
+ SEND_UMP(0x40310203, 0x12345678);
+ EXPECT_MT_PACKETS(2, { 0x20b16500, 0 }, { 0x20b16401, 0 },
+ { 0x20b10640, 0 }, { 0x20b12600, 0 },
+ { 0x20b16302, 0 }, { 0x20b16203, 0 },
+ { 0x20b10609, 0 }, { 0x20b1260d, 0 });
+}
+
TEST_HARNESS_MAIN
--
2.43.0
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-11-22 10:36 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-22 10:36 [PATCH 0/2] ALSA: seq: ump: Convert MIDI 2.0 RPN and NRPN for MIDI 1.0 clients HyeongJun An
2026-11-22 10:36 ` [PATCH 1/2] ALSA: seq: ump: Don't drop " HyeongJun An
2026-11-22 10:36 ` [PATCH 2/2] selftests/alsa: Test the sequencer UMP MIDI 2.0 RPN conversion HyeongJun An
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®