From patchwork Tue Jun 21 07:18:34 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ohad Ben Cohen X-Patchwork-Id: 908152 Received: from devils.ext.ti.com (devils.ext.ti.com [198.47.26.153]) by demeter2.kernel.org (8.14.4/8.14.4) with ESMTP id p5N5Cu7P004895 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 23 Jun 2011 05:13:18 GMT Received: from dlep35.itg.ti.com ([157.170.170.118]) by devils.ext.ti.com (8.13.7/8.13.7) with ESMTP id p5N5Bkq3026303 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 23 Jun 2011 00:11:46 -0500 Received: from linux.omap.com (smtp-le.itg.ti.com [157.170.170.27]) by dlep35.itg.ti.com (8.13.7/8.13.8) with ESMTP id p5N5BkL5016786; Thu, 23 Jun 2011 00:11:46 -0500 (CDT) Received: from linux.omap.com (localhost [127.0.0.1]) by linux.omap.com (Postfix) with ESMTP id A959A8062C; Thu, 23 Jun 2011 00:11:45 -0500 (CDT) X-Original-To: davinci-linux-open-source@linux.davincidsp.com Delivered-To: davinci-linux-open-source@linux.davincidsp.com Received: from dflp53.itg.ti.com (dflp53.itg.ti.com [128.247.5.6]) by linux.omap.com (Postfix) with ESMTP id AFDE180626 for ; Tue, 21 Jun 2011 02:22:29 -0500 (CDT) Received: from red.ext.ti.com (red.ext.ti.com [192.94.93.37]) by dflp53.itg.ti.com (8.13.8/8.13.8) with ESMTP id p5L7MTg8022188 for ; Tue, 21 Jun 2011 02:22:29 -0500 (CDT) Received: from psmtp.com (na3sys009amx199.postini.com [74.125.149.235]) by red.ext.ti.com (8.13.7/8.13.7) with SMTP id p5L7MSAl008129 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Tue, 21 Jun 2011 02:22:28 -0500 Received: from mail-ww0-f47.google.com ([74.125.82.47]) (using TLSv1) by na3sys009amx199.postini.com ([74.125.148.10]) with SMTP; Tue, 21 Jun 2011 07:22:29 GMT Received: by mail-ww0-f47.google.com with SMTP id 27so2747957wwf.4 for ; Tue, 21 Jun 2011 00:22:28 -0700 (PDT) Received: by 10.227.199.21 with SMTP id eq21mr5900419wbb.101.1308640946573; Tue, 21 Jun 2011 00:22:26 -0700 (PDT) Received: from localhost.localdomain (46-116-119-47.bb.netvision.net.il [46.116.119.47]) by mx.google.com with ESMTPS id o19sm3767542wbh.38.2011.06.21.00.22.24 (version=TLSv1/SSLv3 cipher=OTHER); Tue, 21 Jun 2011 00:22:25 -0700 (PDT) From: Ohad Ben-Cohen To: , , Subject: [RFC 8/8] rpmsg: add omap host backend Date: Tue, 21 Jun 2011 10:18:34 +0300 Message-Id: <1308640714-17961-9-git-send-email-ohad@wizery.com> X-Mailer: git-send-email 1.7.1 In-Reply-To: <1308640714-17961-1-git-send-email-ohad@wizery.com> References: <1308640714-17961-1-git-send-email-ohad@wizery.com> X-pstn-neptune: 0/0/0.00/0 X-pstn-levels: (S:97.71072/99.90000 CV:99.9000 FC:95.5390 LC: 0.0000 R:95.9108 P:95.9108 M:97.0282 C:98.6951 ) X-pstn-settings: 2 (0.5000:0.0750) s cv GT3 gt2 gt1 r p m c X-pstn-addresses: from [35/1] X-Mailman-Approved-At: Thu, 23 Jun 2011 00:07:34 -0500 Cc: Ohad Ben-Cohen , davinci-linux-open-source , Arnd Bergmann , Brian Swetland , Rusty Russell , Grant Likely , akpm@linux-foundation.org X-BeenThere: davinci-linux-open-source@linux.davincidsp.com X-Mailman-Version: 2.1.12 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Sender: davinci-linux-open-source-bounces@linux.davincidsp.com Errors-To: davinci-linux-open-source-bounces@linux.davincidsp.com X-Greylist: Sender succeeded STARTTLS authentication, not delayed by milter-greylist-4.2.6 (demeter2.kernel.org [140.211.167.43]); Thu, 23 Jun 2011 05:13:18 +0000 (UTC) Add rpmsg devices that wires virtio to the OMAP mailbox and enabled A9<->M3 communications on OMAP4. Each M3 core gets its own virtio device, since it is running its own independent instance of RTOS, and is therefore treated as a completely separate processor by rpmsg. Both M3 cores use the same mailbox device to communicate with the A9, and the mailbox payload is used to indicate which of the virtqueues is triggered (each M3 core gets 2 virtqueues, for rx and tx). Map the two vrings and the IPC buffers as noncacheable; use dynamically carved-out physical memory for that. Put this omap implementation in a dedicated drivers/rpmsg/host folder, with the intention of stacking different rpmsg host implementations in one place. Signed-off-by: Ohad Ben-Cohen --- drivers/rpmsg/Kconfig | 2 + drivers/rpmsg/Makefile | 2 + drivers/rpmsg/host/Kconfig | 11 + drivers/rpmsg/host/Makefile | 1 + drivers/rpmsg/host/omap_rpmsg.c | 540 +++++++++++++++++++++++++++++++++++++++ drivers/rpmsg/host/omap_rpmsg.h | 69 +++++ 6 files changed, 625 insertions(+), 0 deletions(-) create mode 100644 drivers/rpmsg/host/Kconfig create mode 100644 drivers/rpmsg/host/Makefile create mode 100644 drivers/rpmsg/host/omap_rpmsg.c create mode 100644 drivers/rpmsg/host/omap_rpmsg.h diff --git a/drivers/rpmsg/Kconfig b/drivers/rpmsg/Kconfig index 41303f5..9544d62 100644 --- a/drivers/rpmsg/Kconfig +++ b/drivers/rpmsg/Kconfig @@ -12,3 +12,5 @@ if RPMSG source "drivers/virtio/Kconfig" endif # RPMSG + +source "drivers/rpmsg/host/Kconfig" diff --git a/drivers/rpmsg/Makefile b/drivers/rpmsg/Makefile index 7617fcb..368a526 100644 --- a/drivers/rpmsg/Makefile +++ b/drivers/rpmsg/Makefile @@ -1 +1,3 @@ obj-$(CONFIG_RPMSG) += virtio_rpmsg_bus.o + +obj-$(CONFIG_RPMSG) += host/ diff --git a/drivers/rpmsg/host/Kconfig b/drivers/rpmsg/host/Kconfig new file mode 100644 index 0000000..fd2140c --- /dev/null +++ b/drivers/rpmsg/host/Kconfig @@ -0,0 +1,11 @@ +config OMAP_RPMSG + tristate "OMAP virtio-based remote processor messaging support" + depends on ARCH_OMAP4 && OMAP_REMOTE_PROC + select CONFIG_OMAP_MBOX_FWK + select RPMSG + help + Say Y if you want to enable OMAP's virtio-based remote-processor + messaging, currently only available on OMAP4. This is required + for offloading cpu-intensive and/or latency-sensitive tasks to + the remote on-chip M3s or C64x+ dsp, usually used by multimedia + frameworks. diff --git a/drivers/rpmsg/host/Makefile b/drivers/rpmsg/host/Makefile new file mode 100644 index 0000000..9823745 --- /dev/null +++ b/drivers/rpmsg/host/Makefile @@ -0,0 +1 @@ +obj-$(CONFIG_OMAP_RPMSG) += omap_rpmsg.o diff --git a/drivers/rpmsg/host/omap_rpmsg.c b/drivers/rpmsg/host/omap_rpmsg.c new file mode 100644 index 0000000..b8c2305 --- /dev/null +++ b/drivers/rpmsg/host/omap_rpmsg.c @@ -0,0 +1,540 @@ +/* + * Remote processor messaging transport (OMAP platform-specific bits) + * + * Copyright (C) 2011 Texas Instruments, Inc. + * Copyright (C) 2011 Google, Inc. + * + * Ohad Ben-Cohen + * Brian Swetland + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#define pr_fmt(fmt) "%s: " fmt, __func__ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "omap_rpmsg.h" + +/** + * struct omap_rpmsg_vproc - omap's virtio remote processor state + * @vdev: virtio device + * @vring: phys address of two vrings; first one used for rx, 2nd one for tx + * @buf_paddr: physical address of the IPC buffer region + * @buf_size: size of IPC buffer region + * @buf_mapped: kernel (ioremap'ed) address of IPC buffer region + * @mbox_name: name of omap mailbox device to use with this vproc + * @rproc_name: name of remote proc device to use with this vproc + * @mbox: omap mailbox handle + * @rproc: remoteproc handle + * @nb: notifier block that will be invoked on inbound mailbox messages + * @vq: virtio's virtqueues + * @base_vq_id: index of first virtqueue that belongs to this vproc + * @num_of_vqs: number of virtqueues this vproc owns + * @static_chnls: table of static channels for this vproc + */ +struct omap_rpmsg_vproc { + struct virtio_device vdev; + unsigned int vring[2]; /* mpu owns first vring, ipu owns the 2nd */ + unsigned int buf_paddr; + unsigned int buf_size; /* size must be page-aligned */ + void *buf_mapped; + char *mbox_name; + char *rproc_name; + struct omap_mbox *mbox; + struct rproc *rproc; + struct notifier_block nb; + struct virtqueue *vq[2]; + int base_vq_id; + int num_of_vqs; + struct rpmsg_channel_info *static_chnls; +}; + +#define to_omap_vproc(vd) container_of(vd, struct omap_rpmsg_vproc, vdev) + +/** + * struct omap_rpmsg_vq_info - virtqueue state + * @num: number of buffers supported by the vring + * @vq_id: a unique index of this virtqueue + * @addr: address where the vring is mapped onto + * @vproc: the virtual remote processor state + * + * Such a struct will be maintained for every virtqueue we're + * using to communicate with the remote processor + */ +struct omap_rpmsg_vq_info { + __u16 num; + __u16 vq_id; + void *addr; + struct omap_rpmsg_vproc *vproc; +}; + +/* + * For now, allocate 256 buffers of 512 bytes for each side. each buffer + * will then have 16B for the msg header and 496B for the payload. + * This will require a total space of 256KB for the buffers themselves, and + * 3 pages for every vring (the size of the vring depends on the number of + * buffers it supports). + */ +#define RPMSG_NUM_BUFS (512) +#define RPMSG_BUF_SIZE (512) +#define RPMSG_BUFS_SPACE (RPMSG_NUM_BUFS * RPMSG_BUF_SIZE) + +/* + * The alignment between the consumer and producer parts of the vring. + * Note: this is part of the "wire" protocol. If you change this, you need + * to update your BIOS image as well + */ +#define RPMSG_VRING_ALIGN (4096) + +/* With 256 buffers, our vring will occupy 3 pages */ +#define RPMSG_RING_SIZE ((DIV_ROUND_UP(vring_size(RPMSG_NUM_BUFS / 2, \ + RPMSG_VRING_ALIGN), PAGE_SIZE)) * PAGE_SIZE) + +/* The total IPC space needed to communicate with a remote processor */ +#define RPMSG_IPC_MEM (RPMSG_BUFS_SPACE + 2 * RPMSG_RING_SIZE) + +/* + * Provide rpmsg core with platform-specific configuration. + * Since user data is at stake here, bugs can't be tolerated. hence + * the BUG_ON approach on invalid lengths. + * + * For more info on these configuration requests, see enum + * rpmsg_platform_requests. + */ +static void omap_rpmsg_get(struct virtio_device *vdev, unsigned int request, + void *buf, unsigned len) +{ + struct omap_rpmsg_vproc *vproc = to_omap_vproc(vdev); + int tmp; + + switch (request) { + case VPROC_BUF_ADDR: + BUG_ON(len != sizeof(vproc->buf_mapped)); + memcpy(buf, &vproc->buf_mapped, len); + break; + case VPROC_BUF_PADDR: + BUG_ON(len != sizeof(vproc->buf_paddr)); + memcpy(buf, &vproc->buf_paddr, len); + break; + case VPROC_BUF_NUM: + BUG_ON(len != sizeof(tmp)); + tmp = RPMSG_NUM_BUFS; + memcpy(buf, &tmp, len); + break; + case VPROC_BUF_SZ: + BUG_ON(len != sizeof(tmp)); + tmp = RPMSG_BUF_SIZE; + memcpy(buf, &tmp, len); + break; + case VPROC_STATIC_CHANNELS: + BUG_ON(len != sizeof(vproc->static_chnls)); + memcpy(buf, &vproc->static_chnls, len); + break; + default: + dev_err(&vdev->dev, "invalid request: %d\n", request); + } +} + +/* kick the remote processor, and let it know which virtqueue to poke at */ +static void omap_rpmsg_notify(struct virtqueue *vq) +{ + struct omap_rpmsg_vq_info *rpvq = vq->priv; + int ret; + + pr_debug("sending mailbox msg: %d\n", rpvq->vq_id); + /* send the index of the triggered virtqueue in the mailbox payload */ + ret = omap_mbox_msg_send(rpvq->vproc->mbox, rpvq->vq_id); + if (ret) + pr_err("ugh, omap_mbox_msg_send() failed: %d\n", ret); +} + +/** + * omap_rpmsg_mbox_callback() - inbound mailbox message handler + * @this: notifier block + * @index: unused + * @data: mailbox payload + * + * This handler is invoked by omap's mailbox driver whenever a mailbox + * message is received. Usually, the mailbox payload simply contains + * the index of the virtqueue that is kicked by the remote processor, + * and we let virtio handle it. + * + * In addition to virtqueue indices, we also have some out-of-band values + * that indicates different events. Those values are deliberately very + * big so they don't coincide with virtqueue indices. Moreover, + * they are rarely used, if used at all, and their necessity should + * be revisited. + */ +static int omap_rpmsg_mbox_callback(struct notifier_block *this, + unsigned long index, void *data) +{ + mbox_msg_t msg = (mbox_msg_t) data; + struct omap_rpmsg_vproc *vproc; + + vproc = container_of(this, struct omap_rpmsg_vproc, nb); + + pr_debug("mbox msg: 0x%x\n", msg); + + switch (msg) { + case RP_MBOX_CRASH: + pr_err("%s has just crashed !\n", vproc->rproc_name); + /* todo: smarter error handling here */ + break; + case RP_MBOX_ECHO_REPLY: + pr_info("received echo reply from %s !\n", vproc->rproc_name); + break; + case RP_MBOX_PENDING_MSG: + /* + * a new inbound message is waiting in our rx vring (1st vring). + * Let's pretend the message explicitly contained the rx vring + * index number and handle it generically. + */ + msg = vproc->base_vq_id; + /* intentional fall-through */ + default: + /* ignore vq indices which are clearly not for us */ + if (msg < vproc->base_vq_id) + break; + + msg -= vproc->base_vq_id; + + /* + * Currently both PENDING_MSG and explicit-virtqueue-index + * messaging are supported. + * Whatever approach is taken, at this point 'msg' contains + * the index of the vring which was just triggered. + */ + if (msg < vproc->num_of_vqs) + vring_interrupt(msg, vproc->vq[msg]); + } + + return NOTIFY_DONE; +} + +/* prepare a virtqueue */ +static struct virtqueue *rp_find_vq(struct virtio_device *vdev, + unsigned index, + void (*callback)(struct virtqueue *vq), + const char *name) +{ + struct omap_rpmsg_vproc *vproc = to_omap_vproc(vdev); + struct omap_rpmsg_vq_info *rpvq; + struct virtqueue *vq; + int err; + + rpvq = kmalloc(sizeof(*rpvq), GFP_KERNEL); + if (!rpvq) + return ERR_PTR(-ENOMEM); + + /* ioremap'ing normal memory, so we cast away sparse's complaints */ + rpvq->addr = (__force void *) ioremap_nocache(vproc->vring[index], + RPMSG_RING_SIZE); + if (!rpvq->addr) { + err = -ENOMEM; + goto free_rpvq; + } + + memset(rpvq->addr, 0, RPMSG_RING_SIZE); + + pr_debug("vring%d: phys 0x%x, virt 0x%x\n", index, vproc->vring[index], + (unsigned int) rpvq->addr); + + vq = vring_new_virtqueue(RPMSG_NUM_BUFS / 2, RPMSG_VRING_ALIGN, vdev, + rpvq->addr, omap_rpmsg_notify, callback, name); + if (!vq) { + pr_err("vring_new_virtqueue failed\n"); + err = -ENOMEM; + goto unmap_vring; + } + + vproc->vq[index] = vq; + vq->priv = rpvq; + /* unique id for this virtqueue */ + rpvq->vq_id = vproc->base_vq_id + index; + rpvq->vproc = vproc; + + return vq; + +unmap_vring: + /* iounmap normal memory, so make sparse happy */ + iounmap((__force void __iomem *) rpvq->addr); +free_rpvq: + kfree(rpvq); + return ERR_PTR(err); +} + +static void omap_rpmsg_del_vqs(struct virtio_device *vdev) +{ + struct virtqueue *vq, *n; + struct omap_rpmsg_vproc *vproc = to_omap_vproc(vdev); + + if (vproc->rproc) + rproc_put(vproc->rproc); + + if (vproc->mbox) + omap_mbox_put(vproc->mbox, &vproc->nb); + + if (vproc->buf_mapped) + /* iounmap normal memory, so make sparse happy */ + iounmap((__force void __iomem *)vproc->buf_mapped); + + list_for_each_entry_safe(vq, n, &vdev->vqs, list) { + struct omap_rpmsg_vq_info *rpvq = vq->priv; + vring_del_virtqueue(vq); + /* iounmap normal memory, so make sparse happy */ + iounmap((__force void __iomem *) rpvq->addr); + kfree(rpvq); + } +} + +static int omap_rpmsg_find_vqs(struct virtio_device *vdev, unsigned nvqs, + struct virtqueue *vqs[], + vq_callback_t *callbacks[], + const char *names[]) +{ + struct omap_rpmsg_vproc *vproc = to_omap_vproc(vdev); + int i, err; + + /* we maintain two virtqueues per remote processor (for RX and TX) */ + if (nvqs != 2) + return -EINVAL; + + for (i = 0; i < nvqs; ++i) { + vqs[i] = rp_find_vq(vdev, i, callbacks[i], names[i]); + if (IS_ERR(vqs[i])) { + err = PTR_ERR(vqs[i]); + goto error; + } + } + + vproc->num_of_vqs = nvqs; + + /* ioremap'ing normal memory, so we cast away sparse's complaints */ + vproc->buf_mapped = (__force void *) ioremap_nocache(vproc->buf_paddr, + vproc->buf_size); + if (!vproc->buf_mapped) { + pr_err("ioremap failed\n"); + err = -ENOMEM; + goto error; + } + + /* for now, use mailbox's notifiers. later that can be optimized */ + vproc->nb.notifier_call = omap_rpmsg_mbox_callback; + vproc->mbox = omap_mbox_get(vproc->mbox_name, &vproc->nb); + if (IS_ERR(vproc->mbox)) { + pr_err("failed to get mailbox %s\n", vproc->mbox_name); + err = -EINVAL; + goto error; + } + + pr_debug("buf: phys 0x%x, virt 0x%x\n", vproc->buf_paddr, + (unsigned int) vproc->buf_mapped); + + /* tell the M3 we're ready (so M3 will know we're sane) */ + err = omap_mbox_msg_send(vproc->mbox, RP_MBOX_READY); + if (err) { + pr_err("ugh, omap_mbox_msg_send() failed: %d\n", err); + goto error; + } + + /* send it the physical address of the vrings + IPC buffer */ + err = omap_mbox_msg_send(vproc->mbox, (mbox_msg_t) vproc->buf_paddr); + if (err) { + pr_err("ugh, omap_mbox_msg_send() failed: %d\n", err); + goto error; + } + + /* ping the remote processor. this is only for sanity-sake; + * there is no functional effect whatsoever */ + err = omap_mbox_msg_send(vproc->mbox, RP_MBOX_ECHO_REQUEST); + if (err) { + pr_err("ugh, omap_mbox_msg_send() failed: %d\n", err); + goto error; + } + + /* now load the firmware, and boot the M3 */ + vproc->rproc = rproc_get(vproc->rproc_name); + if (!vproc->rproc) { + pr_err("failed to get rproc %s\n", vproc->rproc_name); + err = -EINVAL; + goto error; + } + + return 0; + +error: + omap_rpmsg_del_vqs(vdev); + return err; +} + +/* + * should be nice to add firmware support for these handlers. + * for now provide them so virtio doesn't crash + */ +static u8 omap_rpmsg_get_status(struct virtio_device *vdev) +{ + return 0; +} + +static void omap_rpmsg_set_status(struct virtio_device *vdev, u8 status) +{ + dev_dbg(&vdev->dev, "new status: %d\n", status); +} + +static void omap_rpmsg_reset(struct virtio_device *vdev) +{ + dev_dbg(&vdev->dev, "reset !\n"); +} + +static u32 omap_rpmsg_get_features(struct virtio_device *vdev) +{ + /* for now, use hardcoded bitmap. later this should be provided + * by the firmware itself */ + return 1 << VIRTIO_RPMSG_F_NS; +} + +static void omap_rpmsg_finalize_features(struct virtio_device *vdev) +{ + /* Give virtio_ring a chance to accept features */ + vring_transport_features(vdev); +} + +static void omap_rpmsg_vproc_release(struct device *dev) +{ + /* this handler is provided so driver core doesn't yell at us */ +} + +static struct virtio_config_ops omap_rpmsg_config_ops = { + .get_features = omap_rpmsg_get_features, + .finalize_features = omap_rpmsg_finalize_features, + .get = omap_rpmsg_get, + .find_vqs = omap_rpmsg_find_vqs, + .del_vqs = omap_rpmsg_del_vqs, + .reset = omap_rpmsg_reset, + .set_status = omap_rpmsg_set_status, + .get_status = omap_rpmsg_get_status, +}; + +/* + * Populating the static channels table. + * + * This is not always required, and the example below just demonstrates + * how to populate it with a static server channel. + * + * This example should be moved to Documentation/rpmsh.h before merging. + * + * For more info, see 'struct rpmsg_channel_info'. + */ +static struct rpmsg_channel_info omap_ipuc0_static_chnls[] = { + RMSG_SERVER_CHNL("rpmsg-server-sample", 137), + { }, +}; + +static struct rpmsg_channel_info omap_ipuc1_static_chnls[] = { + { }, +}; + +static struct omap_rpmsg_vproc omap_rpmsg_vprocs[] = { + /* ipu_c0's rpmsg backend */ + { + .vdev.id.device = VIRTIO_ID_RPMSG, + .vdev.config = &omap_rpmsg_config_ops, + .mbox_name = "mailbox-1", + .rproc_name = "ipu", + /* core 0 is using indices 0 + 1 for its vqs */ + .base_vq_id = 0, + .static_chnls = omap_ipuc0_static_chnls, + }, + /* ipu_c1's rpmsg backend */ + { + .vdev.id.device = VIRTIO_ID_RPMSG, + .vdev.config = &omap_rpmsg_config_ops, + .mbox_name = "mailbox-1", + .rproc_name = "ipu", + /* core 1 is using indices 2 + 3 for its vqs */ + .base_vq_id = 2, + .static_chnls = omap_ipuc1_static_chnls, + }, +}; + +static int __init omap_rpmsg_ini(void) +{ + int i, ret = 0; + /* + * This whole area generally needs some rework. + * E.g, consider using dma_alloc_coherent for the IPC buffers and + * vrings, use CMA, etc... + */ + phys_addr_t paddr = omap_dsp_get_mempool_base(); + phys_addr_t psize = omap_dsp_get_mempool_size(); + + /* + * allocate carverout memory for the buffers and vring, and + * then register the vproc virtio device + */ + for (i = 0; i < ARRAY_SIZE(omap_rpmsg_vprocs); i++) { + struct omap_rpmsg_vproc *vproc = &omap_rpmsg_vprocs[i]; + + if (psize < RPMSG_IPC_MEM) { + pr_err("out of carveout memory: %d (%d)\n", psize, i); + return -ENOMEM; + } + + vproc->buf_paddr = paddr; + vproc->buf_size = RPMSG_BUFS_SPACE; + vproc->vring[0] = paddr + RPMSG_BUFS_SPACE; + vproc->vring[1] = paddr + RPMSG_BUFS_SPACE + RPMSG_RING_SIZE; + + paddr += RPMSG_IPC_MEM; + psize -= RPMSG_IPC_MEM; + + pr_debug("vproc%d: buf 0x%x, vring0 0x%x, vring1 0x%x\n", i, + vproc->buf_paddr, vproc->vring[0], vproc->vring[1]); + + vproc->vdev.dev.release = omap_rpmsg_vproc_release; + + ret = register_virtio_device(&vproc->vdev); + if (ret) { + pr_err("failed to register vproc: %d\n", ret); + break; + } + } + + return ret; +} +module_init(omap_rpmsg_ini); + +static void __exit omap_rpmsg_fini(void) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(omap_rpmsg_vprocs); i++) { + struct omap_rpmsg_vproc *vproc = &omap_rpmsg_vprocs[i]; + + unregister_virtio_device(&vproc->vdev); + } +} +module_exit(omap_rpmsg_fini); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("OMAP Remote processor messaging virtio device"); diff --git a/drivers/rpmsg/host/omap_rpmsg.h b/drivers/rpmsg/host/omap_rpmsg.h new file mode 100644 index 0000000..f6d2036 --- /dev/null +++ b/drivers/rpmsg/host/omap_rpmsg.h @@ -0,0 +1,69 @@ +/* + * Remote processor messaging + * + * Copyright (C) 2011 Texas Instruments, Inc. + * Copyright (C) 2011 Google, Inc. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in + * the documentation and/or other materials provided with the + * distribution. + * * Neither the name Texas Instruments nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#ifndef _OMAP_RPMSG_H +#define _OMAP_RPMSG_H + +/* + * enum - Predefined Mailbox Messages + * + * @RP_MBOX_READY: informs the M3's that we're up and running. this is + * part of the init sequence sent that the M3 expects to see immediately + * after it is booted. + * + * @RP_MBOX_PENDING_MSG: informs the receiver that there is an inbound + * message waiting in its own receive-side vring. please note that currently + * this message is optional: alternatively, one can explicitly send the index + * of the triggered virtqueue itself. the preferred approach will be decided + * as we progress and experiment with those two different approaches. + * + * @RP_MBOX_CRASH: this message is sent if BIOS crashes + * + * @RP_MBOX_ECHO_REQUEST: a mailbox-level "ping" message. + * + * @RP_MBOX_ECHO_REPLY: a mailbox-level reply to a "ping" + * + * @RP_MBOX_ABORT_REQUEST: a "please crash" request, used for testing the + * recovery mechanism (to some extent). + */ +enum omap_rp_mbox_messages { + RP_MBOX_READY = 0xFFFFFF00, + RP_MBOX_PENDING_MSG = 0xFFFFFF01, + RP_MBOX_CRASH = 0xFFFFFF02, + RP_MBOX_ECHO_REQUEST = 0xFFFFFF03, + RP_MBOX_ECHO_REPLY = 0xFFFFFF04, + RP_MBOX_ABORT_REQUEST = 0xFFFFFF05, +}; + +#endif /* _OMAP_RPMSG_H */