| /* |
| * Intel SST Firmware Loader |
| * |
| * Copyright (C) 2013, Intel Corporation. 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 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/kernel.h> |
| #include <linux/slab.h> |
| #include <linux/sched.h> |
| #include <linux/firmware.h> |
| #include <linux/export.h> |
| #include <linux/platform_device.h> |
| #include <linux/dma-mapping.h> |
| #include <linux/dmaengine.h> |
| #include <linux/pci.h> |
| #include <linux/acpi.h> |
| |
| /* supported DMA engine drivers */ |
| #include <linux/platform_data/dma-dw.h> |
| #include <linux/dma/dw.h> |
| |
| #include <asm/page.h> |
| #include <asm/pgtable.h> |
| |
| #include "sst-dsp.h" |
| #include "sst-dsp-priv.h" |
| |
| #define SST_DMA_RESOURCES 2 |
| #define SST_DSP_DMA_MAX_BURST 0x3 |
| #define SST_HSW_BLOCK_ANY 0xffffffff |
| |
| #define SST_HSW_MASK_DMA_ADDR_DSP 0xfff00000 |
| |
| struct sst_dma { |
| struct sst_dsp *sst; |
| |
| struct dw_dma_chip *chip; |
| |
| struct dma_async_tx_descriptor *desc; |
| struct dma_chan *ch; |
| }; |
| |
| static inline void sst_memcpy32(volatile void __iomem *dest, void *src, u32 bytes) |
| { |
| /* __iowrite32_copy use 32bit size values so divide by 4 */ |
| __iowrite32_copy((void *)dest, src, bytes/4); |
| } |
| |
| static void sst_dma_transfer_complete(void *arg) |
| { |
| struct sst_dsp *sst = (struct sst_dsp *)arg; |
| |
| dev_dbg(sst->dev, "DMA: callback\n"); |
| } |
| |
| static int sst_dsp_dma_copy(struct sst_dsp *sst, dma_addr_t dest_addr, |
| dma_addr_t src_addr, size_t size) |
| { |
| struct dma_async_tx_descriptor *desc; |
| struct sst_dma *dma = sst->dma; |
| |
| if (dma->ch == NULL) { |
| dev_err(sst->dev, "error: no DMA channel\n"); |
| return -ENODEV; |
| } |
| |
| dev_dbg(sst->dev, "DMA: src: 0x%lx dest 0x%lx size %zu\n", |
| (unsigned long)src_addr, (unsigned long)dest_addr, size); |
| |
| desc = dma->ch->device->device_prep_dma_memcpy(dma->ch, dest_addr, |
| src_addr, size, DMA_CTRL_ACK); |
| if (!desc){ |
| dev_err(sst->dev, "error: dma prep memcpy failed\n"); |
| return -EINVAL; |
| } |
| |
| desc->callback = sst_dma_transfer_complete; |
| desc->callback_param = sst; |
| |
| desc->tx_submit(desc); |
| dma_wait_for_async_tx(desc); |
| |
| return 0; |
| } |
| |
| /* copy to DSP */ |
| int sst_dsp_dma_copyto(struct sst_dsp *sst, dma_addr_t dest_addr, |
| dma_addr_t src_addr, size_t size) |
| { |
| return sst_dsp_dma_copy(sst, dest_addr | SST_HSW_MASK_DMA_ADDR_DSP, |
| src_addr, size); |
| } |
| EXPORT_SYMBOL_GPL(sst_dsp_dma_copyto); |
| |
| /* copy from DSP */ |
| int sst_dsp_dma_copyfrom(struct sst_dsp *sst, dma_addr_t dest_addr, |
| dma_addr_t src_addr, size_t size) |
| { |
| return sst_dsp_dma_copy(sst, dest_addr, |
| src_addr | SST_HSW_MASK_DMA_ADDR_DSP, size); |
| } |
| EXPORT_SYMBOL_GPL(sst_dsp_dma_copyfrom); |
| |
| /* remove module from memory - callers hold locks */ |
| static void block_list_remove(struct sst_dsp *dsp, |
| struct list_head *block_list) |
| { |
| struct sst_mem_block *block, *tmp; |
| int err; |
| |
| /* disable each block */ |
| list_for_each_entry(block, block_list, module_list) { |
| |
| if (block->ops && block->ops->disable) { |
| err = block->ops->disable(block); |
| if (err < 0) |
| dev_err(dsp->dev, |
| "error: cant disable block %d:%d\n", |
| block->type, block->index); |
| } |
| } |
| |
| /* mark each block as free */ |
| list_for_each_entry_safe(block, tmp, block_list, module_list) { |
| list_del(&block->module_list); |
| list_move(&block->list, &dsp->free_block_list); |
| dev_dbg(dsp->dev, "block freed %d:%d at offset 0x%x\n", |
| block->type, block->index, block->offset); |
| } |
| } |
| |
| /* prepare the memory block to receive data from host - callers hold locks */ |
| static int block_list_prepare(struct sst_dsp *dsp, |
| struct list_head *block_list) |
| { |
| struct sst_mem_block *block; |
| int ret = 0; |
| |
| /* enable each block so that's it'e ready for data */ |
| list_for_each_entry(block, block_list, module_list) { |
| |
| if (block->ops && block->ops->enable && !block->users) { |
| ret = block->ops->enable(block); |
| if (ret < 0) { |
| dev_err(dsp->dev, |
| "error: cant disable block %d:%d\n", |
| block->type, block->index); |
| goto err; |
| } |
| } |
| } |
| return ret; |
| |
| err: |
| list_for_each_entry(block, block_list, module_list) { |
| if (block->ops && block->ops->disable) |
| block->ops->disable(block); |
| } |
| return ret; |
| } |
| |
| static struct dw_dma_platform_data dw_pdata = { |
| .is_private = 1, |
| .chan_allocation_order = CHAN_ALLOCATION_ASCENDING, |
| .chan_priority = CHAN_PRIORITY_ASCENDING, |
| }; |
| |
| static struct dw_dma_chip *dw_probe(struct device *dev, struct resource *mem, |
| int irq) |
| { |
| struct dw_dma_chip *chip; |
| int err; |
| |
| chip = devm_kzalloc(dev, sizeof(*chip), GFP_KERNEL); |
| if (!chip) |
| return ERR_PTR(-ENOMEM); |
| |
| chip->irq = irq; |
| chip->regs = devm_ioremap_resource(dev, mem); |
| if (IS_ERR(chip->regs)) |
| return ERR_CAST(chip->regs); |
| |
| err = dma_coerce_mask_and_coherent(dev, DMA_BIT_MASK(31)); |
| if (err) |
| return ERR_PTR(err); |
| |
| chip->dev = dev; |
| err = dw_dma_probe(chip, &dw_pdata); |
| if (err) |
| return ERR_PTR(err); |
| |
| return chip; |
| } |
| |
| static void dw_remove(struct dw_dma_chip *chip) |
| { |
| dw_dma_remove(chip); |
| } |
| |
| static bool dma_chan_filter(struct dma_chan *chan, void *param) |
| { |
| struct sst_dsp *dsp = (struct sst_dsp *)param; |
| |
| return chan->device->dev == dsp->dma_dev; |
| } |
| |
| int sst_dsp_dma_get_channel(struct sst_dsp *dsp, int chan_id) |
| { |
| struct sst_dma *dma = dsp->dma; |
| struct dma_slave_config slave; |
| dma_cap_mask_t mask; |
| int ret; |
| |
| /* The Intel MID DMA engine driver needs the slave config set but |
| * Synopsis DMA engine driver safely ignores the slave config */ |
| dma_cap_zero(mask); |
| dma_cap_set(DMA_SLAVE, mask); |
| dma_cap_set(DMA_MEMCPY, mask); |
| |
| dma->ch = dma_request_channel(mask, dma_chan_filter, dsp); |
| if (dma->ch == NULL) { |
| dev_err(dsp->dev, "error: DMA request channel failed\n"); |
| return -EIO; |
| } |
| |
| memset(&slave, 0, sizeof(slave)); |
| slave.direction = DMA_MEM_TO_DEV; |
| slave.src_addr_width = |
| slave.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES; |
| slave.src_maxburst = slave.dst_maxburst = SST_DSP_DMA_MAX_BURST; |
| |
| ret = dmaengine_slave_config(dma->ch, &slave); |
| if (ret) { |
| dev_err(dsp->dev, "error: unable to set DMA slave config %d\n", |
| ret); |
| dma_release_channel(dma->ch); |
| dma->ch = NULL; |
| } |
| |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_dsp_dma_get_channel); |
| |
| void sst_dsp_dma_put_channel(struct sst_dsp *dsp) |
| { |
| struct sst_dma *dma = dsp->dma; |
| |
| if (!dma->ch) |
| return; |
| |
| dma_release_channel(dma->ch); |
| dma->ch = NULL; |
| } |
| EXPORT_SYMBOL_GPL(sst_dsp_dma_put_channel); |
| |
| int sst_dma_new(struct sst_dsp *sst) |
| { |
| struct sst_pdata *sst_pdata = sst->pdata; |
| struct sst_dma *dma; |
| struct resource mem; |
| const char *dma_dev_name; |
| int ret = 0; |
| |
| /* configure the correct platform data for whatever DMA engine |
| * is attached to the ADSP IP. */ |
| switch (sst->pdata->dma_engine) { |
| case SST_DMA_TYPE_DW: |
| dma_dev_name = "dw_dmac"; |
| break; |
| case SST_DMA_TYPE_MID: |
| dma_dev_name = "Intel MID DMA"; |
| break; |
| default: |
| dev_err(sst->dev, "error: invalid DMA engine %d\n", |
| sst->pdata->dma_engine); |
| return -EINVAL; |
| } |
| |
| dma = devm_kzalloc(sst->dev, sizeof(struct sst_dma), GFP_KERNEL); |
| if (!dma) |
| return -ENOMEM; |
| |
| dma->sst = sst; |
| |
| memset(&mem, 0, sizeof(mem)); |
| |
| mem.start = sst->addr.lpe_base + sst_pdata->dma_base; |
| mem.end = sst->addr.lpe_base + sst_pdata->dma_base + sst_pdata->dma_size - 1; |
| mem.flags = IORESOURCE_MEM; |
| |
| /* now register DMA engine device */ |
| dma->chip = dw_probe(sst->dma_dev, &mem, sst_pdata->irq); |
| if (IS_ERR(dma->chip)) { |
| dev_err(sst->dev, "error: DMA device register failed\n"); |
| ret = PTR_ERR(dma->chip); |
| goto err_dma_dev; |
| } |
| |
| sst->dma = dma; |
| sst->fw_use_dma = true; |
| return 0; |
| |
| err_dma_dev: |
| devm_kfree(sst->dev, dma); |
| return ret; |
| } |
| EXPORT_SYMBOL(sst_dma_new); |
| |
| void sst_dma_free(struct sst_dma *dma) |
| { |
| |
| if (dma == NULL) |
| return; |
| |
| if (dma->ch) |
| dma_release_channel(dma->ch); |
| |
| if (dma->chip) |
| dw_remove(dma->chip); |
| |
| } |
| EXPORT_SYMBOL(sst_dma_free); |
| |
| /* create new generic firmware object */ |
| struct sst_fw *sst_fw_new(struct sst_dsp *dsp, |
| const struct firmware *fw, void *private) |
| { |
| struct sst_fw *sst_fw; |
| int err; |
| |
| if (!dsp->ops->parse_fw) |
| return NULL; |
| |
| sst_fw = kzalloc(sizeof(*sst_fw), GFP_KERNEL); |
| if (sst_fw == NULL) |
| return NULL; |
| |
| sst_fw->dsp = dsp; |
| sst_fw->private = private; |
| sst_fw->size = fw->size; |
| |
| /* allocate DMA buffer to store FW data */ |
| sst_fw->dma_buf = dma_alloc_coherent(dsp->dma_dev, sst_fw->size, |
| &sst_fw->dmable_fw_paddr, GFP_DMA | GFP_KERNEL); |
| if (!sst_fw->dma_buf) { |
| dev_err(dsp->dev, "error: DMA alloc failed\n"); |
| kfree(sst_fw); |
| return NULL; |
| } |
| |
| /* copy FW data to DMA-able memory */ |
| memcpy((void *)sst_fw->dma_buf, (void *)fw->data, fw->size); |
| |
| if (dsp->fw_use_dma) { |
| err = sst_dsp_dma_get_channel(dsp, 0); |
| if (err < 0) |
| goto chan_err; |
| } |
| |
| /* call core specific FW paser to load FW data into DSP */ |
| err = dsp->ops->parse_fw(sst_fw); |
| if (err < 0) { |
| dev_err(dsp->dev, "error: parse fw failed %d\n", err); |
| goto parse_err; |
| } |
| |
| if (dsp->fw_use_dma) |
| sst_dsp_dma_put_channel(dsp); |
| |
| mutex_lock(&dsp->mutex); |
| list_add(&sst_fw->list, &dsp->fw_list); |
| mutex_unlock(&dsp->mutex); |
| |
| return sst_fw; |
| |
| parse_err: |
| if (dsp->fw_use_dma) |
| sst_dsp_dma_put_channel(dsp); |
| chan_err: |
| dma_free_coherent(dsp->dma_dev, sst_fw->size, |
| sst_fw->dma_buf, |
| sst_fw->dmable_fw_paddr); |
| sst_fw->dma_buf = NULL; |
| kfree(sst_fw); |
| return NULL; |
| } |
| EXPORT_SYMBOL_GPL(sst_fw_new); |
| |
| int sst_fw_reload(struct sst_fw *sst_fw) |
| { |
| struct sst_dsp *dsp = sst_fw->dsp; |
| int ret; |
| |
| dev_dbg(dsp->dev, "reloading firmware\n"); |
| |
| /* call core specific FW paser to load FW data into DSP */ |
| ret = dsp->ops->parse_fw(sst_fw); |
| if (ret < 0) |
| dev_err(dsp->dev, "error: parse fw failed %d\n", ret); |
| |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_fw_reload); |
| |
| void sst_fw_unload(struct sst_fw *sst_fw) |
| { |
| struct sst_dsp *dsp = sst_fw->dsp; |
| struct sst_module *module, *mtmp; |
| struct sst_module_runtime *runtime, *rtmp; |
| |
| dev_dbg(dsp->dev, "unloading firmware\n"); |
| |
| mutex_lock(&dsp->mutex); |
| |
| /* check module by module */ |
| list_for_each_entry_safe(module, mtmp, &dsp->module_list, list) { |
| if (module->sst_fw == sst_fw) { |
| |
| /* remove runtime modules */ |
| list_for_each_entry_safe(runtime, rtmp, &module->runtime_list, list) { |
| |
| block_list_remove(dsp, &runtime->block_list); |
| list_del(&runtime->list); |
| kfree(runtime); |
| } |
| |
| /* now remove the module */ |
| block_list_remove(dsp, &module->block_list); |
| list_del(&module->list); |
| kfree(module); |
| } |
| } |
| |
| /* remove all scratch blocks */ |
| block_list_remove(dsp, &dsp->scratch_block_list); |
| |
| mutex_unlock(&dsp->mutex); |
| } |
| EXPORT_SYMBOL_GPL(sst_fw_unload); |
| |
| /* free single firmware object */ |
| void sst_fw_free(struct sst_fw *sst_fw) |
| { |
| struct sst_dsp *dsp = sst_fw->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| list_del(&sst_fw->list); |
| mutex_unlock(&dsp->mutex); |
| |
| if (sst_fw->dma_buf) |
| dma_free_coherent(dsp->dma_dev, sst_fw->size, sst_fw->dma_buf, |
| sst_fw->dmable_fw_paddr); |
| kfree(sst_fw); |
| } |
| EXPORT_SYMBOL_GPL(sst_fw_free); |
| |
| /* free all firmware objects */ |
| void sst_fw_free_all(struct sst_dsp *dsp) |
| { |
| struct sst_fw *sst_fw, *t; |
| |
| mutex_lock(&dsp->mutex); |
| list_for_each_entry_safe(sst_fw, t, &dsp->fw_list, list) { |
| |
| list_del(&sst_fw->list); |
| dma_free_coherent(dsp->dev, sst_fw->size, sst_fw->dma_buf, |
| sst_fw->dmable_fw_paddr); |
| kfree(sst_fw); |
| } |
| mutex_unlock(&dsp->mutex); |
| } |
| EXPORT_SYMBOL_GPL(sst_fw_free_all); |
| |
| /* create a new SST generic module from FW template */ |
| struct sst_module *sst_module_new(struct sst_fw *sst_fw, |
| struct sst_module_template *template, void *private) |
| { |
| struct sst_dsp *dsp = sst_fw->dsp; |
| struct sst_module *sst_module; |
| |
| sst_module = kzalloc(sizeof(*sst_module), GFP_KERNEL); |
| if (sst_module == NULL) |
| return NULL; |
| |
| sst_module->id = template->id; |
| sst_module->dsp = dsp; |
| sst_module->sst_fw = sst_fw; |
| sst_module->scratch_size = template->scratch_size; |
| sst_module->persistent_size = template->persistent_size; |
| |
| INIT_LIST_HEAD(&sst_module->block_list); |
| INIT_LIST_HEAD(&sst_module->runtime_list); |
| |
| mutex_lock(&dsp->mutex); |
| list_add(&sst_module->list, &dsp->module_list); |
| mutex_unlock(&dsp->mutex); |
| |
| return sst_module; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_new); |
| |
| /* free firmware module and remove from available list */ |
| void sst_module_free(struct sst_module *sst_module) |
| { |
| struct sst_dsp *dsp = sst_module->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| list_del(&sst_module->list); |
| mutex_unlock(&dsp->mutex); |
| |
| kfree(sst_module); |
| } |
| EXPORT_SYMBOL_GPL(sst_module_free); |
| |
| struct sst_module_runtime *sst_module_runtime_new(struct sst_module *module, |
| int id, void *private) |
| { |
| struct sst_dsp *dsp = module->dsp; |
| struct sst_module_runtime *runtime; |
| |
| runtime = kzalloc(sizeof(*runtime), GFP_KERNEL); |
| if (runtime == NULL) |
| return NULL; |
| |
| runtime->id = id; |
| runtime->dsp = dsp; |
| runtime->module = module; |
| INIT_LIST_HEAD(&runtime->block_list); |
| |
| mutex_lock(&dsp->mutex); |
| list_add(&runtime->list, &module->runtime_list); |
| mutex_unlock(&dsp->mutex); |
| |
| return runtime; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_new); |
| |
| void sst_module_runtime_free(struct sst_module_runtime *runtime) |
| { |
| struct sst_dsp *dsp = runtime->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| list_del(&runtime->list); |
| mutex_unlock(&dsp->mutex); |
| |
| kfree(runtime); |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_free); |
| |
| static struct sst_mem_block *find_block(struct sst_dsp *dsp, |
| struct sst_block_allocator *ba) |
| { |
| struct sst_mem_block *block; |
| |
| list_for_each_entry(block, &dsp->free_block_list, list) { |
| if (block->type == ba->type && block->offset == ba->offset) |
| return block; |
| } |
| |
| return NULL; |
| } |
| |
| /* Block allocator must be on block boundary */ |
| static int block_alloc_contiguous(struct sst_dsp *dsp, |
| struct sst_block_allocator *ba, struct list_head *block_list) |
| { |
| struct list_head tmp = LIST_HEAD_INIT(tmp); |
| struct sst_mem_block *block; |
| u32 block_start = SST_HSW_BLOCK_ANY; |
| int size = ba->size, offset = ba->offset; |
| |
| while (ba->size > 0) { |
| |
| block = find_block(dsp, ba); |
| if (!block) { |
| list_splice(&tmp, &dsp->free_block_list); |
| |
| ba->size = size; |
| ba->offset = offset; |
| return -ENOMEM; |
| } |
| |
| list_move_tail(&block->list, &tmp); |
| ba->offset += block->size; |
| ba->size -= block->size; |
| } |
| ba->size = size; |
| ba->offset = offset; |
| |
| list_for_each_entry(block, &tmp, list) { |
| |
| if (block->offset < block_start) |
| block_start = block->offset; |
| |
| list_add(&block->module_list, block_list); |
| |
| dev_dbg(dsp->dev, "block allocated %d:%d at offset 0x%x\n", |
| block->type, block->index, block->offset); |
| } |
| |
| list_splice(&tmp, &dsp->used_block_list); |
| return 0; |
| } |
| |
| /* allocate first free DSP blocks for data - callers hold locks */ |
| static int block_alloc(struct sst_dsp *dsp, struct sst_block_allocator *ba, |
| struct list_head *block_list) |
| { |
| struct sst_mem_block *block, *tmp; |
| int ret = 0; |
| |
| if (ba->size == 0) |
| return 0; |
| |
| /* find first free whole blocks that can hold module */ |
| list_for_each_entry_safe(block, tmp, &dsp->free_block_list, list) { |
| |
| /* ignore blocks with wrong type */ |
| if (block->type != ba->type) |
| continue; |
| |
| if (ba->size > block->size) |
| continue; |
| |
| ba->offset = block->offset; |
| block->bytes_used = ba->size % block->size; |
| list_add(&block->module_list, block_list); |
| list_move(&block->list, &dsp->used_block_list); |
| dev_dbg(dsp->dev, "block allocated %d:%d at offset 0x%x\n", |
| block->type, block->index, block->offset); |
| return 0; |
| } |
| |
| /* then find free multiple blocks that can hold module */ |
| list_for_each_entry_safe(block, tmp, &dsp->free_block_list, list) { |
| |
| /* ignore blocks with wrong type */ |
| if (block->type != ba->type) |
| continue; |
| |
| /* do we span > 1 blocks */ |
| if (ba->size > block->size) { |
| |
| /* align ba to block boundary */ |
| ba->offset = block->offset; |
| |
| ret = block_alloc_contiguous(dsp, ba, block_list); |
| if (ret == 0) |
| return ret; |
| |
| } |
| } |
| |
| /* not enough free block space */ |
| return -ENOMEM; |
| } |
| |
| int sst_alloc_blocks(struct sst_dsp *dsp, struct sst_block_allocator *ba, |
| struct list_head *block_list) |
| { |
| int ret; |
| |
| dev_dbg(dsp->dev, "block request 0x%x bytes at offset 0x%x type %d\n", |
| ba->size, ba->offset, ba->type); |
| |
| mutex_lock(&dsp->mutex); |
| |
| ret = block_alloc(dsp, ba, block_list); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: can't alloc blocks %d\n", ret); |
| goto out; |
| } |
| |
| /* prepare DSP blocks for module usage */ |
| ret = block_list_prepare(dsp, block_list); |
| if (ret < 0) |
| dev_err(dsp->dev, "error: prepare failed\n"); |
| |
| out: |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_alloc_blocks); |
| |
| int sst_free_blocks(struct sst_dsp *dsp, struct list_head *block_list) |
| { |
| mutex_lock(&dsp->mutex); |
| block_list_remove(dsp, block_list); |
| mutex_unlock(&dsp->mutex); |
| return 0; |
| } |
| EXPORT_SYMBOL_GPL(sst_free_blocks); |
| |
| /* allocate memory blocks for static module addresses - callers hold locks */ |
| static int block_alloc_fixed(struct sst_dsp *dsp, struct sst_block_allocator *ba, |
| struct list_head *block_list) |
| { |
| struct sst_mem_block *block, *tmp; |
| u32 end = ba->offset + ba->size, block_end; |
| int err; |
| |
| /* only IRAM/DRAM blocks are managed */ |
| if (ba->type != SST_MEM_IRAM && ba->type != SST_MEM_DRAM) |
| return 0; |
| |
| /* are blocks already attached to this module */ |
| list_for_each_entry_safe(block, tmp, block_list, module_list) { |
| |
| /* ignore blocks with wrong type */ |
| if (block->type != ba->type) |
| continue; |
| |
| block_end = block->offset + block->size; |
| |
| /* find block that holds section */ |
| if (ba->offset >= block->offset && end <= block_end) |
| return 0; |
| |
| /* does block span more than 1 section */ |
| if (ba->offset >= block->offset && ba->offset < block_end) { |
| |
| /* align ba to block boundary */ |
| ba->size -= block_end - ba->offset; |
| ba->offset = block_end; |
| err = block_alloc_contiguous(dsp, ba, block_list); |
| if (err < 0) |
| return -ENOMEM; |
| |
| /* module already owns blocks */ |
| return 0; |
| } |
| } |
| |
| /* find first free blocks that can hold section in free list */ |
| list_for_each_entry_safe(block, tmp, &dsp->free_block_list, list) { |
| block_end = block->offset + block->size; |
| |
| /* ignore blocks with wrong type */ |
| if (block->type != ba->type) |
| continue; |
| |
| /* find block that holds section */ |
| if (ba->offset >= block->offset && end <= block_end) { |
| |
| /* add block */ |
| list_move(&block->list, &dsp->used_block_list); |
| list_add(&block->module_list, block_list); |
| dev_dbg(dsp->dev, "block allocated %d:%d at offset 0x%x\n", |
| block->type, block->index, block->offset); |
| return 0; |
| } |
| |
| /* does block span more than 1 section */ |
| if (ba->offset >= block->offset && ba->offset < block_end) { |
| |
| /* add block */ |
| list_move(&block->list, &dsp->used_block_list); |
| list_add(&block->module_list, block_list); |
| /* align ba to block boundary */ |
| ba->size -= block_end - ba->offset; |
| ba->offset = block_end; |
| |
| err = block_alloc_contiguous(dsp, ba, block_list); |
| if (err < 0) |
| return -ENOMEM; |
| |
| return 0; |
| } |
| } |
| |
| return -ENOMEM; |
| } |
| |
| /* Load fixed module data into DSP memory blocks */ |
| int sst_module_alloc_blocks(struct sst_module *module) |
| { |
| struct sst_dsp *dsp = module->dsp; |
| struct sst_fw *sst_fw = module->sst_fw; |
| struct sst_block_allocator ba; |
| int ret; |
| |
| ba.size = module->size; |
| ba.type = module->type; |
| ba.offset = module->offset; |
| |
| dev_dbg(dsp->dev, "block request 0x%x bytes at offset 0x%x type %d\n", |
| ba.size, ba.offset, ba.type); |
| |
| mutex_lock(&dsp->mutex); |
| |
| /* alloc blocks that includes this section */ |
| ret = block_alloc_fixed(dsp, &ba, &module->block_list); |
| if (ret < 0) { |
| dev_err(dsp->dev, |
| "error: no free blocks for section at offset 0x%x size 0x%x\n", |
| module->offset, module->size); |
| mutex_unlock(&dsp->mutex); |
| return -ENOMEM; |
| } |
| |
| /* prepare DSP blocks for module copy */ |
| ret = block_list_prepare(dsp, &module->block_list); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: fw module prepare failed\n"); |
| goto err; |
| } |
| |
| /* copy partial module data to blocks */ |
| if (dsp->fw_use_dma) { |
| ret = sst_dsp_dma_copyto(dsp, |
| dsp->addr.lpe_base + module->offset, |
| sst_fw->dmable_fw_paddr + module->data_offset, |
| module->size); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: module copy failed\n"); |
| goto err; |
| } |
| } else |
| sst_memcpy32(dsp->addr.lpe + module->offset, module->data, |
| module->size); |
| |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| |
| err: |
| block_list_remove(dsp, &module->block_list); |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_alloc_blocks); |
| |
| /* Unload entire module from DSP memory */ |
| int sst_module_free_blocks(struct sst_module *module) |
| { |
| struct sst_dsp *dsp = module->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| block_list_remove(dsp, &module->block_list); |
| mutex_unlock(&dsp->mutex); |
| return 0; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_free_blocks); |
| |
| int sst_module_runtime_alloc_blocks(struct sst_module_runtime *runtime, |
| int offset) |
| { |
| struct sst_dsp *dsp = runtime->dsp; |
| struct sst_module *module = runtime->module; |
| struct sst_block_allocator ba; |
| int ret; |
| |
| if (module->persistent_size == 0) |
| return 0; |
| |
| ba.size = module->persistent_size; |
| ba.type = SST_MEM_DRAM; |
| |
| mutex_lock(&dsp->mutex); |
| |
| /* do we need to allocate at a fixed address ? */ |
| if (offset != 0) { |
| |
| ba.offset = offset; |
| |
| dev_dbg(dsp->dev, "persistent fixed block request 0x%x bytes type %d offset 0x%x\n", |
| ba.size, ba.type, ba.offset); |
| |
| /* alloc blocks that includes this section */ |
| ret = block_alloc_fixed(dsp, &ba, &runtime->block_list); |
| |
| } else { |
| dev_dbg(dsp->dev, "persistent block request 0x%x bytes type %d\n", |
| ba.size, ba.type); |
| |
| /* alloc blocks that includes this section */ |
| ret = block_alloc(dsp, &ba, &runtime->block_list); |
| } |
| if (ret < 0) { |
| dev_err(dsp->dev, |
| "error: no free blocks for runtime module size 0x%x\n", |
| module->persistent_size); |
| mutex_unlock(&dsp->mutex); |
| return -ENOMEM; |
| } |
| runtime->persistent_offset = ba.offset; |
| |
| /* prepare DSP blocks for module copy */ |
| ret = block_list_prepare(dsp, &runtime->block_list); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: runtime block prepare failed\n"); |
| goto err; |
| } |
| |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| |
| err: |
| block_list_remove(dsp, &module->block_list); |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_alloc_blocks); |
| |
| int sst_module_runtime_free_blocks(struct sst_module_runtime *runtime) |
| { |
| struct sst_dsp *dsp = runtime->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| block_list_remove(dsp, &runtime->block_list); |
| mutex_unlock(&dsp->mutex); |
| return 0; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_free_blocks); |
| |
| int sst_module_runtime_save(struct sst_module_runtime *runtime, |
| struct sst_module_runtime_context *context) |
| { |
| struct sst_dsp *dsp = runtime->dsp; |
| struct sst_module *module = runtime->module; |
| int ret = 0; |
| |
| dev_dbg(dsp->dev, "saving runtime %d memory at 0x%x size 0x%x\n", |
| runtime->id, runtime->persistent_offset, |
| module->persistent_size); |
| |
| context->buffer = dma_alloc_coherent(dsp->dma_dev, |
| module->persistent_size, |
| &context->dma_buffer, GFP_DMA | GFP_KERNEL); |
| if (!context->buffer) { |
| dev_err(dsp->dev, "error: DMA context alloc failed\n"); |
| return -ENOMEM; |
| } |
| |
| mutex_lock(&dsp->mutex); |
| |
| if (dsp->fw_use_dma) { |
| |
| ret = sst_dsp_dma_get_channel(dsp, 0); |
| if (ret < 0) |
| goto err; |
| |
| ret = sst_dsp_dma_copyfrom(dsp, context->dma_buffer, |
| dsp->addr.lpe_base + runtime->persistent_offset, |
| module->persistent_size); |
| sst_dsp_dma_put_channel(dsp); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: context copy failed\n"); |
| goto err; |
| } |
| } else |
| sst_memcpy32(context->buffer, dsp->addr.lpe + |
| runtime->persistent_offset, |
| module->persistent_size); |
| |
| err: |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_save); |
| |
| int sst_module_runtime_restore(struct sst_module_runtime *runtime, |
| struct sst_module_runtime_context *context) |
| { |
| struct sst_dsp *dsp = runtime->dsp; |
| struct sst_module *module = runtime->module; |
| int ret = 0; |
| |
| dev_dbg(dsp->dev, "restoring runtime %d memory at 0x%x size 0x%x\n", |
| runtime->id, runtime->persistent_offset, |
| module->persistent_size); |
| |
| mutex_lock(&dsp->mutex); |
| |
| if (!context->buffer) { |
| dev_info(dsp->dev, "no context buffer need to restore!\n"); |
| goto err; |
| } |
| |
| if (dsp->fw_use_dma) { |
| |
| ret = sst_dsp_dma_get_channel(dsp, 0); |
| if (ret < 0) |
| goto err; |
| |
| ret = sst_dsp_dma_copyto(dsp, |
| dsp->addr.lpe_base + runtime->persistent_offset, |
| context->dma_buffer, module->persistent_size); |
| sst_dsp_dma_put_channel(dsp); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: module copy failed\n"); |
| goto err; |
| } |
| } else |
| sst_memcpy32(dsp->addr.lpe + runtime->persistent_offset, |
| context->buffer, module->persistent_size); |
| |
| dma_free_coherent(dsp->dma_dev, module->persistent_size, |
| context->buffer, context->dma_buffer); |
| context->buffer = NULL; |
| |
| err: |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_restore); |
| |
| /* register a DSP memory block for use with FW based modules */ |
| struct sst_mem_block *sst_mem_block_register(struct sst_dsp *dsp, u32 offset, |
| u32 size, enum sst_mem_type type, struct sst_block_ops *ops, u32 index, |
| void *private) |
| { |
| struct sst_mem_block *block; |
| |
| block = kzalloc(sizeof(*block), GFP_KERNEL); |
| if (block == NULL) |
| return NULL; |
| |
| block->offset = offset; |
| block->size = size; |
| block->index = index; |
| block->type = type; |
| block->dsp = dsp; |
| block->private = private; |
| block->ops = ops; |
| |
| mutex_lock(&dsp->mutex); |
| list_add(&block->list, &dsp->free_block_list); |
| mutex_unlock(&dsp->mutex); |
| |
| return block; |
| } |
| EXPORT_SYMBOL_GPL(sst_mem_block_register); |
| |
| /* unregister all DSP memory blocks */ |
| void sst_mem_block_unregister_all(struct sst_dsp *dsp) |
| { |
| struct sst_mem_block *block, *tmp; |
| |
| mutex_lock(&dsp->mutex); |
| |
| /* unregister used blocks */ |
| list_for_each_entry_safe(block, tmp, &dsp->used_block_list, list) { |
| list_del(&block->list); |
| kfree(block); |
| } |
| |
| /* unregister free blocks */ |
| list_for_each_entry_safe(block, tmp, &dsp->free_block_list, list) { |
| list_del(&block->list); |
| kfree(block); |
| } |
| |
| mutex_unlock(&dsp->mutex); |
| } |
| EXPORT_SYMBOL_GPL(sst_mem_block_unregister_all); |
| |
| /* allocate scratch buffer blocks */ |
| int sst_block_alloc_scratch(struct sst_dsp *dsp) |
| { |
| struct sst_module *module; |
| struct sst_block_allocator ba; |
| int ret; |
| |
| mutex_lock(&dsp->mutex); |
| |
| /* calculate required scratch size */ |
| dsp->scratch_size = 0; |
| list_for_each_entry(module, &dsp->module_list, list) { |
| dev_dbg(dsp->dev, "module %d scratch req 0x%x bytes\n", |
| module->id, module->scratch_size); |
| if (dsp->scratch_size < module->scratch_size) |
| dsp->scratch_size = module->scratch_size; |
| } |
| |
| dev_dbg(dsp->dev, "scratch buffer required is 0x%x bytes\n", |
| dsp->scratch_size); |
| |
| if (dsp->scratch_size == 0) { |
| dev_info(dsp->dev, "no modules need scratch buffer\n"); |
| mutex_unlock(&dsp->mutex); |
| return 0; |
| } |
| |
| /* allocate blocks for module scratch buffers */ |
| dev_dbg(dsp->dev, "allocating scratch blocks\n"); |
| |
| ba.size = dsp->scratch_size; |
| ba.type = SST_MEM_DRAM; |
| |
| /* do we need to allocate at fixed offset */ |
| if (dsp->scratch_offset != 0) { |
| |
| dev_dbg(dsp->dev, "block request 0x%x bytes type %d at 0x%x\n", |
| ba.size, ba.type, ba.offset); |
| |
| ba.offset = dsp->scratch_offset; |
| |
| /* alloc blocks that includes this section */ |
| ret = block_alloc_fixed(dsp, &ba, &dsp->scratch_block_list); |
| |
| } else { |
| dev_dbg(dsp->dev, "block request 0x%x bytes type %d\n", |
| ba.size, ba.type); |
| |
| ba.offset = 0; |
| ret = block_alloc(dsp, &ba, &dsp->scratch_block_list); |
| } |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: can't alloc scratch blocks\n"); |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| |
| ret = block_list_prepare(dsp, &dsp->scratch_block_list); |
| if (ret < 0) { |
| dev_err(dsp->dev, "error: scratch block prepare failed\n"); |
| mutex_unlock(&dsp->mutex); |
| return ret; |
| } |
| |
| /* assign the same offset of scratch to each module */ |
| dsp->scratch_offset = ba.offset; |
| mutex_unlock(&dsp->mutex); |
| return dsp->scratch_size; |
| } |
| EXPORT_SYMBOL_GPL(sst_block_alloc_scratch); |
| |
| /* free all scratch blocks */ |
| void sst_block_free_scratch(struct sst_dsp *dsp) |
| { |
| mutex_lock(&dsp->mutex); |
| block_list_remove(dsp, &dsp->scratch_block_list); |
| mutex_unlock(&dsp->mutex); |
| } |
| EXPORT_SYMBOL_GPL(sst_block_free_scratch); |
| |
| /* get a module from it's unique ID */ |
| struct sst_module *sst_module_get_from_id(struct sst_dsp *dsp, u32 id) |
| { |
| struct sst_module *module; |
| |
| mutex_lock(&dsp->mutex); |
| |
| list_for_each_entry(module, &dsp->module_list, list) { |
| if (module->id == id) { |
| mutex_unlock(&dsp->mutex); |
| return module; |
| } |
| } |
| |
| mutex_unlock(&dsp->mutex); |
| return NULL; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_get_from_id); |
| |
| struct sst_module_runtime *sst_module_runtime_get_from_id( |
| struct sst_module *module, u32 id) |
| { |
| struct sst_module_runtime *runtime; |
| struct sst_dsp *dsp = module->dsp; |
| |
| mutex_lock(&dsp->mutex); |
| |
| list_for_each_entry(runtime, &module->runtime_list, list) { |
| if (runtime->id == id) { |
| mutex_unlock(&dsp->mutex); |
| return runtime; |
| } |
| } |
| |
| mutex_unlock(&dsp->mutex); |
| return NULL; |
| } |
| EXPORT_SYMBOL_GPL(sst_module_runtime_get_from_id); |
| |
| /* returns block address in DSP address space */ |
| u32 sst_dsp_get_offset(struct sst_dsp *dsp, u32 offset, |
| enum sst_mem_type type) |
| { |
| switch (type) { |
| case SST_MEM_IRAM: |
| return offset - dsp->addr.iram_offset + |
| dsp->addr.dsp_iram_offset; |
| case SST_MEM_DRAM: |
| return offset - dsp->addr.dram_offset + |
| dsp->addr.dsp_dram_offset; |
| default: |
| return 0; |
| } |
| } |
| EXPORT_SYMBOL_GPL(sst_dsp_get_offset); |