Skip to contentSkip to conversation

Add ASoC support for AMD APUs [v2]

PATCH In review Alex Deucher 2 participants 11 messages Aug 20, 2015 – Aug 21, 2015

Subsystems (7)
lore
Share & tools
lore
#0 lore raw

cover Add ASoC support for AMD APUs [v2]

This patch set implements support for i2s audio and new AMD GPUs. The i2s codec is fed by a DMA engine on the GPU. To handle this we create an mfd cell which we can hang the i2s codec on. Because of this, this patch set covers two subsystems: drm and alsa. The drm patches add support for the ACP hw block which provides the DMA engine for the i2s codec. The alsa patches add the ASoC driver for the i2s codec. Since the alsa changes depend on the drm changes in this patch set as well as some other drm changes queued for 4.3, I'd like to take the alsa patches in via the drm tree.

V2 changes:
- Use the MFD subsystem rather than adding our own bus
- Squash all sub-feature patches together
- fix comments mentioned in previous review

There was a question about the code and the module licensing. IANAL, but this follows the same pattern that most if not all drm drivers use.

There was also a comment about the patch changelog. I think this is more of a subsystem preference kind of thing. Generally in the drm we include the patch changelog and versions in the commit messages so we have a history of the patch if we need to reference it in the future. If you have a strong preference for the audio patch, we can drop that.

Patch 1 adds the register headers for the ACP block which is a pretty big patch so I've excluded it from email. The entire patch set can be viewed here:
http://cgit.freedesktop.org/~agd5f/linux/log/?h=acp-upstream2

Thanks,

Alex

Chunming Zhou (1):
drm/amd: add ACP 2.x register headers

Maruthi Bayyavarapu (1):
drm/amd: add ACP driver support [v6]

Maruthi Srinivas Bayyavarapu (1):
ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

 drivers/gpu/drm/Kconfig                            |    2 +
 drivers/gpu/drm/amd/acp/Kconfig                    |   10 +
 drivers/gpu/drm/amd/acp/Makefile                   |    9 +
 drivers/gpu/drm/amd/acp/acp_hw.c                   | 1133 ++++++++++
 drivers/gpu/drm/amd/acp/acp_hw.h                   |   91 +
 drivers/gpu/drm/amd/acp/include/acp_gfx_if.h       |   49 +
 drivers/gpu/drm/amd/amdgpu/Makefile                |   13 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu.h                |    5 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c            |  208 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h            |   40 +
 drivers/gpu/drm/amd/amdgpu/vi.c                    |   12 +
 drivers/gpu/drm/amd/include/amd_shared.h           |    1 +
 .../gpu/drm/amd/include/asic_reg/acp/acp_2_1_d.h   |  437 ++++
 .../drm/amd/include/asic_reg/acp/acp_2_1_enum.h    | 1198 ++++++++++
 .../drm/amd/include/asic_reg/acp/acp_2_1_sh_mask.h | 1568 +++++++++++++
 .../gpu/drm/amd/include/asic_reg/acp/acp_2_2_d.h   |  609 ++++++
 .../drm/amd/include/asic_reg/acp/acp_2_2_enum.h    | 1068 +++++++++
 .../drm/amd/include/asic_reg/acp/acp_2_2_sh_mask.h | 2292 ++++++++++++++++++++
 .../gpu/drm/amd/include/asic_reg/acp/acp_2_3_d.h   |  582 +++++
 .../drm/amd/include/asic_reg/acp/acp_2_3_enum.h    | 1079 +++++++++
 .../drm/amd/include/asic_reg/acp/acp_2_3_sh_mask.h | 2192 +++++++++++++++++++
 include/linux/mfd/amd_acp.h                        |  211 ++
 sound/soc/Kconfig                                  |    1 +
 sound/soc/Makefile                                 |    1 +
 sound/soc/amd/Kconfig                              |    5 +
 sound/soc/amd/Makefile                             |    3 +
 sound/soc/amd/acp-pcm-dma.c                        |  676 ++++++
 27 files changed, 13494 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/acp/Kconfig
 create mode 100644 drivers/gpu/drm/amd/acp/Makefile
 create mode 100644 drivers/gpu/drm/amd/acp/acp_hw.c
 create mode 100644 drivers/gpu/drm/amd/acp/acp_hw.h
 create mode 100644 drivers/gpu/drm/amd/acp/include/acp_gfx_if.h
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_1_d.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_1_enum.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_1_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_2_d.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_2_enum.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_3_d.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_3_enum.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/acp/acp_2_3_sh_mask.h
 create mode 100644 include/linux/mfd/amd_acp.h
 create mode 100644 sound/soc/amd/Kconfig
 create mode 100644 sound/soc/amd/Makefile
 create mode 100644 sound/soc/amd/acp-pcm-dma.c
Signature · 6 lines
1.8.3.1

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#1 lore raw

PATCH 2/3 drm/amd: add ACP driver support [v6]

From: Maruthi Bayyavarapu <[email protected]>

This adds the ACP (Audio CoProcessor) IP driver and wires it up to the amdgpu driver. The ACP block provides the DMA engine and bus for the i2s codec which is supported by an alsa driver. This is required for audio on APUs that utilize an i2s codec.

v2: integrate i2s/az check patch
v3: s/amd_acp/amdgpu_acp/
v4: update copyright notice
v5: squash multiple patches, convert to mfd
v6: squash in dynamic cell allocation from Maruthi

---
 drivers/gpu/drm/Kconfig                      |    2 +
 drivers/gpu/drm/amd/acp/Kconfig              |   10 +
 drivers/gpu/drm/amd/acp/Makefile             |    9 +
 drivers/gpu/drm/amd/acp/acp_hw.c             | 1133 ++++++++++++++++++++++++++
 drivers/gpu/drm/amd/acp/acp_hw.h             |   91 +++
 drivers/gpu/drm/amd/acp/include/acp_gfx_if.h |   49 ++
 drivers/gpu/drm/amd/amdgpu/Makefile          |   13 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu.h          |    5 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c      |  208 +++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h      |   40 +
 drivers/gpu/drm/amd/amdgpu/vi.c              |   12 +
 drivers/gpu/drm/amd/include/amd_shared.h     |    1 +
 include/linux/mfd/amd_acp.h                  |  211 +++++
 13 files changed, 1783 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/acp/Kconfig
 create mode 100644 drivers/gpu/drm/amd/acp/Makefile
 create mode 100644 drivers/gpu/drm/amd/acp/acp_hw.c
 create mode 100644 drivers/gpu/drm/amd/acp/acp_hw.h
 create mode 100644 drivers/gpu/drm/amd/acp/include/acp_gfx_if.h
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h
 create mode 100644 include/linux/mfd/amd_acp.h
drivers/gpu/drm/Kconfig +2 -0
index df99b01..2e3df0e 100644
@@ -161,6 +161,8 @@ config DRM_AMDGPU
 
 source "drivers/gpu/drm/amd/amdgpu/Kconfig"
 
+source "drivers/gpu/drm/amd/acp/Kconfig"
+
 source "drivers/gpu/drm/nouveau/Kconfig"
 
 config DRM_I810
drivers/gpu/drm/amd/acp/Kconfig new file +10 -0
new file mode 100644
index 0000000..1de4fe7
@@ -0,0 +1,10 @@
+menu "ACP Configuration"
+
+config DRM_AMD_ACP
+       bool "Enable ACP IP support"
+       default y
+       depends on MFD_CORE
+       help
+	Choose this option to enable ACP IP support for AMD SOCs.
+
+endmenu
drivers/gpu/drm/amd/acp/Makefile new file +9 -0
new file mode 100644
index 0000000..c8c3303
@@ -0,0 +1,9 @@
+#
+# Makefile for the ACP, which is a sub-component
+# of AMDSOC/AMDGPU drm driver.
+# It provides the HW control for ACP related functionalities.
+
+ccflags-y += -Idrivers/gpu/drm/amd/include/asic_reg/acp
+subdir-ccflags-y += -I$(AMDACPPATH)/ -I$(AMDACPPATH)/include
+
+AMD_ACP_FILES := $(AMDACPPATH)/acp_hw.o
drivers/gpu/drm/amd/acp/acp_hw.c new file +1133 -0

1,133 lines not on this page · raw message

drivers/gpu/drm/amd/acp/acp_hw.h new file +91 -0

91 lines not on this page · raw message

drivers/gpu/drm/amd/acp/include/acp_gfx_if.h new file +49 -0

49 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/Makefile +12 -1
index 04c2707..44de879 100644
@@ -5,7 +5,8 @@
 ccflags-y := -Iinclude/drm -Idrivers/gpu/drm/amd/include/asic_reg \
 	-Idrivers/gpu/drm/amd/include \
 	-Idrivers/gpu/drm/amd/amdgpu \
-	-Idrivers/gpu/drm/amd/scheduler
+	-Idrivers/gpu/drm/amd/scheduler \
+	-Idrivers/gpu/drm/amd/acp/include
 
 amdgpu-y := amdgpu_drv.o
 
@@ -89,6 +90,16 @@ amdgpu-y += \
 	../scheduler/sched_fence.o \
 	amdgpu_sched.o
 
+# ACP componet
+ifneq ($(CONFIG_DRM_AMD_ACP),)
+amdgpu-y += amdgpu_acp.o
+
+AMDACPPATH := ../acp
+include drivers/gpu/drm/amd/acp/Makefile
+
+amdgpu-y += $(AMD_ACP_FILES)
+endif
+
 amdgpu-$(CONFIG_COMPAT) += amdgpu_ioc32.o
 amdgpu-$(CONFIG_VGA_SWITCHEROO) += amdgpu_atpx_handler.o
 amdgpu-$(CONFIG_ACPI) += amdgpu_acpi.o
drivers/gpu/drm/amd/amdgpu/amdgpu.h +5 -0
index 2fc58e6..261ae46 100644
@@ -52,6 +52,7 @@
 #include "amdgpu_irq.h"
 #include "amdgpu_ucode.h"
 #include "amdgpu_gds.h"
+#include "amdgpu_acp.h"
 
 #include "gpu_scheduler.h"
 
@@ -1935,6 +1936,10 @@ struct amdgpu_device {
 	struct pci_dev			*pdev;
 	struct rw_semaphore		exclusive_lock;
 
+#ifdef CONFIG_DRM_AMD_ACP
+	struct amdgpu_acp		acp;
+#endif
+
 	/* ASIC */
 	enum amd_asic_type		asic_type;
 	uint32_t			family;
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c new file +208 -0

208 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h new file +40 -0
new file mode 100644
index 0000000..77b70e6
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __AMDGPU_ACP_H__
+#define __AMDGPU_ACP_H__
+
+#include <linux/mfd/core.h>
+
+struct amdgpu_acp {
+	struct device *parent;
+	void *cgs_device;
+	struct amd_acp_private *private;
+	struct mfd_cell *acp_cell;
+};
+
+extern const struct amd_ip_funcs acp_ip_funcs;
+
+#endif /* __AMDGPU_ACP_H__ */
drivers/gpu/drm/amd/amdgpu/vi.c +12 -0
index 552d9e75..faa142d 100644
@@ -71,6 +71,9 @@
 #include "uvd_v5_0.h"
 #include "uvd_v6_0.h"
 #include "vce_v3_0.h"
+#if defined(CONFIG_DRM_AMD_ACP)
+#include "amdgpu_acp.h"
+#endif
 
 /*
  * Indirect registers accessor
@@ -1298,6 +1301,15 @@ static const struct amdgpu_ip_block_version cz_ip_blocks[] =
 		.rev = 0,
 		.funcs = &vce_v3_0_ip_funcs,
 	},
+#if defined(CONFIG_DRM_AMD_ACP)
+	{
+		.type = AMD_IP_BLOCK_TYPE_ACP,
+		.major = 2,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &acp_ip_funcs,
+	},
+#endif
 };
 
 int vi_set_ip_blocks(struct amdgpu_device *adev)
drivers/gpu/drm/amd/include/amd_shared.h +1 -0
index 68a8eaa..54bf96a 100644
@@ -72,6 +72,7 @@ enum amd_ip_block_type {
 	AMD_IP_BLOCK_TYPE_SDMA,
 	AMD_IP_BLOCK_TYPE_UVD,
 	AMD_IP_BLOCK_TYPE_VCE,
+	AMD_IP_BLOCK_TYPE_ACP,
 };
 
 enum amd_clockgating_state {
include/linux/mfd/amd_acp.h new file +211 -0

211 lines not on this page · raw message

Signature · 6 lines
1.8.3.1

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
#2 lore raw

PATCH 3/3 ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

In reply to Alex Deucher

From: Maruthi Srinivas Bayyavarapu <[email protected]>

ACP IP block consists of dedicated DMA and I2S blocks. The PCM driver provides the DMA and CPU DAI components to ALSA core.

v2: squash in Kconfig fix
v3: squash additional commits, convert to mfd, drop rt286 changes
v4: squash in naming fixes

---
 sound/soc/Kconfig           |   1 +
 sound/soc/Makefile          |   1 +
 sound/soc/amd/Kconfig       |   5 +
 sound/soc/amd/Makefile      |   3 +
 sound/soc/amd/acp-pcm-dma.c | 676 ++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 686 insertions(+)
 create mode 100644 sound/soc/amd/Kconfig
 create mode 100644 sound/soc/amd/Makefile
 create mode 100644 sound/soc/amd/acp-pcm-dma.c
sound/soc/Kconfig +1 -0
index 2ae9619..1b9ce31 100644
@@ -32,6 +32,7 @@ config SND_SOC_GENERIC_DMAENGINE_PCM
 
 # All the supported SoCs
 source "sound/soc/adi/Kconfig"
+source "sound/soc/amd/Kconfig"
 source "sound/soc/atmel/Kconfig"
 source "sound/soc/au1x/Kconfig"
 source "sound/soc/bcm/Kconfig"
sound/soc/Makefile +1 -0
index e189903..a6cbb99 100644
@@ -14,6 +14,7 @@ obj-$(CONFIG_SND_SOC)	+= snd-soc-core.o
 obj-$(CONFIG_SND_SOC)	+= codecs/
 obj-$(CONFIG_SND_SOC)	+= generic/
 obj-$(CONFIG_SND_SOC)	+= adi/
+obj-$(CONFIG_SND_SOC)	+= amd/
 obj-$(CONFIG_SND_SOC)	+= atmel/
 obj-$(CONFIG_SND_SOC)	+= au1x/
 obj-$(CONFIG_SND_SOC)	+= bcm/
sound/soc/amd/Kconfig new file +5 -0
new file mode 100644
index 0000000..ce39979
@@ -0,0 +1,5 @@
+config SND_SOC_AMD_ACP
+	tristate "AMD Audio Coprocessor support"
+	depends on MFD_CORE
+	help
+	 This option enables ACP support (DMA,I2S) on AMD platforms.
sound/soc/amd/Makefile new file +3 -0
new file mode 100644
index 0000000..1a66ec0
@@ -0,0 +1,3 @@
+snd-soc-acp-pcm-objs	:= acp-pcm-dma.o
+
+obj-$(CONFIG_SND_SOC_AMD_ACP) += snd-soc-acp-pcm.o
sound/soc/amd/acp-pcm-dma.c new file +676 -0

676 lines not on this page · raw message

Signature · 6 lines
1.8.3.1

_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel
maintainer
#3 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 8 lines · Alex Deucher

From: Maruthi Srinivas Bayyavarapu <[email protected]>

ACP IP block consists of dedicated DMA and I2S blocks. The PCM driver provides the DMA and CPU DAI components to ALSA core.

v2: squash in Kconfig fix
v3: squash additional commits, convert to mfd, drop rt286 changes
v4: squash in naming fixes

To repeat what I said last time:

| Please follow the patch submission process in SubmittingPatches: put any
| versioning in the subject line inside the [] and put noise like inter
| version changelogs after the ---.
Quoted 14 lines

+/*
+ * AMD ALSA SoC PCM Driver
+ *
+ * Copyright 2014-2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.

Quoted 8 lines

+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.

I also remain very concerned about this non-GPL license you are using. Please do not ignore review comments like this :(

#4 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 10 lines · Mark Brown
Quoted 8 lines · Alex Deucher

From: Maruthi Srinivas Bayyavarapu <[email protected]>

ACP IP block consists of dedicated DMA and I2S blocks. The PCM driver provides the DMA and CPU DAI components to ALSA core.

v2: squash in Kconfig fix
v3: squash additional commits, convert to mfd, drop rt286 changes
v4: squash in naming fixes

Quoted 5 lines

To repeat what I said last time:

| Please follow the patch submission process in SubmittingPatches: put any
| versioning in the subject line inside the [] and put noise like inter
| version changelogs after the ---.

I addressed this in my coverletter. Some subsystems prefer to retain that information. If you'd prefer to not have them in the audio patch I will remove them.

Quoted 24 lines
Quoted 15 lines

+/*
+ * AMD ALSA SoC PCM Driver
+ *
+ * Copyright 2014-2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.

Quoted 8 lines

+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.

Quoted 2 lines

I also remain very concerned about this non-GPL license you are using. Please do not ignore review comments like this :(

I mentioned this in the cover letter as well. Most if not all of the drm drivers are licensed the same way. IANAL, but I am not aware of any concerns about them.

Alex
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

maintainer
#5 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 6 lines · Alex Deucher
Quoted 3 lines · Mark Brown
Quoted 1 line

v4: squash in naming fixes

To repeat what I said last time:

Quoted 7 lines
Quoted 3 lines
| Please follow the patch submission process in SubmittingPatches: put any
| versioning in the subject line inside the [] and put noise like inter
| version changelogs after the ---.

I addressed this in my coverletter. Some subsystems prefer to retain that information. If you'd prefer to not have them in the audio patch I will remove them.

When I said to follow the standard process here that's what I meant, yes. It's basically just the graphics subsystem that does something different here, and note that if you are going to include a changelog (either in the normal place or the DRM place) it really ought to actually describe the changes that have been made - the above doesn't reflect the changes that were made at all.

Quoted 6 lines
Quoted 2 lines

I also remain very concerned about this non-GPL license you are using. Please do not ignore review comments like this :(

I mentioned this in the cover letter as well. Most if not all of the drm drivers are licensed the same way. IANAL, but I am not aware of any concerns about them.

DRM is a special case here since there has always been work to share bits of the code with other operating systems, the licensing for DRM is very unusual within the kernel. A quick sampling of drivers suggests that this license is not universally used there either.

There is also the issue I raised with the fact that your non-GPL license statement does not appear to correspond to the MODULE_LICENSE() that you've included which claims the code is GPLed.

#6 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 15 lines · Mark Brown
Quoted 6 lines · Alex Deucher
Quoted 3 lines · Mark Brown
Quoted 1 line

v4: squash in naming fixes

To repeat what I said last time:

Quoted 7 lines
Quoted 3 lines
| Please follow the patch submission process in SubmittingPatches: put any
| versioning in the subject line inside the [] and put noise like inter
| version changelogs after the ---.

I addressed this in my coverletter. Some subsystems prefer to retain that information. If you'd prefer to not have them in the audio patch I will remove them.

Quoted 6 lines

When I said to follow the standard process here that's what I meant, yes. It's basically just the graphics subsystem that does something different here, and note that if you are going to include a changelog (either in the normal place or the DRM place) it really ought to actually describe the changes that have been made - the above doesn't reflect the changes that were made at all.

OK. I'll respin the patch.

Quoted 8 lines
Quoted 7 lines
Quoted 3 lines

I also remain very concerned about this non-GPL license you are using. Please do not ignore review comments like this :(

I mentioned this in the cover letter as well. Most if not all of the drm drivers are licensed the same way. IANAL, but I am not aware of any concerns about them.

Quoted 8 lines

DRM is a special case here since there has always been work to share bits of the code with other operating systems, the licensing for DRM is very unusual within the kernel. A quick sampling of drivers suggests that this license is not universally used there either.

There is also the issue I raised with the fact that your non-GPL license statement does not appear to correspond to the MODULE_LICENSE() that you've included which claims the code is GPLed.

Once again this is exactly what the other drm drivers and support modules do. AMD generally prefers to contribute code licensed under an MIT/X11 style licence which is why we've done this. IANAL, but I think that is compatible with "GPL and additional rights".

Alex
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

maintainer
#7 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 7 lines · Alex Deucher
Quoted 4 lines · Mark Brown

DRM is a special case here since there has always been work to share bits of the code with other operating systems, the licensing for DRM is very unusual within the kernel. A quick sampling of drivers suggests that this license is not universally used there either.

Quoted 8 lines
Quoted 3 lines

There is also the issue I raised with the fact that your non-GPL license statement does not appear to correspond to the MODULE_LICENSE() that you've included which claims the code is GPLed.

Once again this is exactly what the other drm drivers and support modules do. AMD generally prefers to contribute code licensed under an MIT/X11 style licence which is why we've done this. IANAL, but I think that is compatible with "GPL and additional rights".

This is not a DRM driver, it is an ASoC driver, and like I say DRM is a bit special. I do note that AMD has contributed at least its CCP and cpufreq drivers (indeed all their non-DRM code I could find with explicit license statements) under a normal kernel license. As far as I can tell this licensing is entirely confined to the DRM drivers.

I am not readily able to convince myself that this is compatible with the intent of exporting the ASoC APIs _GPL(), this looks like it can be used as the basis for a shim layer for non-GPL code (the licensing strategy seems very similar). I'd need to think through this carefully.

#8 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 23 lines · Mark Brown
Quoted 7 lines · Alex Deucher
Quoted 4 lines · Mark Brown

DRM is a special case here since there has always been work to share bits of the code with other operating systems, the licensing for DRM is very unusual within the kernel. A quick sampling of drivers suggests that this license is not universally used there either.

Quoted 8 lines
Quoted 3 lines

There is also the issue I raised with the fact that your non-GPL license statement does not appear to correspond to the MODULE_LICENSE() that you've included which claims the code is GPLed.

Once again this is exactly what the other drm drivers and support modules do. AMD generally prefers to contribute code licensed under an MIT/X11 style licence which is why we've done this. IANAL, but I think that is compatible with "GPL and additional rights".

This is not a DRM driver, it is an ASoC driver, and like I say DRM is a bit special. I do note that AMD has contributed at least its CCP and cpufreq drivers (indeed all their non-DRM code I could find with explicit license statements) under a normal kernel license. As far as I can tell this licensing is entirely confined to the DRM drivers.

Quoted 4 lines

I am not readily able to convince myself that this is compatible with the intent of exporting the ASoC APIs _GPL(), this looks like it can be used as the basis for a shim layer for non-GPL code (the licensing strategy seems very similar). I'd need to think through this carefully.

If you have strong concerns, I can double check with our lawyers and see about changing it, I'm just not sure how long that will take :( The licensing tends to vary based on the teams involved. The ACP audio and i2s was developed by the GPU teams and we've mostly worked on drm and Xorg stuff until now.

Alex
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel

maintainer
#9 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 8 lines · Alex Deucher
Quoted 5 lines · Mark Brown

This is not a DRM driver, it is an ASoC driver, and like I say DRM is a bit special. I do note that AMD has contributed at least its CCP and cpufreq drivers (indeed all their non-DRM code I could find with explicit license statements) under a normal kernel license. As far as I can tell this licensing is entirely confined to the DRM drivers.

Quoted 7 lines
Quoted 4 lines

I am not readily able to convince myself that this is compatible with the intent of exporting the ASoC APIs _GPL(), this looks like it can be used as the basis for a shim layer for non-GPL code (the licensing strategy seems very similar). I'd need to think through this carefully.

If you have strong concerns, I can double check with our lawyers and see about changing it, I'm just not sure how long that will take :(

I at the very least need to think about this, it's really important to me that we don't want to see proprietary drivers.

Quoted 3 lines

The licensing tends to vary based on the teams involved. The ACP audio and i2s was developed by the GPU teams and we've mostly worked on drm and Xorg stuff until now.

For reference the GPL code in the kernel includes at least:

drivers/crypto/ccp/ccp-crypto-*.c
drivers/cpufreq/powernow-k8.c
drivers/pinctrl/pinctrl-amd.c

and there's direct contributions from AMD there AFAICT, not just outside developers contributing support for AMD chips.

#10 lore raw

Re: [PATCH 3/3] ASoC: AMD: add AMD ASoC ACP-I2S driver [v4]

Quoted 14 lines · Mark Brown
Quoted 12 lines · Alex Deucher
Quoted 10 lines · Mark Brown

This is not a DRM driver, it is an ASoC driver, and like I say DRM is a bit special. I do note that AMD has contributed at least its CCP and cpufreq drivers (indeed all their non-DRM code I could find with explicit license statements) under a normal kernel license. As far as I can tell this licensing is entirely confined to the DRM drivers.

I am not readily able to convince myself that this is compatible with the intent of exporting the ASoC APIs _GPL(), this looks like it can be used as the basis for a shim layer for non-GPL code (the licensing strategy seems very similar). I'd need to think through this carefully.

Quoted 5 lines
Quoted 2 lines

If you have strong concerns, I can double check with our lawyers and see about changing it, I'm just not sure how long that will take :(

I at the very least need to think about this, it's really important to me that we don't want to see proprietary drivers.

That's definitely not our intention. Historically on the GPU side, we've preferred the X11/MIT license because it lets other projects like the BSDs more easily leverage our code.

Quoted 5 lines
Quoted 4 lines

The licensing tends to vary based on the teams involved. The ACP audio and i2s was developed by the GPU teams and we've mostly worked on drm and Xorg stuff until now.

Quoted 8 lines

For reference the GPL code in the kernel includes at least:

drivers/crypto/ccp/ccp-crypto-*.c
drivers/cpufreq/powernow-k8.c
drivers/pinctrl/pinctrl-amd.c

and there's direct contributions from AMD there AFAICT, not just outside developers contributing support for AMD chips.

Right, different teams within AMD. My team has historically only worked on GPU stuff and since we contributed to Xorg and the drm we've been approved to release code under the X11/MIT license.

Alex
_______________________________________________
dri-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/dri-devel