Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Bard liao]

> Date: Wed, 12 Dec 2018 12:23:33 +0100
> From: Takashi Iwai <tiwai@suse.de>
> To: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> Cc: alsa-devel@alsa-project.org, andriy.shevchenko@intel.com, Daniel
> 	Baluta <daniel.baluta@gmail.com>, liam.r.girdwood@linux.intel.com,
> 	vkoul@kernel.org, broonie@kernel.org, Alan Cox <alan@linux.intel.com>,
> 	sound-open-firmware@alsa-project.org
> Subject: Re: [alsa-devel] [PATCH v3 09/14] ASoC: SOF: Add firmware
> 	loader	support
> Message-ID: <s5htvjjj8ei.wl-tiwai@suse.de>
> Content-Type: text/plain; charset=US-ASCII
>
> On Tue, 11 Dec 2018 22:23:13 +0100,
> Pierre-Louis Bossart wrote:
>
>> +
>> +		snd_sof_dsp_block_write(sdev, offset,
>> +					(void *)block + sizeof(*block),
>> +					block->size);
>> +
>> +		/* next block */
>> +		block = (void *)block + sizeof(*block) + block->size;
> This may lead to an unaligned access.

Did you mean we should double check the block->size to
prevent access to an invalid address?

> Also how is the endianess guaranteed?

Did you mean we should guarantee the driver can work no
matter what kernel's endianess is?
ie. Use le32_to_cpu() to handle it?

>
>
> thanks,
>
> Takashi
>
>

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Takashi Iwai]

On Thu, 20 Dec 2018 03:11:55 +0100,
Bard liao wrote:
> 
> 
> > Date: Wed, 12 Dec 2018 12:23:33 +0100
> > From: Takashi Iwai <tiwai@suse.de>
> > To: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> > Cc: alsa-devel@alsa-project.org, andriy.shevchenko@intel.com, Daniel
> > 	Baluta <daniel.baluta@gmail.com>, liam.r.girdwood@linux.intel.com,
> > 	vkoul@kernel.org, broonie@kernel.org, Alan Cox <alan@linux.intel.com>,
> > 	sound-open-firmware@alsa-project.org
> > Subject: Re: [alsa-devel] [PATCH v3 09/14] ASoC: SOF: Add firmware
> > 	loader	support
> > Message-ID: <s5htvjjj8ei.wl-tiwai@suse.de>
> > Content-Type: text/plain; charset=US-ASCII
> >
> > On Tue, 11 Dec 2018 22:23:13 +0100,
> > Pierre-Louis Bossart wrote:
> >
> >> +
> >> +		snd_sof_dsp_block_write(sdev, offset,
> >> +					(void *)block + sizeof(*block),
> >> +					block->size);
> >> +
> >> +		/* next block */
> >> +		block = (void *)block + sizeof(*block) + block->size;
> > This may lead to an unaligned access.
> 
> Did you mean we should double check the block->size to
> prevent access to an invalid address?

You need two types of checks for the given data:
- The bounce check of block->size;
  We need to avoid out-of-bounce access.

- Alignment of block->size;
  For some non-x86 platforms, the access to an unaligned address might
  be illegal.

Oh, and recently another thing is sometimes needed for avoiding
Spectre.  This can be covered by array_index_nospec().


> > Also how is the endianess guaranteed?
> 
> Did you mean we should guarantee the driver can work no
> matter what kernel's endianess is?
> ie. Use le32_to_cpu() to handle it?

Depends on the implementation.  IIRC, topology API refuses the data
in a different endianess by checking the magic number at beginning.


thanks,

Takashi

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Pierre-Louis Bossart]

>>>> +		snd_sof_dsp_block_write(sdev, offset,
>>>> +					(void *)block + sizeof(*block),
>>>> +					block->size);
>>>> +
>>>> +		/* next block */
>>>> +		block = (void *)block + sizeof(*block) + block->size;
>>> This may lead to an unaligned access.
>> Did you mean we should double check the block->size to
>> prevent access to an invalid address?
> You need two types of checks for the given data:
> - The bounce check of block->size;
>    We need to avoid out-of-bounce access.
s/bounce/bounds ?
>
> - Alignment of block->size;
>    For some non-x86 platforms, the access to an unaligned address might
>    be illegal.

Maybe I am missing something but I don't see any sort of explicit 
restriction on alignment in the SOF tools. it looks implicit based on 
address offsets and bases.

Liam, do you see any negative side effects if we enforce a 32-bit 
alignment for all blocks (which essentially means all block sizes are 
multiple of 4)? we can try and experiment but it's better if we have an 
agreement on the design.

>
> Oh, and recently another thing is sometimes needed for avoiding
> Spectre.  This can be covered by array_index_nospec().
>
>
>>> Also how is the endianess guaranteed?
>> Did you mean we should guarantee the driver can work no
>> matter what kernel's endianess is?
>> ie. Use le32_to_cpu() to handle it?
> Depends on the implementation.  IIRC, topology API refuses the data
> in a different endianess by checking the magic number at beginning.
snd_sof_dsp_block_write() is implemented with a platform-specific 
callback, I'd expect any endianess issues to be handled in that 
platform-specific code?

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Bard liao]

On 12/20/2018 11:07 PM, Pierre-Louis Bossart wrote:
>
>>>>> + snd_sof_dsp_block_write(sdev, offset,
>>>>> +                    (void *)block + sizeof(*block),
>>>>> +                    block->size);
>>>>> +
>>>>> +        /* next block */
>>>>> +        block = (void *)block + sizeof(*block) + block->size;
>>>> This may lead to an unaligned access.
>>> Did you mean we should double check the block->size to
>>> prevent access to an invalid address?
>> You need two types of checks for the given data:
>> - The bounce check of block->size;
>>    We need to avoid out-of-bounce access.
> s/bounce/bounds ?
>>
>> - Alignment of block->size;
>>    For some non-x86 platforms, the access to an unaligned address might
>>    be illegal.
>
> Maybe I am missing something but I don't see any sort of explicit 
> restriction on alignment in the SOF tools. it looks implicit based on 
> address offsets and bases.

Maybe get_unaligned() is good enough to avoid unaligned access?

>
> Liam, do you see any negative side effects if we enforce a 32-bit 
> alignment for all blocks (which essentially means all block sizes are 
> multiple of 4)? we can try and experiment but it's better if we have 
> an agreement on the design.
>
>
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Takashi Iwai]

On Fri, 21 Dec 2018 10:05:04 +0100,
Bard liao wrote:
> 
> 
> On 12/20/2018 11:07 PM, Pierre-Louis Bossart wrote:
> >
> >>>>> + snd_sof_dsp_block_write(sdev, offset,
> >>>>> +                    (void *)block + sizeof(*block),
> >>>>> +                    block->size);
> >>>>> +
> >>>>> +        /* next block */
> >>>>> +        block = (void *)block + sizeof(*block) + block->size;
> >>>> This may lead to an unaligned access.
> >>> Did you mean we should double check the block->size to
> >>> prevent access to an invalid address?
> >> You need two types of checks for the given data:
> >> - The bounce check of block->size;
> >>    We need to avoid out-of-bounce access.
> > s/bounce/bounds ?
> >>
> >> - Alignment of block->size;
> >>    For some non-x86 platforms, the access to an unaligned address might
> >>    be illegal.
> >
> > Maybe I am missing something but I don't see any sort of explicit
> > restriction on alignment in the SOF tools. it looks implicit based
> > on address offsets and bases.
> 
> Maybe get_unaligned() is good enough to avoid unaligned access?

That's another option, but you'd need to put everywhere.

An alternative is to just copy the data on a new block header on the
stack temporarily and refer it.

Or simply assuring the alignment by checking block->size as Pierre
suggested...


Takashi


> 
> >
> > Liam, do you see any negative side effects if we enforce a 32-bit
> > alignment for all blocks (which essentially means all block sizes
> > are multiple of 4)? we can try and experiment but it's better if we
> > have an agreement on the design.
> >
> >
> 
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware, loader support

[Liam Girdwood]

On Fri, 2018-12-21 at 10:57 +0100, Takashi Iwai wrote:
> On Fri, 21 Dec 2018 10:05:04 +0100,
> Bard liao wrote:
> > 
> > On 12/20/2018 11:07 PM, Pierre-Louis Bossart wrote:
> > > > > > > + snd_sof_dsp_block_write(sdev, offset,
> > > > > > > +                    (void *)block + sizeof(*block),
> > > > > > > +                    block->size);
> > > > > > > +
> > > > > > > +        /* next block */
> > > > > > > +        block = (void *)block + sizeof(*block) + block->size;
> > > > > > This may lead to an unaligned access.
> > > > > Did you mean we should double check the block->size to
> > > > > prevent access to an invalid address?
> > > > You need two types of checks for the given data:
> > > > - The bounce check of block->size;
> > > >    We need to avoid out-of-bounce access.
> > > s/bounce/bounds ?
> > > > - Alignment of block->size;
> > > >    For some non-x86 platforms, the access to an unaligned address might
> > > >    be illegal.
> > > 
> > > Maybe I am missing something but I don't see any sort of explicit
> > > restriction on alignment in the SOF tools. it looks implicit based
> > > on address offsets and bases.
> > 
> > Maybe get_unaligned() is good enough to avoid unaligned access?
> 
> That's another option, but you'd need to put everywhere.
> 
> An alternative is to just copy the data on a new block header on the
> stack temporarily and refer it.
> 
> Or simply assuring the alignment by checking block->size as Pierre
> suggested...

Yep, checking size is good for me too.

> 
> 
> Takashi
> 
> 
> > > Liam, do you see any negative side effects if we enforce a 32-bit
> > > alignment for all blocks (which essentially means all block sizes
> > > are multiple of 4)? we can try and experiment but it's better if we
> > > have an agreement on the design.


This shouldn't cause any problems, iirc all block sizes are rounded up to
nearest uint32_t anyway (as we have exception handler blocks that are 3 bytes of
assembly in size and rounded up to 4).

Liam

> > > 
> > > 
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Pierre-Louis Bossart]

Thanks for the reviews Mark, much appreciated.

+int snd_sof_load_firmware_memcpy(struct snd_sof_dev *sdev)
>> +{
>> +	struct snd_sof_pdata *plat_data = dev_get_platdata(sdev->dev);
>> +	const char *fw_filename;
>> +	int ret;
> This never actually calls the load_firmware() operation for the DSP
> AFAICT?

it's actually the implementation of the load_firmware callback, see e.g. 
for Baytrail/Broadwell

/*Firmware loading */
     .load_firmware    = snd_sof_load_firmware_memcpy,

>
>> +	ret = request_firmware(&plat_data->fw, fw_filename, sdev->dev);
>> +
>> +	if (ret < 0) {
>> +		dev_err(sdev->dev, "error: request firmware failed err: %d\n",
>> +			ret);
>> +		return ret;
>> +	}
> I'd suggest logging the name of the firmware we tried to load, users
> will thank you.
yes indeed, thanks for the suggestion, will add this right away.
>
>> +	/* create fw_version debugfs to store boot version info */
>> +	if (sdev->first_boot) {
>> +		ret = snd_sof_debugfs_buf_create_item(sdev, &sdev->fw_version,
>> +						      sizeof(sdev->fw_version),
>> +						      "fw_version");
>> +
>> +		if (ret < 0) {
>> +			dev_err(sdev->dev, "error: cannot create debugfs for fw_version\n");
>> +			return ret;
>> +		}
>> +	}
> As Andy said elsewhere debugfs stuff like that probably shouldn't be
> fatal.
yes, we've fixed this already.

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 1043 bytes --]

On Tue, Dec 11, 2018 at 03:23:13PM -0600, Pierre-Louis Bossart wrote:

> +int snd_sof_load_firmware_memcpy(struct snd_sof_dev *sdev)
> +{
> +	struct snd_sof_pdata *plat_data = dev_get_platdata(sdev->dev);
> +	const char *fw_filename;
> +	int ret;

This never actually calls the load_firmware() operation for the DSP
AFAICT?

> +	ret = request_firmware(&plat_data->fw, fw_filename, sdev->dev);
> +
> +	if (ret < 0) {
> +		dev_err(sdev->dev, "error: request firmware failed err: %d\n",
> +			ret);
> +		return ret;
> +	}

I'd suggest logging the name of the firmware we tried to load, users
will thank you.

> +	/* create fw_version debugfs to store boot version info */
> +	if (sdev->first_boot) {
> +		ret = snd_sof_debugfs_buf_create_item(sdev, &sdev->fw_version,
> +						      sizeof(sdev->fw_version),
> +						      "fw_version");
> +
> +		if (ret < 0) {
> +			dev_err(sdev->dev, "error: cannot create debugfs for fw_version\n");
> +			return ret;
> +		}
> +	}

As Andy said elsewhere debugfs stuff like that probably shouldn't be
fatal.

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 1276 bytes --]

On Tue, Dec 11, 2018 at 05:54:02PM -0600, Pierre-Louis Bossart wrote:
> On 12/11/18 4:38 PM, Andy Shevchenko wrote:
> > On Tue, Dec 11, 2018 at 03:23:13PM -0600, Pierre-Louis Bossart wrote:

> > > +	struct sof_ipc_window *w = (struct sof_ipc_window *)ext_hdr;
> > > +
> > > +	int ret = 0;

> > I don't see how it's used. Perhaps you need to check code with `make W=1`.

> Darn, we missed this one. I thought we were using W=1 on github but we
> aren't, this will be fixed.

> Though W=1 doesn't report this one, so need to re-inspect this. Thanks for
> the sighting.

This is one reason not to unconditionally init stuff on declaration -
it masks some warnings.

> > > +int snd_sof_load_firmware(struct snd_sof_dev *sdev)
> > > +{
> > > +	dev_dbg(sdev->dev, "loading firmware\n");
> > Noise.
> > Better to introduce a trace points and drop all these kind of messages.

> it's not that bad, most people understand what dmesg is, it happens once and
> only if you have dynamic debug.

Then everyone adds their print on boot and the log gets huge (and slow
if you push debug through serial during development).  As a rule of
thumb I tend to suggest that if you're not reporting something you
discovered at runtime there's probably already other ways of getting
equivalent trace.

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Takashi Iwai]

On Tue, 11 Dec 2018 22:23:13 +0100,
Pierre-Louis Bossart wrote:
> 
> +/* generic module parser for mmaped DSPs */
> +int snd_sof_parse_module_memcpy(struct snd_sof_dev *sdev,
> +				struct snd_sof_mod_hdr *module)
> +{
> +	struct snd_sof_blk_hdr *block;
> +	int count;
> +	u32 offset;
> +
> +	dev_dbg(sdev->dev, "new module size 0x%x blocks 0x%x type 0x%x\n",
> +		module->size, module->num_blocks, module->type);
> +
> +	block = (void *)module + sizeof(*module);
> +
> +	for (count = 0; count < module->num_blocks; count++) {

Need a sanity check that it won't go beyond the actual firmware size.
User may pass a malicious module data, e.g. with extra large
num_blocks.

> +		if (block->size == 0) {
> +			dev_warn(sdev->dev,
> +				 "warning: block %d size zero\n", count);
> +			dev_warn(sdev->dev, " type 0x%x offset 0x%x\n",
> +				 block->type, block->offset);
> +			continue;
> +		}
> +
> +		switch (block->type) {
> +		case SOF_BLK_IMAGE:
> +		case SOF_BLK_CACHE:
> +		case SOF_BLK_REGS:
> +		case SOF_BLK_SIG:
> +		case SOF_BLK_ROM:
> +			continue;	/* not handled atm */
> +		case SOF_BLK_TEXT:
> +		case SOF_BLK_DATA:
> +			offset = block->offset;
> +			break;
> +		default:
> +			dev_err(sdev->dev, "error: bad type 0x%x for block 0x%x\n",
> +				block->type, count);
> +			return -EINVAL;
> +		}
> +
> +		dev_dbg(sdev->dev,
> +			"block %d type 0x%x size 0x%x ==>  offset 0x%x\n",
> +			count, block->type, block->size, offset);
> +
> +		snd_sof_dsp_block_write(sdev, offset,
> +					(void *)block + sizeof(*block),
> +					block->size);
> +
> +		/* next block */
> +		block = (void *)block + sizeof(*block) + block->size;

This may lead to an unaligned access.
Also how is the endianess guaranteed?


thanks,

Takashi

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Pierre-Louis Bossart]

On 12/11/18 4:38 PM, Andy Shevchenko wrote:
> On Tue, Dec 11, 2018 at 03:23:13PM -0600, Pierre-Louis Bossart wrote:
>> From: Liam Girdwood <liam.r.girdwood@linux.intel.com>
>>
>> The firmware loader exports APIs that can be called by core to load and
>> process multiple different file formats.
>> +static int get_ext_windows(struct snd_sof_dev *sdev,
>> +			   struct sof_ipc_ext_data_hdr *ext_hdr)
>> +{
>> +	struct sof_ipc_window *w = (struct sof_ipc_window *)ext_hdr;
>> +
>> +	int ret = 0;
> I don't see how it's used. Perhaps you need to check code with `make W=1`.

Darn, we missed this one. I thought we were using W=1 on github but we 
aren't, this will be fixed.

Though W=1 doesn't report this one, so need to re-inspect this. Thanks 
for the sighting.

>
>> +	size_t size;
>> +
>> +	if (w->num_windows == 0 || w->num_windows > SOF_IPC_MAX_ELEMS)
>> +		return -EINVAL;
>> +
>> +	size = sizeof(*w) + sizeof(struct sof_ipc_window_elem) * w->num_windows;
>> +
>> +	/* keep a local copy of the data */
>> +	sdev->info_window = kmemdup(w, size, GFP_KERNEL);
>> +	if (!sdev->info_window)
>> +		return -ENOMEM;
>> +
>> +	return ret;
>> +}
>> +			dev_warn(sdev->dev,
>> +				 "warning: block %d size zero\n", count);
>> +			dev_warn(sdev->dev, " type 0x%x offset 0x%x\n",
>> +				 block->type, block->offset);
> Hmm... Why do we need a kernel level duplication in words?
Not sure I get this one, it's just a warning message where you don't 
copy a block of size zero into the target memory.
>
>> +int snd_sof_load_firmware(struct snd_sof_dev *sdev)
>> +{
>> +	dev_dbg(sdev->dev, "loading firmware\n");
> Noise.
> Better to introduce a trace points and drop all these kind of messages.

it's not that bad, most people understand what dmesg is, it happens once 
and only if you have dynamic debug.

At some point we also thought about something like drm_debug where you 
can specify which parts you are interested in, in addition to the 
verbosity level.

>
>> +
>> +	if (sdev->ops->load_firmware)
>> +		return sdev->ops->load_firmware(sdev);
>> +	return 0;
>> +}
>> +EXPORT_SYMBOL(snd_sof_load_firmware);

Re: [PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Andy Shevchenko]

On Tue, Dec 11, 2018 at 03:23:13PM -0600, Pierre-Louis Bossart wrote:
> From: Liam Girdwood <liam.r.girdwood@linux.intel.com>
> 
> The firmware loader exports APIs that can be called by core to load and
> process multiple different file formats.

> +static int get_ext_windows(struct snd_sof_dev *sdev,
> +			   struct sof_ipc_ext_data_hdr *ext_hdr)
> +{
> +	struct sof_ipc_window *w = (struct sof_ipc_window *)ext_hdr;
> +

> +	int ret = 0;

I don't see how it's used. Perhaps you need to check code with `make W=1`.

> +	size_t size;
> +
> +	if (w->num_windows == 0 || w->num_windows > SOF_IPC_MAX_ELEMS)
> +		return -EINVAL;
> +
> +	size = sizeof(*w) + sizeof(struct sof_ipc_window_elem) * w->num_windows;
> +
> +	/* keep a local copy of the data */
> +	sdev->info_window = kmemdup(w, size, GFP_KERNEL);
> +	if (!sdev->info_window)
> +		return -ENOMEM;
> +
> +	return ret;
> +}

> +			dev_warn(sdev->dev,
> +				 "warning: block %d size zero\n", count);
> +			dev_warn(sdev->dev, " type 0x%x offset 0x%x\n",
> +				 block->type, block->offset);

Hmm... Why do we need a kernel level duplication in words?

> +int snd_sof_load_firmware(struct snd_sof_dev *sdev)
> +{

> +	dev_dbg(sdev->dev, "loading firmware\n");

Noise.
Better to introduce a trace points and drop all these kind of messages.

> +
> +	if (sdev->ops->load_firmware)
> +		return sdev->ops->load_firmware(sdev);
> +	return 0;
> +}
> +EXPORT_SYMBOL(snd_sof_load_firmware);

-- 
With Best Regards,
Andy Shevchenko

[PATCH v3 09/14] ASoC: SOF: Add firmware loader support

[Pierre-Louis Bossart]

From: Liam Girdwood <liam.r.girdwood@linux.intel.com>

The firmware loader exports APIs that can be called by core to load and
process multiple different file formats.

Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
---
 sound/soc/sof/loader.c | 322 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 322 insertions(+)
 create mode 100644 sound/soc/sof/loader.c

diff --git a/sound/soc/sof/loader.c b/sound/soc/sof/loader.c
new file mode 100644
index 000000000000..dbadd8c521e6
--- /dev/null
+++ b/sound/soc/sof/loader.c
@@ -0,0 +1,322 @@
+// SPDX-License-Identifier: (GPL-2.0 OR BSD-3-Clause)
+//
+// This file is provided under a dual BSD/GPLv2 license.  When using or
+// redistributing this file, you may do so under either license.
+//
+// Copyright(c) 2018 Intel Corporation. All rights reserved.
+//
+// Author: Liam Girdwood <liam.r.girdwood@linux.intel.com>
+//
+// Generic firmware loader.
+//
+
+#include <linux/firmware.h>
+#include <sound/sof.h>
+#include "ops.h"
+
+static int get_ext_windows(struct snd_sof_dev *sdev,
+			   struct sof_ipc_ext_data_hdr *ext_hdr)
+{
+	struct sof_ipc_window *w = (struct sof_ipc_window *)ext_hdr;
+
+	int ret = 0;
+	size_t size;
+
+	if (w->num_windows == 0 || w->num_windows > SOF_IPC_MAX_ELEMS)
+		return -EINVAL;
+
+	size = sizeof(*w) + sizeof(struct sof_ipc_window_elem) * w->num_windows;
+
+	/* keep a local copy of the data */
+	sdev->info_window = kmemdup(w, size, GFP_KERNEL);
+	if (!sdev->info_window)
+		return -ENOMEM;
+
+	return ret;
+}
+
+/* parse the extended FW boot data structures from FW boot message */
+int snd_sof_fw_parse_ext_data(struct snd_sof_dev *sdev, u32 offset)
+{
+	struct sof_ipc_ext_data_hdr *ext_hdr;
+	void *ext_data;
+	int ret = 0;
+
+	ext_data = kzalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!ext_data)
+		return -ENOMEM;
+
+	/* get first header */
+	snd_sof_dsp_block_read(sdev, offset, ext_data, sizeof(*ext_hdr));
+	ext_hdr = (struct sof_ipc_ext_data_hdr *)ext_data;
+
+	while (ext_hdr->hdr.cmd == SOF_IPC_FW_READY) {
+		/* read in ext structure */
+		offset += sizeof(*ext_hdr);
+		snd_sof_dsp_block_read(sdev, offset,
+				       ext_data + sizeof(*ext_hdr),
+				       ext_hdr->hdr.size - sizeof(*ext_hdr));
+
+		dev_dbg(sdev->dev, "found ext header type %d size 0x%x\n",
+			ext_hdr->type, ext_hdr->hdr.size);
+
+		/* process structure data */
+		switch (ext_hdr->type) {
+		case SOF_IPC_EXT_DMA_BUFFER:
+			break;
+		case SOF_IPC_EXT_WINDOW:
+			ret = get_ext_windows(sdev, ext_hdr);
+			break;
+		default:
+			break;
+		}
+
+		if (ret < 0) {
+			dev_err(sdev->dev, "error: failed to parse ext data type %d\n",
+				ext_hdr->type);
+			goto out;
+		}
+
+		/* move to next header */
+		offset += ext_hdr->hdr.size;
+		snd_sof_dsp_block_read(sdev, offset, ext_data,
+				       sizeof(*ext_hdr));
+		ext_hdr = (struct sof_ipc_ext_data_hdr *)ext_data;
+	}
+out:
+	kfree(ext_data);
+	return ret;
+}
+EXPORT_SYMBOL(snd_sof_fw_parse_ext_data);
+
+/* generic module parser for mmaped DSPs */
+int snd_sof_parse_module_memcpy(struct snd_sof_dev *sdev,
+				struct snd_sof_mod_hdr *module)
+{
+	struct snd_sof_blk_hdr *block;
+	int count;
+	u32 offset;
+
+	dev_dbg(sdev->dev, "new module size 0x%x blocks 0x%x type 0x%x\n",
+		module->size, module->num_blocks, module->type);
+
+	block = (void *)module + sizeof(*module);
+
+	for (count = 0; count < module->num_blocks; count++) {
+		if (block->size == 0) {
+			dev_warn(sdev->dev,
+				 "warning: block %d size zero\n", count);
+			dev_warn(sdev->dev, " type 0x%x offset 0x%x\n",
+				 block->type, block->offset);
+			continue;
+		}
+
+		switch (block->type) {
+		case SOF_BLK_IMAGE:
+		case SOF_BLK_CACHE:
+		case SOF_BLK_REGS:
+		case SOF_BLK_SIG:
+		case SOF_BLK_ROM:
+			continue;	/* not handled atm */
+		case SOF_BLK_TEXT:
+		case SOF_BLK_DATA:
+			offset = block->offset;
+			break;
+		default:
+			dev_err(sdev->dev, "error: bad type 0x%x for block 0x%x\n",
+				block->type, count);
+			return -EINVAL;
+		}
+
+		dev_dbg(sdev->dev,
+			"block %d type 0x%x size 0x%x ==>  offset 0x%x\n",
+			count, block->type, block->size, offset);
+
+		snd_sof_dsp_block_write(sdev, offset,
+					(void *)block + sizeof(*block),
+					block->size);
+
+		/* next block */
+		block = (void *)block + sizeof(*block) + block->size;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(snd_sof_parse_module_memcpy);
+
+static int check_header(struct snd_sof_dev *sdev, const struct firmware *fw)
+{
+	struct snd_sof_fw_header *header;
+
+	/* Read the header information from the data pointer */
+	header = (struct snd_sof_fw_header *)fw->data;
+
+	/* verify FW sig */
+	if (strncmp(header->sig, SND_SOF_FW_SIG, SND_SOF_FW_SIG_SIZE) != 0) {
+		dev_err(sdev->dev, "error: invalid firmware signature\n");
+		return -EINVAL;
+	}
+
+	/* check size is valid */
+	if (fw->size != header->file_size + sizeof(*header)) {
+		dev_err(sdev->dev, "error: invalid filesize mismatch got 0x%zx expected 0x%zx\n",
+			fw->size, header->file_size + sizeof(*header));
+		return -EINVAL;
+	}
+
+	dev_dbg(sdev->dev, "header size=0x%x modules=0x%x abi=0x%x size=%zu\n",
+		header->file_size, header->num_modules,
+		header->abi, sizeof(*header));
+
+	return 0;
+}
+
+static int load_modules(struct snd_sof_dev *sdev, const struct firmware *fw)
+{
+	struct snd_sof_fw_header *header;
+	struct snd_sof_mod_hdr *module;
+	int (*load_module)(struct snd_sof_dev *sof_dev,
+			   struct snd_sof_mod_hdr *hdr);
+	int ret, count;
+
+	header = (struct snd_sof_fw_header *)fw->data;
+	load_module = sdev->ops->load_module;
+	if (!load_module)
+		return -EINVAL;
+
+	/* parse each module */
+	module = (void *)fw->data + sizeof(*header);
+	for (count = 0; count < header->num_modules; count++) {
+		/* module */
+		ret = load_module(sdev, module);
+		if (ret < 0) {
+			dev_err(sdev->dev, "error: invalid module %d\n", count);
+			return ret;
+		}
+		module = (void *)module + sizeof(*module) + module->size;
+	}
+
+	return 0;
+}
+
+int snd_sof_load_firmware_memcpy(struct snd_sof_dev *sdev)
+{
+	struct snd_sof_pdata *plat_data = dev_get_platdata(sdev->dev);
+	const char *fw_filename;
+	int ret;
+
+	/* set code loading condition to true */
+	sdev->code_loading = 1;
+	fw_filename = plat_data->machine->sof_fw_filename;
+
+	ret = request_firmware(&plat_data->fw, fw_filename, sdev->dev);
+
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: request firmware failed err: %d\n",
+			ret);
+		return ret;
+	}
+
+	/* make sure the FW header and file is valid */
+	ret = check_header(sdev, plat_data->fw);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: invalid FW header\n");
+		goto error;
+	}
+
+	/* prepare the DSP for FW loading */
+	ret = snd_sof_dsp_reset(sdev);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: failed to reset DSP\n");
+		goto error;
+	}
+
+	/* parse and load firmware modules to DSP */
+	ret = load_modules(sdev, plat_data->fw);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: invalid FW modules\n");
+		goto error;
+	}
+
+	return 0;
+
+error:
+	release_firmware(plat_data->fw);
+	return ret;
+
+}
+EXPORT_SYMBOL(snd_sof_load_firmware_memcpy);
+
+int snd_sof_load_firmware(struct snd_sof_dev *sdev)
+{
+	dev_dbg(sdev->dev, "loading firmware\n");
+
+	if (sdev->ops->load_firmware)
+		return sdev->ops->load_firmware(sdev);
+	return 0;
+}
+EXPORT_SYMBOL(snd_sof_load_firmware);
+
+int snd_sof_run_firmware(struct snd_sof_dev *sdev)
+{
+	int ret;
+
+	init_waitqueue_head(&sdev->boot_wait);
+	sdev->boot_complete = false;
+
+	/* create fw_version debugfs to store boot version info */
+	if (sdev->first_boot) {
+		ret = snd_sof_debugfs_buf_create_item(sdev, &sdev->fw_version,
+						      sizeof(sdev->fw_version),
+						      "fw_version");
+
+		if (ret < 0) {
+			dev_err(sdev->dev, "error: cannot create debugfs for fw_version\n");
+			return ret;
+		}
+	}
+
+	/* perform pre fw run operations */
+	ret = snd_sof_dsp_pre_fw_run(sdev);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: failed pre fw run op\n");
+		return ret;
+	}
+
+	dev_dbg(sdev->dev, "booting DSP firmware\n");
+
+	/* boot the firmware on the DSP */
+	ret = snd_sof_dsp_run(sdev);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: failed to reset DSP\n");
+		return ret;
+	}
+
+	/* now wait for the DSP to boot */
+	ret = wait_event_timeout(sdev->boot_wait, sdev->boot_complete,
+				 msecs_to_jiffies(sdev->boot_timeout));
+	if (ret == 0) {
+		dev_err(sdev->dev, "error: firmware boot timeout\n");
+		snd_sof_dsp_dbg_dump(sdev, SOF_DBG_REGS | SOF_DBG_MBOX |
+			SOF_DBG_TEXT | SOF_DBG_PCI);
+		return -EIO;
+	}
+
+	dev_info(sdev->dev, "firmware boot complete\n");
+
+	/* perform post fw run operations */
+	ret = snd_sof_dsp_post_fw_run(sdev);
+	if (ret < 0) {
+		dev_err(sdev->dev, "error: failed post fw run op\n");
+		return ret;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(snd_sof_run_firmware);
+
+void snd_sof_fw_unload(struct snd_sof_dev *sdev)
+{
+	/* TODO: support module unloading at runtime */
+}
+EXPORT_SYMBOL(snd_sof_fw_unload);
-- 
2.17.1