soc: qcom: Add Shared Memory Manager driver
authorBjorn Andersson <bjorn.andersson@sonymobile.com>
Fri, 26 Jun 2015 21:50:10 +0000 (14:50 -0700)
committerAndy Gross <agross@codeaurora.org>
Tue, 28 Jul 2015 20:50:16 +0000 (15:50 -0500)
The Shared Memory Manager driver implements an interface for allocating
and accessing items in the memory area shared among all of the
processors in a Qualcomm platform.

Signed-off-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Acked-by: Andy Gross <agross@codeaurora.org>
Signed-off-by: Andy Gross <agross@codeaurora.org>
drivers/soc/qcom/Kconfig
drivers/soc/qcom/Makefile
drivers/soc/qcom/smem.c [new file with mode: 0644]
include/linux/soc/qcom/smem.h [new file with mode: 0644]

index 5eea374c8fa621ac91056fdd9b1853dd772d358c..8544e1594c2cb0c0bbf34308d62200d57f550b38 100644 (file)
@@ -17,3 +17,11 @@ config QCOM_PM
          QCOM Platform specific power driver to manage cores and L2 low power
          modes. It interface with various system drivers to put the cores in
          low power modes.
+
+config QCOM_SMEM
+       tristate "Qualcomm Shared Memory Manager (SMEM)"
+       depends on ARCH_QCOM
+       help
+         Say y here to enable support for the Qualcomm Shared Memory Manager.
+         The driver provides an interface to items in a heap shared among all
+         processors in a Qualcomm platform.
index 931d385386c535b846548a3c6b9070c774013fa1..3a033c43c0ef95ddae9478d695f52508901300aa 100644 (file)
@@ -1,2 +1,3 @@
 obj-$(CONFIG_QCOM_GSBI)        +=      qcom_gsbi.o
 obj-$(CONFIG_QCOM_PM)  +=      spm.o
+obj-$(CONFIG_QCOM_SMEM) +=     smem.o
diff --git a/drivers/soc/qcom/smem.c b/drivers/soc/qcom/smem.c
new file mode 100644 (file)
index 0000000..7c2c324
--- /dev/null
@@ -0,0 +1,775 @@
+/*
+ * Copyright (c) 2015, Sony Mobile Communications AB.
+ * Copyright (c) 2012-2013, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * 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.
+ */
+
+#include <linux/hwspinlock.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/soc/qcom/smem.h>
+
+/*
+ * The Qualcomm shared memory system is a allocate only heap structure that
+ * consists of one of more memory areas that can be accessed by the processors
+ * in the SoC.
+ *
+ * All systems contains a global heap, accessible by all processors in the SoC,
+ * with a table of contents data structure (@smem_header) at the beginning of
+ * the main shared memory block.
+ *
+ * The global header contains meta data for allocations as well as a fixed list
+ * of 512 entries (@smem_global_entry) that can be initialized to reference
+ * parts of the shared memory space.
+ *
+ *
+ * In addition to this global heap a set of "private" heaps can be set up at
+ * boot time with access restrictions so that only certain processor pairs can
+ * access the data.
+ *
+ * These partitions are referenced from an optional partition table
+ * (@smem_ptable), that is found 4kB from the end of the main smem region. The
+ * partition table entries (@smem_ptable_entry) lists the involved processors
+ * (or hosts) and their location in the main shared memory region.
+ *
+ * Each partition starts with a header (@smem_partition_header) that identifies
+ * the partition and holds properties for the two internal memory regions. The
+ * two regions are cached and non-cached memory respectively. Each region
+ * contain a link list of allocation headers (@smem_private_entry) followed by
+ * their data.
+ *
+ * Items in the non-cached region are allocated from the start of the partition
+ * while items in the cached region are allocated from the end. The free area
+ * is hence the region between the cached and non-cached offsets.
+ *
+ *
+ * To synchronize allocations in the shared memory heaps a remote spinlock must
+ * be held - currently lock number 3 of the sfpb or tcsr is used for this on all
+ * platforms.
+ *
+ */
+
+/*
+ * Item 3 of the global heap contains an array of versions for the various
+ * software components in the SoC. We verify that the boot loader version is
+ * what the expected version (SMEM_EXPECTED_VERSION) as a sanity check.
+ */
+#define SMEM_ITEM_VERSION      3
+#define  SMEM_MASTER_SBL_VERSION_INDEX 7
+#define  SMEM_EXPECTED_VERSION         11
+
+/*
+ * The first 8 items are only to be allocated by the boot loader while
+ * initializing the heap.
+ */
+#define SMEM_ITEM_LAST_FIXED   8
+
+/* Highest accepted item number, for both global and private heaps */
+#define SMEM_ITEM_COUNT                512
+
+/* Processor/host identifier for the application processor */
+#define SMEM_HOST_APPS         0
+
+/* Max number of processors/hosts in a system */
+#define SMEM_HOST_COUNT                9
+
+/**
+  * struct smem_proc_comm - proc_comm communication struct (legacy)
+  * @command:  current command to be executed
+  * @status:   status of the currently requested command
+  * @params:   parameters to the command
+  */
+struct smem_proc_comm {
+       u32 command;
+       u32 status;
+       u32 params[2];
+};
+
+/**
+ * struct smem_global_entry - entry to reference smem items on the heap
+ * @allocated: boolean to indicate if this entry is used
+ * @offset:    offset to the allocated space
+ * @size:      size of the allocated space, 8 byte aligned
+ * @aux_base:  base address for the memory region used by this unit, or 0 for
+ *             the default region. bits 0,1 are reserved
+ */
+struct smem_global_entry {
+       u32 allocated;
+       u32 offset;
+       u32 size;
+       u32 aux_base; /* bits 1:0 reserved */
+};
+#define AUX_BASE_MASK          0xfffffffc
+
+/**
+ * struct smem_header - header found in beginning of primary smem region
+ * @proc_comm:         proc_comm communication interface (legacy)
+ * @version:           array of versions for the various subsystems
+ * @initialized:       boolean to indicate that smem is initialized
+ * @free_offset:       index of the first unallocated byte in smem
+ * @available:         number of bytes available for allocation
+ * @reserved:          reserved field, must be 0
+ * toc:                        array of references to items
+ */
+struct smem_header {
+       struct smem_proc_comm proc_comm[4];
+       u32 version[32];
+       u32 initialized;
+       u32 free_offset;
+       u32 available;
+       u32 reserved;
+       struct smem_global_entry toc[SMEM_ITEM_COUNT];
+};
+
+/**
+ * struct smem_ptable_entry - one entry in the @smem_ptable list
+ * @offset:    offset, within the main shared memory region, of the partition
+ * @size:      size of the partition
+ * @flags:     flags for the partition (currently unused)
+ * @host0:     first processor/host with access to this partition
+ * @host1:     second processor/host with access to this partition
+ * @reserved:  reserved entries for later use
+ */
+struct smem_ptable_entry {
+       u32 offset;
+       u32 size;
+       u32 flags;
+       u16 host0;
+       u16 host1;
+       u32 reserved[8];
+};
+
+/**
+ * struct smem_ptable - partition table for the private partitions
+ * @magic:     magic number, must be SMEM_PTABLE_MAGIC
+ * @version:   version of the partition table
+ * @num_entries: number of partitions in the table
+ * @reserved:  for now reserved entries
+ * @entry:     list of @smem_ptable_entry for the @num_entries partitions
+ */
+struct smem_ptable {
+       u32 magic;
+       u32 version;
+       u32 num_entries;
+       u32 reserved[5];
+       struct smem_ptable_entry entry[];
+};
+#define SMEM_PTABLE_MAGIC      0x434f5424 /* "$TOC" */
+
+/**
+ * struct smem_partition_header - header of the partitions
+ * @magic:     magic number, must be SMEM_PART_MAGIC
+ * @host0:     first processor/host with access to this partition
+ * @host1:     second processor/host with access to this partition
+ * @size:      size of the partition
+ * @offset_free_uncached: offset to the first free byte of uncached memory in
+ *             this partition
+ * @offset_free_cached: offset to the first free byte of cached memory in this
+ *             partition
+ * @reserved:  for now reserved entries
+ */
+struct smem_partition_header {
+       u32 magic;
+       u16 host0;
+       u16 host1;
+       u32 size;
+       u32 offset_free_uncached;
+       u32 offset_free_cached;
+       u32 reserved[3];
+};
+#define SMEM_PART_MAGIC                0x54525024 /* "$PRT" */
+
+/**
+ * struct smem_private_entry - header of each item in the private partition
+ * @canary:    magic number, must be SMEM_PRIVATE_CANARY
+ * @item:      identifying number of the smem item
+ * @size:      size of the data, including padding bytes
+ * @padding_data: number of bytes of padding of data
+ * @padding_hdr: number of bytes of padding between the header and the data
+ * @reserved:  for now reserved entry
+ */
+struct smem_private_entry {
+       u16 canary;
+       u16 item;
+       u32 size; /* includes padding bytes */
+       u16 padding_data;
+       u16 padding_hdr;
+       u32 reserved;
+};
+#define SMEM_PRIVATE_CANARY    0xa5a5
+
+/**
+ * struct smem_region - representation of a chunk of memory used for smem
+ * @aux_base:  identifier of aux_mem base
+ * @virt_base: virtual base address of memory with this aux_mem identifier
+ * @size:      size of the memory region
+ */
+struct smem_region {
+       u32 aux_base;
+       void __iomem *virt_base;
+       size_t size;
+};
+
+/**
+ * struct qcom_smem - device data for the smem device
+ * @dev:       device pointer
+ * @hwlock:    reference to a hwspinlock
+ * @partitions:        list of pointers to partitions affecting the current
+ *             processor/host
+ * @num_regions: number of @regions
+ * @regions:   list of the memory regions defining the shared memory
+ */
+struct qcom_smem {
+       struct device *dev;
+
+       struct hwspinlock *hwlock;
+
+       struct smem_partition_header *partitions[SMEM_HOST_COUNT];
+
+       unsigned num_regions;
+       struct smem_region regions[0];
+};
+
+/* Pointer to the one and only smem handle */
+static struct qcom_smem *__smem;
+
+/* Timeout (ms) for the trylock of remote spinlocks */
+#define HWSPINLOCK_TIMEOUT     1000
+
+static int qcom_smem_alloc_private(struct qcom_smem *smem,
+                                  unsigned host,
+                                  unsigned item,
+                                  size_t size)
+{
+       struct smem_partition_header *phdr;
+       struct smem_private_entry *hdr;
+       size_t alloc_size;
+       void *p;
+
+       /* We're not going to find it if there's no matching partition */
+       if (host >= SMEM_HOST_COUNT || !smem->partitions[host])
+               return -ENOENT;
+
+       phdr = smem->partitions[host];
+
+       p = (void *)phdr + sizeof(*phdr);
+       while (p < (void *)phdr + phdr->offset_free_uncached) {
+               hdr = p;
+
+               if (hdr->canary != SMEM_PRIVATE_CANARY) {
+                       dev_err(smem->dev,
+                               "Found invalid canary in host %d partition\n",
+                               host);
+                       return -EINVAL;
+               }
+
+               if (hdr->item == item)
+                       return -EEXIST;
+
+               p += sizeof(*hdr) + hdr->padding_hdr + hdr->size;
+       }
+
+       /* Check that we don't grow into the cached region */
+       alloc_size = sizeof(*hdr) + ALIGN(size, 8);
+       if (p + alloc_size >= (void *)phdr + phdr->offset_free_cached) {
+               dev_err(smem->dev, "Out of memory\n");
+               return -ENOSPC;
+       }
+
+       hdr = p;
+       hdr->canary = SMEM_PRIVATE_CANARY;
+       hdr->item = item;
+       hdr->size = ALIGN(size, 8);
+       hdr->padding_data = hdr->size - size;
+       hdr->padding_hdr = 0;
+
+       /*
+        * Ensure the header is written before we advance the free offset, so
+        * that remote processors that does not take the remote spinlock still
+        * gets a consistent view of the linked list.
+        */
+       wmb();
+       phdr->offset_free_uncached += alloc_size;
+
+       return 0;
+}
+
+static int qcom_smem_alloc_global(struct qcom_smem *smem,
+                                 unsigned item,
+                                 size_t size)
+{
+       struct smem_header *header;
+       struct smem_global_entry *entry;
+
+       if (WARN_ON(item >= SMEM_ITEM_COUNT))
+               return -EINVAL;
+
+       header = smem->regions[0].virt_base;
+       entry = &header->toc[item];
+       if (entry->allocated)
+               return -EEXIST;
+
+       size = ALIGN(size, 8);
+       if (WARN_ON(size > header->available))
+               return -ENOMEM;
+
+       entry->offset = header->free_offset;
+       entry->size = size;
+
+       /*
+        * Ensure the header is consistent before we mark the item allocated,
+        * so that remote processors will get a consistent view of the item
+        * even though they do not take the spinlock on read.
+        */
+       wmb();
+       entry->allocated = 1;
+
+       header->free_offset += size;
+       header->available -= size;
+
+       return 0;
+}
+
+/**
+ * qcom_smem_alloc() - allocate space for a smem item
+ * @host:      remote processor id, or -1
+ * @item:      smem item handle
+ * @size:      number of bytes to be allocated
+ *
+ * Allocate space for a given smem item of size @size, given that the item is
+ * not yet allocated.
+ */
+int qcom_smem_alloc(unsigned host, unsigned item, size_t size)
+{
+       unsigned long flags;
+       int ret;
+
+       if (!__smem)
+               return -EPROBE_DEFER;
+
+       if (item < SMEM_ITEM_LAST_FIXED) {
+               dev_err(__smem->dev,
+                       "Rejecting allocation of static entry %d\n", item);
+               return -EINVAL;
+       }
+
+       ret = hwspin_lock_timeout_irqsave(__smem->hwlock,
+                                         HWSPINLOCK_TIMEOUT,
+                                         &flags);
+       if (ret)
+               return ret;
+
+       ret = qcom_smem_alloc_private(__smem, host, item, size);
+       if (ret == -ENOENT)
+               ret = qcom_smem_alloc_global(__smem, item, size);
+
+       hwspin_unlock_irqrestore(__smem->hwlock, &flags);
+
+       return ret;
+}
+EXPORT_SYMBOL(qcom_smem_alloc);
+
+static int qcom_smem_get_global(struct qcom_smem *smem,
+                               unsigned item,
+                               void **ptr,
+                               size_t *size)
+{
+       struct smem_header *header;
+       struct smem_region *area;
+       struct smem_global_entry *entry;
+       u32 aux_base;
+       unsigned i;
+
+       if (WARN_ON(item >= SMEM_ITEM_COUNT))
+               return -EINVAL;
+
+       header = smem->regions[0].virt_base;
+       entry = &header->toc[item];
+       if (!entry->allocated)
+               return -ENXIO;
+
+       if (ptr != NULL) {
+               aux_base = entry->aux_base & AUX_BASE_MASK;
+
+               for (i = 0; i < smem->num_regions; i++) {
+                       area = &smem->regions[i];
+
+                       if (area->aux_base == aux_base || !aux_base) {
+                               *ptr = area->virt_base + entry->offset;
+                               break;
+                       }
+               }
+       }
+       if (size != NULL)
+               *size = entry->size;
+
+       return 0;
+}
+
+static int qcom_smem_get_private(struct qcom_smem *smem,
+                                unsigned host,
+                                unsigned item,
+                                void **ptr,
+                                size_t *size)
+{
+       struct smem_partition_header *phdr;
+       struct smem_private_entry *hdr;
+       void *p;
+
+       /* We're not going to find it if there's no matching partition */
+       if (host >= SMEM_HOST_COUNT || !smem->partitions[host])
+               return -ENOENT;
+
+       phdr = smem->partitions[host];
+
+       p = (void *)phdr + sizeof(*phdr);
+       while (p < (void *)phdr + phdr->offset_free_uncached) {
+               hdr = p;
+
+               if (hdr->canary != SMEM_PRIVATE_CANARY) {
+                       dev_err(smem->dev,
+                               "Found invalid canary in host %d partition\n",
+                               host);
+                       return -EINVAL;
+               }
+
+               if (hdr->item == item) {
+                       if (ptr != NULL)
+                               *ptr = p + sizeof(*hdr) + hdr->padding_hdr;
+
+                       if (size != NULL)
+                               *size = hdr->size - hdr->padding_data;
+
+                       return 0;
+               }
+
+               p += sizeof(*hdr) + hdr->padding_hdr + hdr->size;
+       }
+
+       return -ENOENT;
+}
+
+/**
+ * qcom_smem_get() - resolve ptr of size of a smem item
+ * @host:      the remote processor, or -1
+ * @item:      smem item handle
+ * @ptr:       pointer to be filled out with address of the item
+ * @size:      pointer to be filled out with size of the item
+ *
+ * Looks up pointer and size of a smem item.
+ */
+int qcom_smem_get(unsigned host, unsigned item, void **ptr, size_t *size)
+{
+       unsigned long flags;
+       int ret;
+
+       if (!__smem)
+               return -EPROBE_DEFER;
+
+       ret = hwspin_lock_timeout_irqsave(__smem->hwlock,
+                                         HWSPINLOCK_TIMEOUT,
+                                         &flags);
+       if (ret)
+               return ret;
+
+       ret = qcom_smem_get_private(__smem, host, item, ptr, size);
+       if (ret == -ENOENT)
+               ret = qcom_smem_get_global(__smem, item, ptr, size);
+
+       hwspin_unlock_irqrestore(__smem->hwlock, &flags);
+       return ret;
+
+}
+EXPORT_SYMBOL(qcom_smem_get);
+
+/**
+ * qcom_smem_get_free_space() - retrieve amount of free space in a partition
+ * @host:      the remote processor identifying a partition, or -1
+ *
+ * To be used by smem clients as a quick way to determine if any new
+ * allocations has been made.
+ */
+int qcom_smem_get_free_space(unsigned host)
+{
+       struct smem_partition_header *phdr;
+       struct smem_header *header;
+       unsigned ret;
+
+       if (!__smem)
+               return -EPROBE_DEFER;
+
+       if (host < SMEM_HOST_COUNT && __smem->partitions[host]) {
+               phdr = __smem->partitions[host];
+               ret = phdr->offset_free_cached - phdr->offset_free_uncached;
+       } else {
+               header = __smem->regions[0].virt_base;
+               ret = header->available;
+       }
+
+       return ret;
+}
+EXPORT_SYMBOL(qcom_smem_get_free_space);
+
+static int qcom_smem_get_sbl_version(struct qcom_smem *smem)
+{
+       unsigned *versions;
+       size_t size;
+       int ret;
+
+       ret = qcom_smem_get_global(smem, SMEM_ITEM_VERSION,
+                                  (void **)&versions, &size);
+       if (ret < 0) {
+               dev_err(smem->dev, "Unable to read the version item\n");
+               return -ENOENT;
+       }
+
+       if (size < sizeof(unsigned) * SMEM_MASTER_SBL_VERSION_INDEX) {
+               dev_err(smem->dev, "Version item is too small\n");
+               return -EINVAL;
+       }
+
+       return versions[SMEM_MASTER_SBL_VERSION_INDEX];
+}
+
+static int qcom_smem_enumerate_partitions(struct qcom_smem *smem,
+                                         unsigned local_host)
+{
+       struct smem_partition_header *header;
+       struct smem_ptable_entry *entry;
+       struct smem_ptable *ptable;
+       unsigned remote_host;
+       int i;
+
+       ptable = smem->regions[0].virt_base + smem->regions[0].size - SZ_4K;
+       if (ptable->magic != SMEM_PTABLE_MAGIC)
+               return 0;
+
+       if (ptable->version != 1) {
+               dev_err(smem->dev,
+                       "Unsupported partition header version %d\n",
+                       ptable->version);
+               return -EINVAL;
+       }
+
+       for (i = 0; i < ptable->num_entries; i++) {
+               entry = &ptable->entry[i];
+
+               if (entry->host0 != local_host && entry->host1 != local_host)
+                       continue;
+
+               if (!entry->offset)
+                       continue;
+
+               if (!entry->size)
+                       continue;
+
+               if (entry->host0 == local_host)
+                       remote_host = entry->host1;
+               else
+                       remote_host = entry->host0;
+
+               if (remote_host >= SMEM_HOST_COUNT) {
+                       dev_err(smem->dev,
+                               "Invalid remote host %d\n",
+                               remote_host);
+                       return -EINVAL;
+               }
+
+               if (smem->partitions[remote_host]) {
+                       dev_err(smem->dev,
+                               "Already found a partition for host %d\n",
+                               remote_host);
+                       return -EINVAL;
+               }
+
+               header = smem->regions[0].virt_base + entry->offset;
+
+               if (header->magic != SMEM_PART_MAGIC) {
+                       dev_err(smem->dev,
+                               "Partition %d has invalid magic\n", i);
+                       return -EINVAL;
+               }
+
+               if (header->host0 != local_host && header->host1 != local_host) {
+                       dev_err(smem->dev,
+                               "Partition %d hosts are invalid\n", i);
+                       return -EINVAL;
+               }
+
+               if (header->host0 != remote_host && header->host1 != remote_host) {
+                       dev_err(smem->dev,
+                               "Partition %d hosts are invalid\n", i);
+                       return -EINVAL;
+               }
+
+               if (header->size != entry->size) {
+                       dev_err(smem->dev,
+                               "Partition %d has invalid size\n", i);
+                       return -EINVAL;
+               }
+
+               if (header->offset_free_uncached > header->size) {
+                       dev_err(smem->dev,
+                               "Partition %d has invalid free pointer\n", i);
+                       return -EINVAL;
+               }
+
+               smem->partitions[remote_host] = header;
+       }
+
+       return 0;
+}
+
+static int qcom_smem_count_mem_regions(struct platform_device *pdev)
+{
+       struct resource *res;
+       int num_regions = 0;
+       int i;
+
+       for (i = 0; i < pdev->num_resources; i++) {
+               res = &pdev->resource[i];
+
+               if (resource_type(res) == IORESOURCE_MEM)
+                       num_regions++;
+       }
+
+       return num_regions;
+}
+
+static int qcom_smem_probe(struct platform_device *pdev)
+{
+       struct smem_header *header;
+       struct device_node *np;
+       struct qcom_smem *smem;
+       struct resource *res;
+       struct resource r;
+       size_t array_size;
+       int num_regions = 0;
+       int hwlock_id;
+       u32 version;
+       int ret;
+       int i;
+
+       num_regions = qcom_smem_count_mem_regions(pdev) + 1;
+
+       array_size = num_regions * sizeof(struct smem_region);
+       smem = devm_kzalloc(&pdev->dev, sizeof(*smem) + array_size, GFP_KERNEL);
+       if (!smem)
+               return -ENOMEM;
+
+       smem->dev = &pdev->dev;
+       smem->num_regions = num_regions;
+
+       np = of_parse_phandle(pdev->dev.of_node, "memory-region", 0);
+       if (!np) {
+               dev_err(&pdev->dev, "No memory-region specified\n");
+               return -EINVAL;
+       }
+
+       ret = of_address_to_resource(np, 0, &r);
+       of_node_put(np);
+       if (ret)
+               return ret;
+
+       smem->regions[0].aux_base = (u32)r.start;
+       smem->regions[0].size = resource_size(&r);
+       smem->regions[0].virt_base = devm_ioremap_nocache(&pdev->dev,
+                                                         r.start,
+                                                         resource_size(&r));
+       if (!smem->regions[0].virt_base)
+               return -ENOMEM;
+
+       for (i = 1; i < num_regions; i++) {
+               res = platform_get_resource(pdev, IORESOURCE_MEM, i - 1);
+
+               smem->regions[i].aux_base = (u32)res->start;
+               smem->regions[i].size = resource_size(res);
+               smem->regions[i].virt_base = devm_ioremap_nocache(&pdev->dev,
+                                                                 res->start,
+                                                                 resource_size(res));
+               if (!smem->regions[i].virt_base)
+                       return -ENOMEM;
+       }
+
+       header = smem->regions[0].virt_base;
+       if (header->initialized != 1 || header->reserved) {
+               dev_err(&pdev->dev, "SMEM is not initialized by SBL\n");
+               return -EINVAL;
+       }
+
+       version = qcom_smem_get_sbl_version(smem);
+       if (version >> 16 != SMEM_EXPECTED_VERSION) {
+               dev_err(&pdev->dev, "Unsupported SMEM version 0x%x\n", version);
+               return -EINVAL;
+       }
+
+       ret = qcom_smem_enumerate_partitions(smem, SMEM_HOST_APPS);
+       if (ret < 0)
+               return ret;
+
+       hwlock_id = of_hwspin_lock_get_id(pdev->dev.of_node, 0);
+       if (hwlock_id < 0) {
+               dev_err(&pdev->dev, "failed to retrieve hwlock\n");
+               return hwlock_id;
+       }
+
+       smem->hwlock = hwspin_lock_request_specific(hwlock_id);
+       if (!smem->hwlock)
+               return -ENXIO;
+
+       __smem = smem;
+
+       return 0;
+}
+
+static int qcom_smem_remove(struct platform_device *pdev)
+{
+       __smem = NULL;
+       hwspin_lock_free(__smem->hwlock);
+
+       return 0;
+}
+
+static const struct of_device_id qcom_smem_of_match[] = {
+       { .compatible = "qcom,smem" },
+       {}
+};
+MODULE_DEVICE_TABLE(of, qcom_smem_of_match);
+
+static struct platform_driver qcom_smem_driver = {
+       .probe = qcom_smem_probe,
+       .remove = qcom_smem_remove,
+       .driver  = {
+               .name = "qcom-smem",
+               .of_match_table = qcom_smem_of_match,
+               .suppress_bind_attrs = true,
+       },
+};
+
+static int __init qcom_smem_init(void)
+{
+       return platform_driver_register(&qcom_smem_driver);
+}
+arch_initcall(qcom_smem_init);
+
+static void __exit qcom_smem_exit(void)
+{
+       platform_driver_unregister(&qcom_smem_driver);
+}
+module_exit(qcom_smem_exit)
+
+MODULE_AUTHOR("Bjorn Andersson <bjorn.andersson@sonymobile.com>");
+MODULE_DESCRIPTION("Qualcomm Shared Memory Manager");
+MODULE_LICENSE("GPL v2");
diff --git a/include/linux/soc/qcom/smem.h b/include/linux/soc/qcom/smem.h
new file mode 100644 (file)
index 0000000..bc9630d
--- /dev/null
@@ -0,0 +1,11 @@
+#ifndef __QCOM_SMEM_H__
+#define __QCOM_SMEM_H__
+
+#define QCOM_SMEM_HOST_ANY -1
+
+int qcom_smem_alloc(unsigned host, unsigned item, size_t size);
+int qcom_smem_get(unsigned host, unsigned item, void **ptr, size_t *size);
+
+int qcom_smem_get_free_space(unsigned host);
+
+#endif