All of lore.kernel.org
 help / color / mirror / Atom feed
From: Alex Elder <elder@ieee.org>
To: Sireesh Kodali <sireeshkodali1@gmail.com>,
	phone-devel@vger.kernel.org,
	~postmarketos/upstreaming@lists.sr.ht, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org,
	elder@kernel.org
Cc: "David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>
Subject: Re: [RFC PATCH 09/17] net: ipa: Add support for using BAM as a DMA transport
Date: Wed, 13 Oct 2021 17:30:06 -0500	[thread overview]
Message-ID: <794ca401-294f-54ed-0af2-e3c763723340@ieee.org> (raw)
In-Reply-To: <20210920030811.57273-10-sireeshkodali1@gmail.com>

On 9/19/21 10:08 PM, Sireesh Kodali wrote:
> BAM is used on IPA v2.x. Since BAM already has a nice dmaengine driver,
> the IPA driver only makes calls the dmaengine API.
> Also add BAM transaction support to IPA's trasaction abstraction layer.
> 
> BAM transactions should use NAPI just like GSI transactions, but just
> use callbacks on each transaction for now.

This is where things get a little more complicated.  I'm not really
familiar with the BAM interface and would really like to give this
a much deeper review, and I won't be doing that now.

At first glance, it looks reasonably clean to me, and it surprises
me a little that this different system can be used with a relatively
small amount of change.  Much looks duplicated, so it could be that
a little more work abstracting might avoid that (but I haven't looked
that closely).

					-Alex

> 
> Signed-off-by: Sireesh Kodali <sireeshkodali1@gmail.com>
> ---
>   drivers/net/ipa/Makefile          |   2 +-
>   drivers/net/ipa/bam.c             | 525 ++++++++++++++++++++++++++++++
>   drivers/net/ipa/gsi.c             |   1 +
>   drivers/net/ipa/ipa_data.h        |   1 +
>   drivers/net/ipa/ipa_dma.h         |  18 +-
>   drivers/net/ipa/ipa_dma_private.h |   2 +
>   drivers/net/ipa/ipa_main.c        |  20 +-
>   drivers/net/ipa/ipa_trans.c       |  14 +-
>   drivers/net/ipa/ipa_trans.h       |   4 +
>   9 files changed, 569 insertions(+), 18 deletions(-)
>   create mode 100644 drivers/net/ipa/bam.c
> 
> diff --git a/drivers/net/ipa/Makefile b/drivers/net/ipa/Makefile
> index 3cd021fb992e..4abebc667f77 100644
> --- a/drivers/net/ipa/Makefile
> +++ b/drivers/net/ipa/Makefile
> @@ -2,7 +2,7 @@ obj-$(CONFIG_QCOM_IPA)	+=	ipa.o
>   
>   ipa-y			:=	ipa_main.o ipa_power.o ipa_reg.o ipa_mem.o \
>   				ipa_table.o ipa_interrupt.o gsi.o ipa_trans.o \
> -				ipa_gsi.o ipa_smp2p.o ipa_uc.o \
> +				ipa_gsi.o ipa_smp2p.o ipa_uc.o bam.o \
>   				ipa_endpoint.o ipa_cmd.o ipa_modem.o \
>   				ipa_resource.o ipa_qmi.o ipa_qmi_msg.o \
>   				ipa_sysfs.o
> diff --git a/drivers/net/ipa/bam.c b/drivers/net/ipa/bam.c
> new file mode 100644
> index 000000000000..0726e385fee5
> --- /dev/null
> +++ b/drivers/net/ipa/bam.c
> @@ -0,0 +1,525 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +/* Copyright (c) 2020, The Linux Foundation. All rights reserved.
> + */
> +
> +#include <linux/completion.h>
> +#include <linux/dma-mapping.h>
> +#include <linux/dmaengine.h>
> +#include <linux/interrupt.h>
> +#include <linux/io.h>
> +#include <linux/kernel.h>
> +#include <linux/mutex.h>
> +#include <linux/netdevice.h>
> +#include <linux/platform_device.h>
> +
> +#include "ipa_gsi.h"
> +#include "ipa.h"
> +#include "ipa_dma.h"
> +#include "ipa_dma_private.h"
> +#include "ipa_gsi.h"
> +#include "ipa_trans.h"
> +#include "ipa_data.h"
> +
> +/**
> + * DOC: The IPA Smart Peripheral System Interface
> + *
> + * The Smart Peripheral System is a means to communicate over BAM pipes to
> + * the IPA block. The Modem also uses BAM pipes to communicate with the IPA
> + * core.
> + *
> + * Refer the GSI documentation, because BAM is a precursor to GSI and more or less
> + * the same, conceptually (maybe, IDK, I have no docs to go through).
> + *
> + * Each channel here corresponds to 1 BAM pipe configured in BAM2BAM mode
> + *
> + * IPA cmds are transferred one at a time, each in one BAM transfer.
> + */
> +
> +/* Get and configure the BAM DMA channel */
> +int bam_channel_init_one(struct ipa_dma *bam,
> +			 const struct ipa_gsi_endpoint_data *data, bool command)
> +{
> +	struct dma_slave_config bam_config;
> +	u32 channel_id = data->channel_id;
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +	int ret;
> +
> +	/*TODO: if (!bam_channel_data_valid(bam, data))
> +		return -EINVAL;*/
> +
> +	channel->dma_subsys = bam;
> +	channel->dma_chan = dma_request_chan(bam->dev, data->channel_name);
> +	channel->toward_ipa = data->toward_ipa;
> +	channel->tlv_count = data->channel.tlv_count;
> +	channel->tre_count = data->channel.tre_count;
> +	if (IS_ERR(channel->dma_chan)) {
> +		dev_err(bam->dev, "failed to request BAM channel %s: %d\n",
> +				data->channel_name,
> +				(int) PTR_ERR(channel->dma_chan));
> +		return PTR_ERR(channel->dma_chan);
> +	}
> +
> +	ret = ipa_channel_trans_init(bam, data->channel_id);
> +	if (ret)
> +		goto err_dma_chan_free;
> +
> +	if (data->toward_ipa) {
> +		bam_config.direction = DMA_MEM_TO_DEV;
> +		bam_config.dst_maxburst = channel->tlv_count;
> +	} else {
> +		bam_config.direction = DMA_DEV_TO_MEM;
> +		bam_config.src_maxburst = channel->tlv_count;
> +	}
> +
> +	dmaengine_slave_config(channel->dma_chan, &bam_config);
> +
> +	if (command)
> +		ret = ipa_cmd_pool_init(channel, 256);
> +
> +	if (!ret)
> +		return 0;
> +
> +err_dma_chan_free:
> +	dma_release_channel(channel->dma_chan);
> +	return ret;
> +}
> +
> +static void bam_channel_exit_one(struct ipa_channel *channel)
> +{
> +	if (channel->dma_chan) {
> +		dmaengine_terminate_sync(channel->dma_chan);
> +		dma_release_channel(channel->dma_chan);
> +	}
> +}
> +
> +/* Get channels from BAM_DMA */
> +int bam_channel_init(struct ipa_dma *bam, u32 count,
> +		const struct ipa_gsi_endpoint_data *data)
> +{
> +	int ret = 0;
> +	u32 i;
> +
> +	for (i = 0; i < count; ++i) {
> +		bool command = i == IPA_ENDPOINT_AP_COMMAND_TX;
> +
> +		if (!data[i].channel_name || data[i].ee_id == GSI_EE_MODEM)
> +			continue;
> +
> +		ret = bam_channel_init_one(bam, &data[i], command);
> +		if (ret)
> +			goto err_unwind;
> +	}
> +
> +	return ret;
> +
> +err_unwind:
> +	while (i--) {
> +		if (ipa_gsi_endpoint_data_empty(&data[i]))
> +			continue;
> +
> +		bam_channel_exit_one(&bam->channel[i]);
> +	}
> +	return ret;
> +}
> +
> +/* Inverse of bam_channel_init() */
> +void bam_channel_exit(struct ipa_dma *bam)
> +{
> +	u32 channel_id = BAM_CHANNEL_COUNT_MAX - 1;
> +
> +	do
> +		bam_channel_exit_one(&bam->channel[channel_id]);
> +	while (channel_id--);
> +}
> +
> +/* Inverse of bam_init() */
> +static void bam_exit(struct ipa_dma *bam)
> +{
> +	mutex_destroy(&bam->mutex);
> +	bam_channel_exit(bam);
> +}
> +
> +/* Return the channel id associated with a given channel */
> +static u32 bam_channel_id(struct ipa_channel *channel)
> +{
> +	return channel - &channel->dma_subsys->channel[0];
> +}
> +
> +static void
> +bam_channel_tx_update(struct ipa_channel *channel, struct ipa_trans *trans)
> +{
> +	u64 byte_count = trans->byte_count + trans->len;
> +	u64 trans_count = trans->trans_count + 1;
> +
> +	byte_count -= channel->compl_byte_count;
> +	channel->compl_byte_count += byte_count;
> +	trans_count -= channel->compl_trans_count;
> +	channel->compl_trans_count += trans_count;
> +
> +	ipa_gsi_channel_tx_completed(channel->dma_subsys, bam_channel_id(channel),
> +					   trans_count, byte_count);
> +}
> +
> +static void
> +bam_channel_rx_update(struct ipa_channel *channel, struct ipa_trans *trans)
> +{
> +	/* FIXME */
> +	u64 byte_count = trans->byte_count + trans->len;
> +
> +	channel->byte_count += byte_count;
> +	channel->trans_count++;
> +}
> +
> +/* Consult hardware, move any newly completed transactions to completed list */
> +static void bam_channel_update(struct ipa_channel *channel)
> +{
> +	struct ipa_trans *trans;
> +
> +	list_for_each_entry(trans, &channel->trans_info.pending, links) {
> +		enum dma_status trans_status =
> +				dma_async_is_tx_complete(channel->dma_chan,
> +					trans->cookie, NULL, NULL);
> +		if (trans_status == DMA_COMPLETE)
> +			break;
> +	}
> +	/* Get the transaction for the latest completed event.  Take a
> +	 * reference to keep it from completing before we give the events
> +	 * for this and previous transactions back to the hardware.
> +	 */
> +	refcount_inc(&trans->refcount);
> +
> +	/* For RX channels, update each completed transaction with the number
> +	 * of bytes that were actually received.  For TX channels, report
> +	 * the number of transactions and bytes this completion represents
> +	 * up the network stack.
> +	 */
> +	if (channel->toward_ipa)
> +		bam_channel_tx_update(channel, trans);
> +	else
> +		bam_channel_rx_update(channel, trans);
> +
> +	ipa_trans_move_complete(trans);
> +
> +	ipa_trans_free(trans);
> +}
> +
> +/**
> + * bam_channel_poll_one() - Return a single completed transaction on a channel
> + * @channel:	Channel to be polled
> + *
> + * Return:	Transaction pointer, or null if none are available
> + *
> + * This function returns the first entry on a channel's completed transaction
> + * list.  If that list is empty, the hardware is consulted to determine
> + * whether any new transactions have completed.  If so, they're moved to the
> + * completed list and the new first entry is returned.  If there are no more
> + * completed transactions, a null pointer is returned.
> + */
> +static struct ipa_trans *bam_channel_poll_one(struct ipa_channel *channel)
> +{
> +	struct ipa_trans *trans;
> +
> +	/* Get the first transaction from the completed list */
> +	trans = ipa_channel_trans_complete(channel);
> +	if (!trans) {
> +		bam_channel_update(channel);
> +		trans = ipa_channel_trans_complete(channel);
> +	}
> +
> +	if (trans)
> +		ipa_trans_move_polled(trans);
> +
> +	return trans;
> +}
> +
> +/**
> + * bam_channel_poll() - NAPI poll function for a channel
> + * @napi:	NAPI structure for the channel
> + * @budget:	Budget supplied by NAPI core
> + *
> + * Return:	Number of items polled (<= budget)
> + *
> + * Single transactions completed by hardware are polled until either
> + * the budget is exhausted, or there are no more.  Each transaction
> + * polled is passed to ipa_trans_complete(), to perform remaining
> + * completion processing and retire/free the transaction.
> + */
> +static int bam_channel_poll(struct napi_struct *napi, int budget)
> +{
> +	struct ipa_channel *channel;
> +	int count = 0;
> +
> +	channel = container_of(napi, struct ipa_channel, napi);
> +	while (count < budget) {
> +		struct ipa_trans *trans;
> +
> +		count++;
> +		trans = bam_channel_poll_one(channel);
> +		if (!trans)
> +			break;
> +		ipa_trans_complete(trans);
> +	}
> +
> +	if (count < budget)
> +		napi_complete(&channel->napi);
> +
> +	return count;
> +}
> +
> +/* Setup function for a single channel */
> +static void bam_channel_setup_one(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	if (!channel->dma_subsys)
> +		return;	/* Ignore uninitialized channels */
> +
> +	if (channel->toward_ipa) {
> +		netif_tx_napi_add(&bam->dummy_dev, &channel->napi,
> +				  bam_channel_poll, NAPI_POLL_WEIGHT);
> +	} else {
> +		netif_napi_add(&bam->dummy_dev, &channel->napi,
> +			       bam_channel_poll, NAPI_POLL_WEIGHT);
> +	}
> +	napi_enable(&channel->napi);
> +}
> +
> +static void bam_channel_teardown_one(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	if (!channel->dma_subsys)
> +		return;		/* Ignore uninitialized channels */
> +
> +	netif_napi_del(&channel->napi);
> +}
> +
> +/* Setup function for channels */
> +static int bam_channel_setup(struct ipa_dma *bam)
> +{
> +	u32 channel_id = 0;
> +	int ret;
> +
> +	mutex_lock(&bam->mutex);
> +
> +	do
> +		bam_channel_setup_one(bam, channel_id);
> +	while (++channel_id < BAM_CHANNEL_COUNT_MAX);
> +
> +	/* Make sure no channels were defined that hardware does not support */
> +	while (channel_id < BAM_CHANNEL_COUNT_MAX) {
> +		struct ipa_channel *channel = &bam->channel[channel_id++];
> +
> +		if (!channel->dma_subsys)
> +			continue;	/* Ignore uninitialized channels */
> +
> +		dev_err(bam->dev, "channel %u not supported by hardware\n",
> +			channel_id - 1);
> +		channel_id = BAM_CHANNEL_COUNT_MAX;
> +		goto err_unwind;
> +	}
> +
> +	mutex_unlock(&bam->mutex);
> +
> +	return 0;
> +
> +err_unwind:
> +	while (channel_id--)
> +		bam_channel_teardown_one(bam, channel_id);
> +
> +	mutex_unlock(&bam->mutex);
> +
> +	return ret;
> +}
> +
> +/* Inverse of bam_channel_setup() */
> +static void bam_channel_teardown(struct ipa_dma *bam)
> +{
> +	u32 channel_id;
> +
> +	mutex_lock(&bam->mutex);
> +
> +	channel_id = BAM_CHANNEL_COUNT_MAX;
> +	do
> +		bam_channel_teardown_one(bam, channel_id);
> +	while (channel_id--);
> +
> +	mutex_unlock(&bam->mutex);
> +}
> +
> +static int bam_setup(struct ipa_dma *bam)
> +{
> +	return bam_channel_setup(bam);
> +}
> +
> +static void bam_teardown(struct ipa_dma *bam)
> +{
> +	bam_channel_teardown(bam);
> +}
> +
> +static u32 bam_channel_tre_max(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	/* Hardware limit is channel->tre_count - 1 */
> +	return channel->tre_count - (channel->tlv_count - 1);
> +}
> +
> +static u32 bam_channel_trans_tre_max(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	return channel->tlv_count;
> +}
> +
> +static int bam_channel_start(struct ipa_dma *bam, u32 channel_id)
> +{
> +	return 0;
> +}
> +
> +static int bam_channel_stop(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	return dmaengine_terminate_sync(channel->dma_chan);
> +}
> +
> +static void bam_channel_reset(struct ipa_dma *bam, u32 channel_id, bool doorbell)
> +{
> +	bam_channel_stop(bam, channel_id);
> +}
> +
> +static int bam_channel_suspend(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	return dmaengine_pause(channel->dma_chan);
> +}
> +
> +static int bam_channel_resume(struct ipa_dma *bam, u32 channel_id)
> +{
> +	struct ipa_channel *channel = &bam->channel[channel_id];
> +
> +	return dmaengine_resume(channel->dma_chan);
> +}
> +
> +static void bam_suspend(struct ipa_dma *bam)
> +{
> +	/* No-op for now */
> +}
> +
> +static void bam_resume(struct ipa_dma *bam)
> +{
> +	/* No-op for now */
> +}
> +
> +static void bam_trans_callback(void *arg)
> +{
> +	ipa_trans_complete(arg);
> +}
> +
> +static void bam_trans_commit(struct ipa_trans *trans, bool unused)
> +{
> +	struct ipa_channel *channel = &trans->dma_subsys->channel[trans->channel_id];
> +	enum ipa_cmd_opcode opcode = IPA_CMD_NONE;
> +	struct ipa_cmd_info *info;
> +	struct scatterlist *sg;
> +	u32 byte_count = 0;
> +	u32 i;
> +	enum dma_transfer_direction direction;
> +
> +	if (channel->toward_ipa)
> +		direction = DMA_MEM_TO_DEV;
> +	else
> +		direction = DMA_DEV_TO_MEM;
> +
> +	/* assert(trans->used > 0); */
> +
> +	info = trans->info ? &trans->info[0] : NULL;
> +	for_each_sg(trans->sgl, sg, trans->used, i) {
> +		bool last_tre = i == trans->used - 1;
> +		dma_addr_t addr = sg_dma_address(sg);
> +		u32 len = sg_dma_len(sg);
> +		u32 dma_flags = 0;
> +		struct dma_async_tx_descriptor *desc;
> +
> +		byte_count += len;
> +		if (info)
> +			opcode = info++->opcode;
> +
> +		if (opcode != IPA_CMD_NONE) {
> +			len = opcode;
> +			dma_flags |= DMA_PREP_IMM_CMD;
> +		}
> +
> +		if (last_tre)
> +			dma_flags |= DMA_PREP_INTERRUPT;
> +
> +		desc = dmaengine_prep_slave_single(channel->dma_chan, addr, len,
> +				direction, dma_flags);
> +
> +		if (last_tre) {
> +			desc->callback = bam_trans_callback;
> +			desc->callback_param = trans;
> +		}
> +
> +		desc->cookie = dmaengine_submit(desc);
> +
> +		if (last_tre)
> +			trans->cookie = desc->cookie;
> +
> +		if (direction == DMA_DEV_TO_MEM)
> +			dmaengine_desc_attach_metadata(desc, &trans->len, sizeof(trans->len));
> +	}
> +
> +	if (channel->toward_ipa) {
> +		/* We record TX bytes when they are sent */
> +		trans->len = byte_count;
> +		trans->trans_count = channel->trans_count;
> +		trans->byte_count = channel->byte_count;
> +		channel->trans_count++;
> +		channel->byte_count += byte_count;
> +	}
> +
> +	ipa_trans_move_pending(trans);
> +
> +	dma_async_issue_pending(channel->dma_chan);
> +}
> +
> +/* Initialize the BAM DMA channels
> + * Actual hw init is handled by the BAM_DMA driver
> + */
> +int bam_init(struct ipa_dma *bam, struct platform_device *pdev,
> +		enum ipa_version version, u32 count,
> +		const struct ipa_gsi_endpoint_data *data)
> +{
> +	struct device *dev = &pdev->dev;
> +	int ret;
> +
> +	bam->dev = dev;
> +	bam->version = version;
> +	bam->setup = bam_setup;
> +	bam->teardown = bam_teardown;
> +	bam->exit = bam_exit;
> +	bam->suspend = bam_suspend;
> +	bam->resume = bam_resume;
> +	bam->channel_tre_max = bam_channel_tre_max;
> +	bam->channel_trans_tre_max = bam_channel_trans_tre_max;
> +	bam->channel_start = bam_channel_start;
> +	bam->channel_stop = bam_channel_stop;
> +	bam->channel_reset = bam_channel_reset;
> +	bam->channel_suspend = bam_channel_suspend;
> +	bam->channel_resume = bam_channel_resume;
> +	bam->trans_commit = bam_trans_commit;
> +
> +	init_dummy_netdev(&bam->dummy_dev);
> +
> +	ret = bam_channel_init(bam, count, data);
> +	if (ret)
> +		return ret;
> +
> +	mutex_init(&bam->mutex);
> +
> +	return 0;
> +}
> diff --git a/drivers/net/ipa/gsi.c b/drivers/net/ipa/gsi.c
> index 39d9ca620a9f..ac0b9e748fa1 100644
> --- a/drivers/net/ipa/gsi.c
> +++ b/drivers/net/ipa/gsi.c
> @@ -2210,6 +2210,7 @@ int gsi_init(struct ipa_dma *gsi, struct platform_device *pdev,
>   	gsi->channel_reset = gsi_channel_reset;
>   	gsi->channel_suspend = gsi_channel_suspend;
>   	gsi->channel_resume = gsi_channel_resume;
> +	gsi->trans_commit = gsi_trans_commit;
>   
>   	/* GSI uses NAPI on all channels.  Create a dummy network device
>   	 * for the channel NAPI contexts to be associated with.
> diff --git a/drivers/net/ipa/ipa_data.h b/drivers/net/ipa/ipa_data.h
> index 6d329e9ce5d2..7d62d49f414f 100644
> --- a/drivers/net/ipa/ipa_data.h
> +++ b/drivers/net/ipa/ipa_data.h
> @@ -188,6 +188,7 @@ struct ipa_gsi_endpoint_data {
>   	u8 channel_id;
>   	u8 endpoint_id;
>   	bool toward_ipa;
> +	const char *channel_name;	/* used only for BAM DMA channels */
>   
>   	struct gsi_channel_data channel;
>   	struct ipa_endpoint_data endpoint;
> diff --git a/drivers/net/ipa/ipa_dma.h b/drivers/net/ipa/ipa_dma.h
> index 1a23e6ac5785..3000182ae689 100644
> --- a/drivers/net/ipa/ipa_dma.h
> +++ b/drivers/net/ipa/ipa_dma.h
> @@ -17,7 +17,11 @@
>   
>   /* Maximum number of channels and event rings supported by the driver */
>   #define GSI_CHANNEL_COUNT_MAX	23
> +#define BAM_CHANNEL_COUNT_MAX	20
>   #define GSI_EVT_RING_COUNT_MAX	24
> +#define IPA_CHANNEL_COUNT_MAX	MAX(GSI_CHANNEL_COUNT_MAX, \
> +				    BAM_CHANNEL_COUNT_MAX)
> +#define MAX(a, b)		((a > b) ? a : b)
>   
>   /* Maximum TLV FIFO size for a channel; 64 here is arbitrary (and high) */
>   #define GSI_TLV_MAX		64
> @@ -119,6 +123,8 @@ struct ipa_channel {
>   	struct gsi_ring tre_ring;
>   	u32 evt_ring_id;
>   
> +	struct dma_chan *dma_chan;
> +
>   	u64 byte_count;			/* total # bytes transferred */
>   	u64 trans_count;		/* total # transactions */
>   	/* The following counts are used only for TX endpoints */
> @@ -154,7 +160,7 @@ struct ipa_dma {
>   	u32 irq;
>   	u32 channel_count;
>   	u32 evt_ring_count;
> -	struct ipa_channel channel[GSI_CHANNEL_COUNT_MAX];
> +	struct ipa_channel channel[IPA_CHANNEL_COUNT_MAX];
>   	struct gsi_evt_ring evt_ring[GSI_EVT_RING_COUNT_MAX];
>   	u32 event_bitmap;		/* allocated event rings */
>   	u32 modem_channel_bitmap;	/* modem channels to allocate */
> @@ -303,7 +309,7 @@ static inline void ipa_dma_resume(struct ipa_dma *dma_subsys)
>   }
>   
>   /**
> - * ipa_dma_init() - Initialize the GSI subsystem
> + * ipa_init/bam_init() - Initialize the GSI/BAM subsystem
>    * @dma_subsys:	Address of ipa_dma structure embedded in an IPA structure
>    * @pdev:	IPA platform device
>    * @version:	IPA hardware version (implies GSI version)
> @@ -312,14 +318,18 @@ static inline void ipa_dma_resume(struct ipa_dma *dma_subsys)
>    *
>    * Return:	0 if successful, or a negative error code
>    *
> - * Early stage initialization of the GSI subsystem, performing tasks
> - * that can be done before the GSI hardware is ready to use.
> + * Early stage initialization of the GSI/BAM subsystem, performing tasks
> + * that can be done before the GSI/BAM hardware is ready to use.
>    */
>   
>   int gsi_init(struct ipa_dma *dma_subsys, struct platform_device *pdev,
>   	     enum ipa_version version, u32 count,
>   	     const struct ipa_gsi_endpoint_data *data);
>   
> +int bam_init(struct ipa_dma *dma_subsys, struct platform_device *pdev,
> +	     enum ipa_version version, u32 count,
> +	     const struct ipa_gsi_endpoint_data *data);
> +
>   /**
>    * ipa_dma_exit() - Exit the DMA subsystem
>    * @dma_subsys:	ipa_dma address previously passed to a successful gsi_init() call
> diff --git a/drivers/net/ipa/ipa_dma_private.h b/drivers/net/ipa/ipa_dma_private.h
> index 40148a551b47..1db53e597a61 100644
> --- a/drivers/net/ipa/ipa_dma_private.h
> +++ b/drivers/net/ipa/ipa_dma_private.h
> @@ -16,6 +16,8 @@ struct ipa_channel;
>   
>   #define GSI_RING_ELEMENT_SIZE	16	/* bytes; must be a power of 2 */
>   
> +void gsi_trans_commit(struct ipa_trans *trans, bool ring_db);
> +
>   /* Return the entry that follows one provided in a transaction pool */
>   void *ipa_trans_pool_next(struct ipa_trans_pool *pool, void *element);
>   
> diff --git a/drivers/net/ipa/ipa_main.c b/drivers/net/ipa/ipa_main.c
> index ba06e3ad554c..ea6c4347f2c6 100644
> --- a/drivers/net/ipa/ipa_main.c
> +++ b/drivers/net/ipa/ipa_main.c
> @@ -60,12 +60,15 @@
>    * core.  The GSI implements a set of "channels" used for communication
>    * between the AP and the IPA.
>    *
> - * The IPA layer uses GSI channels to implement its "endpoints".  And while
> - * a GSI channel carries data between the AP and the IPA, a pair of IPA
> - * endpoints is used to carry traffic between two EEs.  Specifically, the main
> - * modem network interface is implemented by two pairs of endpoints:  a TX
> + * The IPA layer uses GSI channels or BAM pipes to implement its "endpoints".
> + * And while a GSI channel carries data between the AP and the IPA, a pair of
> + * IPA endpoints is used to carry traffic between two EEs.  Specifically, the
> + * main modem network interface is implemented by two pairs of endpoints:  a TX
>    * endpoint on the AP coupled with an RX endpoint on the modem; and another
>    * RX endpoint on the AP receiving data from a TX endpoint on the modem.
> + *
> + * For BAM based transport, a pair of BAM pipes are used for TX and RX between
> + * the AP and IPA, and between IPA and other EEs.
>    */
>   
>   /* The name of the GSI firmware file relative to /lib/firmware */
> @@ -716,8 +719,13 @@ static int ipa_probe(struct platform_device *pdev)
>   	if (ret)
>   		goto err_reg_exit;
>   
> -	ret = gsi_init(&ipa->dma_subsys, pdev, ipa->version, data->endpoint_count,
> -		       data->endpoint_data);
> +	if (IPA_HAS_GSI(ipa->version))
> +		ret = gsi_init(&ipa->dma_subsys, pdev, ipa->version, data->endpoint_count,
> +			       data->endpoint_data);
> +	else
> +		ret = bam_init(&ipa->dma_subsys, pdev, ipa->version, data->endpoint_count,
> +			       data->endpoint_data);
> +
>   	if (ret)
>   		goto err_mem_exit;
>   
> diff --git a/drivers/net/ipa/ipa_trans.c b/drivers/net/ipa/ipa_trans.c
> index 22755f3ce3da..444f44846da8 100644
> --- a/drivers/net/ipa/ipa_trans.c
> +++ b/drivers/net/ipa/ipa_trans.c
> @@ -254,7 +254,7 @@ struct ipa_trans *ipa_channel_trans_complete(struct ipa_channel *channel)
>   }
>   
>   /* Move a transaction from the allocated list to the pending list */
> -static void ipa_trans_move_pending(struct ipa_trans *trans)
> +void ipa_trans_move_pending(struct ipa_trans *trans)
>   {
>   	struct ipa_channel *channel = &trans->dma_subsys->channel[trans->channel_id];
>   	struct ipa_trans_info *trans_info = &channel->trans_info;
> @@ -539,7 +539,7 @@ static void gsi_trans_tre_fill(struct gsi_tre *dest_tre, dma_addr_t addr,
>    * pending list.  Finally, updates the channel ring pointer and optionally
>    * rings the doorbell.
>    */
> -static void __gsi_trans_commit(struct ipa_trans *trans, bool ring_db)
> +void gsi_trans_commit(struct ipa_trans *trans, bool ring_db)
>   {
>   	struct ipa_channel *channel = &trans->dma_subsys->channel[trans->channel_id];
>   	struct gsi_ring *ring = &channel->tre_ring;
> @@ -604,9 +604,9 @@ static void __gsi_trans_commit(struct ipa_trans *trans, bool ring_db)
>   /* Commit a GSI transaction */
>   void ipa_trans_commit(struct ipa_trans *trans, bool ring_db)
>   {
> -	if (trans->used)
> -		__gsi_trans_commit(trans, ring_db);
> -	else
> +	if (trans->used) {
> +		trans->dma_subsys->trans_commit(trans, ring_db);
> +	} else
>   		ipa_trans_free(trans);
>   }
>   
> @@ -618,7 +618,7 @@ void ipa_trans_commit_wait(struct ipa_trans *trans)
>   
>   	refcount_inc(&trans->refcount);
>   
> -	__gsi_trans_commit(trans, true);
> +	trans->dma_subsys->trans_commit(trans, true);
>   
>   	wait_for_completion(&trans->completion);
>   
> @@ -638,7 +638,7 @@ int ipa_trans_commit_wait_timeout(struct ipa_trans *trans,
>   
>   	refcount_inc(&trans->refcount);
>   
> -	__gsi_trans_commit(trans, true);
> +	trans->dma_subsys->trans_commit(trans, true);
>   
>   	remaining = wait_for_completion_timeout(&trans->completion,
>   						timeout_jiffies);
> diff --git a/drivers/net/ipa/ipa_trans.h b/drivers/net/ipa/ipa_trans.h
> index b93342414360..5f41e3e6f92a 100644
> --- a/drivers/net/ipa/ipa_trans.h
> +++ b/drivers/net/ipa/ipa_trans.h
> @@ -10,6 +10,7 @@
>   #include <linux/refcount.h>
>   #include <linux/completion.h>
>   #include <linux/dma-direction.h>
> +#include <linux/dmaengine.h>
>   
>   #include "ipa_cmd.h"
>   
> @@ -61,6 +62,7 @@ struct ipa_trans {
>   	struct scatterlist *sgl;
>   	struct ipa_cmd_info *info;	/* array of entries, or null */
>   	enum dma_data_direction direction;
> +	dma_cookie_t cookie;
>   
>   	refcount_t refcount;
>   	struct completion completion;
> @@ -149,6 +151,8 @@ struct ipa_trans *ipa_channel_trans_alloc(struct ipa_dma *dma_subsys, u32 channe
>    */
>   void ipa_trans_free(struct ipa_trans *trans);
>   
> +void ipa_trans_move_pending(struct ipa_trans *trans);
> +
>   /**
>    * ipa_trans_cmd_add() - Add an immediate command to a transaction
>    * @trans:	Transaction
> 


  parent reply	other threads:[~2021-10-13 22:30 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-20  3:07 [RFC PATCH 00/17] net: ipa: Add support for IPA v2.x Sireesh Kodali
2021-09-20  3:07 ` [RFC PATCH 01/17] net: ipa: Correct ipa_status_opcode enumeration Sireesh Kodali
2021-10-13 22:28   ` Alex Elder
2021-10-18 16:12     ` Sireesh Kodali
2021-09-20  3:07 ` [RFC PATCH 02/17] net: ipa: revert to IPA_TABLE_ENTRY_SIZE for 32-bit IPA support Sireesh Kodali
2021-10-13 22:28   ` Alex Elder
2021-10-18 16:16     ` Sireesh Kodali
2021-09-20  3:07 ` [RFC PATCH 03/17] net: ipa: Refactor GSI code Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-09-20  3:07 ` [RFC PATCH 04/17] net: ipa: Establish ipa_dma interface Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-10-18 16:45     ` Sireesh Kodali
2021-09-20  3:07 ` [RFC PATCH 05/17] net: ipa: Check interrupts for availability Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-09-20  3:08 ` [RFC PATCH 06/17] net: ipa: Add timeout for ipa_cmd_pipeline_clear_wait Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-10-18 17:02     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 07/17] net: ipa: Add IPA v2.x register definitions Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-10-18 17:25     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 08/17] net: ipa: Add support for IPA v2.x interrupts Sireesh Kodali
2021-10-13 22:29   ` Alex Elder
2021-09-20  3:08 ` [RFC PATCH 09/17] net: ipa: Add support for using BAM as a DMA transport Sireesh Kodali
2021-09-20 14:31   ` kernel test robot
2021-10-13 22:30   ` Alex Elder [this message]
2021-10-18 17:30     ` Sireesh Kodali
2021-09-20  3:08 ` [PATCH 10/17] net: ipa: Add support for IPA v2.x commands and table init Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-10-18 18:13     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 11/17] net: ipa: Add support for IPA v2.x endpoints Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-10-18 18:17     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 12/17] net: ipa: Add support for IPA v2.x memory map Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-10-18 18:19     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 13/17] net: ipa: Add support for IPA v2.x in the driver's QMI interface Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-10-18 18:22     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 14/17] net: ipa: Add support for IPA v2 microcontroller Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-09-20  3:08 ` [RFC PATCH 15/17] net: ipa: Add IPA v2.6L initialization sequence support Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-09-20  3:08 ` [RFC PATCH 16/17] net: ipa: Add hw config describing IPA v2.x hardware Sireesh Kodali
2021-10-13 22:30   ` Alex Elder
2021-10-18 18:35     ` Sireesh Kodali
2021-09-20  3:08 ` [RFC PATCH 17/17] dt-bindings: net: qcom,ipa: Add support for MSM8953 and MSM8996 IPA Sireesh Kodali
2021-09-23 12:42   ` Rob Herring
2021-10-13 22:31   ` Alex Elder
2021-10-13 22:27 ` [RFC PATCH 00/17] net: ipa: Add support for IPA v2.x Alex Elder

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=794ca401-294f-54ed-0af2-e3c763723340@ieee.org \
    --to=elder@ieee.org \
    --cc=davem@davemloft.net \
    --cc=elder@kernel.org \
    --cc=kuba@kernel.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=phone-devel@vger.kernel.org \
    --cc=sireeshkodali1@gmail.com \
    --cc=~postmarketos/upstreaming@lists.sr.ht \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.