Author | Tokens | Token Proportion | Commits | Commit Proportion |
---|---|---|---|---|
Kalle Valo | 4219 | 72.02% | 27 | 36.49% |
Vasanthakumar Thiagarajan | 1003 | 17.12% | 20 | 27.03% |
Raja Mani | 386 | 6.59% | 8 | 10.81% |
James Minor | 50 | 0.85% | 3 | 4.05% |
Chilam Ng | 35 | 0.60% | 2 | 2.70% |
Santosh Sajjan | 27 | 0.46% | 1 | 1.35% |
Ming Jiang | 23 | 0.39% | 1 | 1.35% |
Naveen Gangadharan | 22 | 0.38% | 1 | 1.35% |
Ray Chen | 17 | 0.29% | 1 | 1.35% |
Bala Shanmugam | 17 | 0.29% | 1 | 1.35% |
Adam Williamson | 11 | 0.19% | 1 | 1.35% |
Guy Chronister | 11 | 0.19% | 1 | 1.35% |
Srinivas Kandagatla | 11 | 0.19% | 1 | 1.35% |
Vivek Natarajan | 10 | 0.17% | 1 | 1.35% |
Geert Uytterhoeven | 5 | 0.09% | 1 | 1.35% |
Mohammed Shafi Shajakhan | 5 | 0.09% | 1 | 1.35% |
Paul Gortmaker | 3 | 0.05% | 1 | 1.35% |
Andi Kleen | 2 | 0.03% | 1 | 1.35% |
Wei Yongjun | 1 | 0.02% | 1 | 1.35% |
Total | 5858 | 74 |
/* * Copyright (c) 2004-2011 Atheros Communications Inc. * Copyright (c) 2011-2012 Qualcomm Atheros, Inc. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include <linux/module.h> #include <linux/mmc/card.h> #include <linux/mmc/mmc.h> #include <linux/mmc/host.h> #include <linux/mmc/sdio_func.h> #include <linux/mmc/sdio_ids.h> #include <linux/mmc/sdio.h> #include <linux/mmc/sd.h> #include "hif.h" #include "hif-ops.h" #include "target.h" #include "debug.h" #include "cfg80211.h" #include "trace.h" struct ath6kl_sdio { struct sdio_func *func; /* protects access to bus_req_freeq */ spinlock_t lock; /* free list */ struct list_head bus_req_freeq; /* available bus requests */ struct bus_request bus_req[BUS_REQUEST_MAX_NUM]; struct ath6kl *ar; u8 *dma_buffer; /* protects access to dma_buffer */ struct mutex dma_buffer_mutex; /* scatter request list head */ struct list_head scat_req; atomic_t irq_handling; wait_queue_head_t irq_wq; /* protects access to scat_req */ spinlock_t scat_lock; bool scatter_enabled; bool is_disabled; const struct sdio_device_id *id; struct work_struct wr_async_work; struct list_head wr_asyncq; /* protects access to wr_asyncq */ spinlock_t wr_async_lock; }; #define CMD53_ARG_READ 0 #define CMD53_ARG_WRITE 1 #define CMD53_ARG_BLOCK_BASIS 1 #define CMD53_ARG_FIXED_ADDRESS 0 #define CMD53_ARG_INCR_ADDRESS 1 static int ath6kl_sdio_config(struct ath6kl *ar); static inline struct ath6kl_sdio *ath6kl_sdio_priv(struct ath6kl *ar) { return ar->hif_priv; } /* * Macro to check if DMA buffer is WORD-aligned and DMA-able. * Most host controllers assume the buffer is DMA'able and will * bug-check otherwise (i.e. buffers on the stack). virt_addr_valid * check fails on stack memory. */ static inline bool buf_needs_bounce(u8 *buf) { return ((unsigned long) buf & 0x3) || !virt_addr_valid(buf); } static void ath6kl_sdio_set_mbox_info(struct ath6kl *ar) { struct ath6kl_mbox_info *mbox_info = &ar->mbox_info; /* EP1 has an extended range */ mbox_info->htc_addr = HIF_MBOX_BASE_ADDR; mbox_info->htc_ext_addr = HIF_MBOX0_EXT_BASE_ADDR; mbox_info->htc_ext_sz = HIF_MBOX0_EXT_WIDTH; mbox_info->block_size = HIF_MBOX_BLOCK_SIZE; mbox_info->gmbox_addr = HIF_GMBOX_BASE_ADDR; mbox_info->gmbox_sz = HIF_GMBOX_WIDTH; } static inline void ath6kl_sdio_set_cmd53_arg(u32 *arg, u8 rw, u8 func, u8 mode, u8 opcode, u32 addr, u16 blksz) { *arg = (((rw & 1) << 31) | ((func & 0x7) << 28) | ((mode & 1) << 27) | ((opcode & 1) << 26) | ((addr & 0x1FFFF) << 9) | (blksz & 0x1FF)); } static inline void ath6kl_sdio_set_cmd52_arg(u32 *arg, u8 write, u8 raw, unsigned int address, unsigned char val) { const u8 func = 0; *arg = ((write & 1) << 31) | ((func & 0x7) << 28) | ((raw & 1) << 27) | (1 << 26) | ((address & 0x1FFFF) << 9) | (1 << 8) | (val & 0xFF); } static int ath6kl_sdio_func0_cmd52_wr_byte(struct mmc_card *card, unsigned int address, unsigned char byte) { struct mmc_command io_cmd; memset(&io_cmd, 0, sizeof(io_cmd)); ath6kl_sdio_set_cmd52_arg(&io_cmd.arg, 1, 0, address, byte); io_cmd.opcode = SD_IO_RW_DIRECT; io_cmd.flags = MMC_RSP_R5 | MMC_CMD_AC; return mmc_wait_for_cmd(card->host, &io_cmd, 0); } static int ath6kl_sdio_io(struct sdio_func *func, u32 request, u32 addr, u8 *buf, u32 len) { int ret = 0; sdio_claim_host(func); if (request & HIF_WRITE) { /* FIXME: looks like ugly workaround for something */ if (addr >= HIF_MBOX_BASE_ADDR && addr <= HIF_MBOX_END_ADDR) addr += (HIF_MBOX_WIDTH - len); /* FIXME: this also looks like ugly workaround */ if (addr == HIF_MBOX0_EXT_BASE_ADDR) addr += HIF_MBOX0_EXT_WIDTH - len; if (request & HIF_FIXED_ADDRESS) ret = sdio_writesb(func, addr, buf, len); else ret = sdio_memcpy_toio(func, addr, buf, len); } else { if (request & HIF_FIXED_ADDRESS) ret = sdio_readsb(func, buf, addr, len); else ret = sdio_memcpy_fromio(func, buf, addr, len); } sdio_release_host(func); ath6kl_dbg(ATH6KL_DBG_SDIO, "%s addr 0x%x%s buf 0x%p len %d\n", request & HIF_WRITE ? "wr" : "rd", addr, request & HIF_FIXED_ADDRESS ? " (fixed)" : "", buf, len); ath6kl_dbg_dump(ATH6KL_DBG_SDIO_DUMP, NULL, "sdio ", buf, len); trace_ath6kl_sdio(addr, request, buf, len); return ret; } static struct bus_request *ath6kl_sdio_alloc_busreq(struct ath6kl_sdio *ar_sdio) { struct bus_request *bus_req; spin_lock_bh(&ar_sdio->lock); if (list_empty(&ar_sdio->bus_req_freeq)) { spin_unlock_bh(&ar_sdio->lock); return NULL; } bus_req = list_first_entry(&ar_sdio->bus_req_freeq, struct bus_request, list); list_del(&bus_req->list); spin_unlock_bh(&ar_sdio->lock); ath6kl_dbg(ATH6KL_DBG_SCATTER, "%s: bus request 0x%p\n", __func__, bus_req); return bus_req; } static void ath6kl_sdio_free_bus_req(struct ath6kl_sdio *ar_sdio, struct bus_request *bus_req) { ath6kl_dbg(ATH6KL_DBG_SCATTER, "%s: bus request 0x%p\n", __func__, bus_req); spin_lock_bh(&ar_sdio->lock); list_add_tail(&bus_req->list, &ar_sdio->bus_req_freeq); spin_unlock_bh(&ar_sdio->lock); } static void ath6kl_sdio_setup_scat_data(struct hif_scatter_req *scat_req, struct mmc_data *data) { struct scatterlist *sg; int i; data->blksz = HIF_MBOX_BLOCK_SIZE; data->blocks = scat_req->len / HIF_MBOX_BLOCK_SIZE; ath6kl_dbg(ATH6KL_DBG_SCATTER, "hif-scatter: (%s) addr: 0x%X, (block len: %d, block count: %d) , (tot:%d,sg:%d)\n", (scat_req->req & HIF_WRITE) ? "WR" : "RD", scat_req->addr, data->blksz, data->blocks, scat_req->len, scat_req->scat_entries); data->flags = (scat_req->req & HIF_WRITE) ? MMC_DATA_WRITE : MMC_DATA_READ; /* fill SG entries */ sg = scat_req->sgentries; sg_init_table(sg, scat_req->scat_entries); /* assemble SG list */ for (i = 0; i < scat_req->scat_entries; i++, sg++) { ath6kl_dbg(ATH6KL_DBG_SCATTER, "%d: addr:0x%p, len:%d\n", i, scat_req->scat_list[i].buf, scat_req->scat_list[i].len); sg_set_buf(sg, scat_req->scat_list[i].buf, scat_req->scat_list[i].len); } /* set scatter-gather table for request */ data->sg = scat_req->sgentries; data->sg_len = scat_req->scat_entries; } static int ath6kl_sdio_scat_rw(struct ath6kl_sdio *ar_sdio, struct bus_request *req) { struct mmc_request mmc_req; struct mmc_command cmd; struct mmc_data data; struct hif_scatter_req *scat_req; u8 opcode, rw; int status, len; scat_req = req->scat_req; if (scat_req->virt_scat) { len = scat_req->len; if (scat_req->req & HIF_BLOCK_BASIS) len = round_down(len, HIF_MBOX_BLOCK_SIZE); status = ath6kl_sdio_io(ar_sdio->func, scat_req->req, scat_req->addr, scat_req->virt_dma_buf, len); goto scat_complete; } memset(&mmc_req, 0, sizeof(struct mmc_request)); memset(&cmd, 0, sizeof(struct mmc_command)); memset(&data, 0, sizeof(struct mmc_data)); ath6kl_sdio_setup_scat_data(scat_req, &data); opcode = (scat_req->req & HIF_FIXED_ADDRESS) ? CMD53_ARG_FIXED_ADDRESS : CMD53_ARG_INCR_ADDRESS; rw = (scat_req->req & HIF_WRITE) ? CMD53_ARG_WRITE : CMD53_ARG_READ; /* Fixup the address so that the last byte will fall on MBOX EOM */ if (scat_req->req & HIF_WRITE) { if (scat_req->addr == HIF_MBOX_BASE_ADDR) scat_req->addr += HIF_MBOX_WIDTH - scat_req->len; else /* Uses extended address range */ scat_req->addr += HIF_MBOX0_EXT_WIDTH - scat_req->len; } /* set command argument */ ath6kl_sdio_set_cmd53_arg(&cmd.arg, rw, ar_sdio->func->num, CMD53_ARG_BLOCK_BASIS, opcode, scat_req->addr, data.blocks); cmd.opcode = SD_IO_RW_EXTENDED; cmd.flags = MMC_RSP_SPI_R5 | MMC_RSP_R5 | MMC_CMD_ADTC; mmc_req.cmd = &cmd; mmc_req.data = &data; sdio_claim_host(ar_sdio->func); mmc_set_data_timeout(&data, ar_sdio->func->card); trace_ath6kl_sdio_scat(scat_req->addr, scat_req->req, scat_req->len, scat_req->scat_entries, scat_req->scat_list); /* synchronous call to process request */ mmc_wait_for_req(ar_sdio->func->card->host, &mmc_req); sdio_release_host(ar_sdio->func); status = cmd.error ? cmd.error : data.error; scat_complete: scat_req->status = status; if (scat_req->status) ath6kl_err("Scatter write request failed:%d\n", scat_req->status); if (scat_req->req & HIF_ASYNCHRONOUS) scat_req->complete(ar_sdio->ar->htc_target, scat_req); return status; } static int ath6kl_sdio_alloc_prep_scat_req(struct ath6kl_sdio *ar_sdio, int n_scat_entry, int n_scat_req, bool virt_scat) { struct hif_scatter_req *s_req; struct bus_request *bus_req; int i, scat_req_sz, scat_list_sz, size; u8 *virt_buf; scat_list_sz = n_scat_entry * sizeof(struct hif_scatter_item); scat_req_sz = sizeof(*s_req) + scat_list_sz; if (!virt_scat) size = sizeof(struct scatterlist) * n_scat_entry; else size = 2 * L1_CACHE_BYTES + ATH6KL_MAX_TRANSFER_SIZE_PER_SCATTER; for (i = 0; i < n_scat_req; i++) { /* allocate the scatter request */ s_req = kzalloc(scat_req_sz, GFP_KERNEL); if (!s_req) return -ENOMEM; if (virt_scat) { virt_buf = kzalloc(size, GFP_KERNEL); if (!virt_buf) { kfree(s_req); return -ENOMEM; } s_req->virt_dma_buf = (u8 *)L1_CACHE_ALIGN((unsigned long)virt_buf); } else { /* allocate sglist */ s_req->sgentries = kzalloc(size, GFP_KERNEL); if (!s_req->sgentries) { kfree(s_req); return -ENOMEM; } } /* allocate a bus request for this scatter request */ bus_req = ath6kl_sdio_alloc_busreq(ar_sdio); if (!bus_req) { kfree(s_req->sgentries); kfree(s_req->virt_dma_buf); kfree(s_req); return -ENOMEM; } /* assign the scatter request to this bus request */ bus_req->scat_req = s_req; s_req->busrequest = bus_req; s_req->virt_scat = virt_scat; /* add it to the scatter pool */ hif_scatter_req_add(ar_sdio->ar, s_req); } return 0; } static int ath6kl_sdio_read_write_sync(struct ath6kl *ar, u32 addr, u8 *buf, u32 len, u32 request) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); u8 *tbuf = NULL; int ret; bool bounced = false; if (request & HIF_BLOCK_BASIS) len = round_down(len, HIF_MBOX_BLOCK_SIZE); if (buf_needs_bounce(buf)) { if (!ar_sdio->dma_buffer) return -ENOMEM; mutex_lock(&ar_sdio->dma_buffer_mutex); tbuf = ar_sdio->dma_buffer; if (request & HIF_WRITE) memcpy(tbuf, buf, len); bounced = true; } else { tbuf = buf; } ret = ath6kl_sdio_io(ar_sdio->func, request, addr, tbuf, len); if ((request & HIF_READ) && bounced) memcpy(buf, tbuf, len); if (bounced) mutex_unlock(&ar_sdio->dma_buffer_mutex); return ret; } static void __ath6kl_sdio_write_async(struct ath6kl_sdio *ar_sdio, struct bus_request *req) { if (req->scat_req) { ath6kl_sdio_scat_rw(ar_sdio, req); } else { void *context; int status; status = ath6kl_sdio_read_write_sync(ar_sdio->ar, req->address, req->buffer, req->length, req->request); context = req->packet; ath6kl_sdio_free_bus_req(ar_sdio, req); ath6kl_hif_rw_comp_handler(context, status); } } static void ath6kl_sdio_write_async_work(struct work_struct *work) { struct ath6kl_sdio *ar_sdio; struct bus_request *req, *tmp_req; ar_sdio = container_of(work, struct ath6kl_sdio, wr_async_work); spin_lock_bh(&ar_sdio->wr_async_lock); list_for_each_entry_safe(req, tmp_req, &ar_sdio->wr_asyncq, list) { list_del(&req->list); spin_unlock_bh(&ar_sdio->wr_async_lock); __ath6kl_sdio_write_async(ar_sdio, req); spin_lock_bh(&ar_sdio->wr_async_lock); } spin_unlock_bh(&ar_sdio->wr_async_lock); } static void ath6kl_sdio_irq_handler(struct sdio_func *func) { int status; struct ath6kl_sdio *ar_sdio; ath6kl_dbg(ATH6KL_DBG_SDIO, "irq\n"); ar_sdio = sdio_get_drvdata(func); atomic_set(&ar_sdio->irq_handling, 1); /* * Release the host during interrups so we can pick it back up when * we process commands. */ sdio_release_host(ar_sdio->func); status = ath6kl_hif_intr_bh_handler(ar_sdio->ar); sdio_claim_host(ar_sdio->func); atomic_set(&ar_sdio->irq_handling, 0); wake_up(&ar_sdio->irq_wq); WARN_ON(status && status != -ECANCELED); } static int ath6kl_sdio_power_on(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct sdio_func *func = ar_sdio->func; int ret = 0; if (!ar_sdio->is_disabled) return 0; ath6kl_dbg(ATH6KL_DBG_BOOT, "sdio power on\n"); sdio_claim_host(func); ret = sdio_enable_func(func); if (ret) { ath6kl_err("Unable to enable sdio func: %d)\n", ret); sdio_release_host(func); return ret; } sdio_release_host(func); /* * Wait for hardware to initialise. It should take a lot less than * 10 ms but let's be conservative here. */ msleep(10); ret = ath6kl_sdio_config(ar); if (ret) { ath6kl_err("Failed to config sdio: %d\n", ret); goto out; } ar_sdio->is_disabled = false; out: return ret; } static int ath6kl_sdio_power_off(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); int ret; if (ar_sdio->is_disabled) return 0; ath6kl_dbg(ATH6KL_DBG_BOOT, "sdio power off\n"); /* Disable the card */ sdio_claim_host(ar_sdio->func); ret = sdio_disable_func(ar_sdio->func); sdio_release_host(ar_sdio->func); if (ret) return ret; ar_sdio->is_disabled = true; return ret; } static int ath6kl_sdio_write_async(struct ath6kl *ar, u32 address, u8 *buffer, u32 length, u32 request, struct htc_packet *packet) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct bus_request *bus_req; bus_req = ath6kl_sdio_alloc_busreq(ar_sdio); if (WARN_ON_ONCE(!bus_req)) return -ENOMEM; bus_req->address = address; bus_req->buffer = buffer; bus_req->length = length; bus_req->request = request; bus_req->packet = packet; spin_lock_bh(&ar_sdio->wr_async_lock); list_add_tail(&bus_req->list, &ar_sdio->wr_asyncq); spin_unlock_bh(&ar_sdio->wr_async_lock); queue_work(ar->ath6kl_wq, &ar_sdio->wr_async_work); return 0; } static void ath6kl_sdio_irq_enable(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); int ret; sdio_claim_host(ar_sdio->func); /* Register the isr */ ret = sdio_claim_irq(ar_sdio->func, ath6kl_sdio_irq_handler); if (ret) ath6kl_err("Failed to claim sdio irq: %d\n", ret); sdio_release_host(ar_sdio->func); } static bool ath6kl_sdio_is_on_irq(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); return !atomic_read(&ar_sdio->irq_handling); } static void ath6kl_sdio_irq_disable(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); int ret; sdio_claim_host(ar_sdio->func); if (atomic_read(&ar_sdio->irq_handling)) { sdio_release_host(ar_sdio->func); ret = wait_event_interruptible(ar_sdio->irq_wq, ath6kl_sdio_is_on_irq(ar)); if (ret) return; sdio_claim_host(ar_sdio->func); } ret = sdio_release_irq(ar_sdio->func); if (ret) ath6kl_err("Failed to release sdio irq: %d\n", ret); sdio_release_host(ar_sdio->func); } static struct hif_scatter_req *ath6kl_sdio_scatter_req_get(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct hif_scatter_req *node = NULL; spin_lock_bh(&ar_sdio->scat_lock); if (!list_empty(&ar_sdio->scat_req)) { node = list_first_entry(&ar_sdio->scat_req, struct hif_scatter_req, list); list_del(&node->list); node->scat_q_depth = get_queue_depth(&ar_sdio->scat_req); } spin_unlock_bh(&ar_sdio->scat_lock); return node; } static void ath6kl_sdio_scatter_req_add(struct ath6kl *ar, struct hif_scatter_req *s_req) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); spin_lock_bh(&ar_sdio->scat_lock); list_add_tail(&s_req->list, &ar_sdio->scat_req); spin_unlock_bh(&ar_sdio->scat_lock); } /* scatter gather read write request */ static int ath6kl_sdio_async_rw_scatter(struct ath6kl *ar, struct hif_scatter_req *scat_req) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); u32 request = scat_req->req; int status = 0; if (!scat_req->len) return -EINVAL; ath6kl_dbg(ATH6KL_DBG_SCATTER, "hif-scatter: total len: %d scatter entries: %d\n", scat_req->len, scat_req->scat_entries); if (request & HIF_SYNCHRONOUS) { status = ath6kl_sdio_scat_rw(ar_sdio, scat_req->busrequest); } else { spin_lock_bh(&ar_sdio->wr_async_lock); list_add_tail(&scat_req->busrequest->list, &ar_sdio->wr_asyncq); spin_unlock_bh(&ar_sdio->wr_async_lock); queue_work(ar->ath6kl_wq, &ar_sdio->wr_async_work); } return status; } /* clean up scatter support */ static void ath6kl_sdio_cleanup_scatter(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct hif_scatter_req *s_req, *tmp_req; /* empty the free list */ spin_lock_bh(&ar_sdio->scat_lock); list_for_each_entry_safe(s_req, tmp_req, &ar_sdio->scat_req, list) { list_del(&s_req->list); spin_unlock_bh(&ar_sdio->scat_lock); /* * FIXME: should we also call completion handler with * ath6kl_hif_rw_comp_handler() with status -ECANCELED so * that the packet is properly freed? */ if (s_req->busrequest) { s_req->busrequest->scat_req = NULL; ath6kl_sdio_free_bus_req(ar_sdio, s_req->busrequest); } kfree(s_req->virt_dma_buf); kfree(s_req->sgentries); kfree(s_req); spin_lock_bh(&ar_sdio->scat_lock); } spin_unlock_bh(&ar_sdio->scat_lock); ar_sdio->scatter_enabled = false; } /* setup of HIF scatter resources */ static int ath6kl_sdio_enable_scatter(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct htc_target *target = ar->htc_target; int ret = 0; bool virt_scat = false; if (ar_sdio->scatter_enabled) return 0; ar_sdio->scatter_enabled = true; /* check if host supports scatter and it meets our requirements */ if (ar_sdio->func->card->host->max_segs < MAX_SCATTER_ENTRIES_PER_REQ) { ath6kl_err("host only supports scatter of :%d entries, need: %d\n", ar_sdio->func->card->host->max_segs, MAX_SCATTER_ENTRIES_PER_REQ); virt_scat = true; } if (!virt_scat) { ret = ath6kl_sdio_alloc_prep_scat_req(ar_sdio, MAX_SCATTER_ENTRIES_PER_REQ, MAX_SCATTER_REQUESTS, virt_scat); if (!ret) { ath6kl_dbg(ATH6KL_DBG_BOOT, "hif-scatter enabled requests %d entries %d\n", MAX_SCATTER_REQUESTS, MAX_SCATTER_ENTRIES_PER_REQ); target->max_scat_entries = MAX_SCATTER_ENTRIES_PER_REQ; target->max_xfer_szper_scatreq = MAX_SCATTER_REQ_TRANSFER_SIZE; } else { ath6kl_sdio_cleanup_scatter(ar); ath6kl_warn("hif scatter resource setup failed, trying virtual scatter method\n"); } } if (virt_scat || ret) { ret = ath6kl_sdio_alloc_prep_scat_req(ar_sdio, ATH6KL_SCATTER_ENTRIES_PER_REQ, ATH6KL_SCATTER_REQS, virt_scat); if (ret) { ath6kl_err("failed to alloc virtual scatter resources !\n"); ath6kl_sdio_cleanup_scatter(ar); return ret; } ath6kl_dbg(ATH6KL_DBG_BOOT, "virtual scatter enabled requests %d entries %d\n", ATH6KL_SCATTER_REQS, ATH6KL_SCATTER_ENTRIES_PER_REQ); target->max_scat_entries = ATH6KL_SCATTER_ENTRIES_PER_REQ; target->max_xfer_szper_scatreq = ATH6KL_MAX_TRANSFER_SIZE_PER_SCATTER; } return 0; } static int ath6kl_sdio_config(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct sdio_func *func = ar_sdio->func; int ret; sdio_claim_host(func); if ((ar_sdio->id->device & MANUFACTURER_ID_ATH6KL_BASE_MASK) >= MANUFACTURER_ID_AR6003_BASE) { /* enable 4-bit ASYNC interrupt on AR6003 or later */ ret = ath6kl_sdio_func0_cmd52_wr_byte(func->card, CCCR_SDIO_IRQ_MODE_REG, SDIO_IRQ_MODE_ASYNC_4BIT_IRQ); if (ret) { ath6kl_err("Failed to enable 4-bit async irq mode %d\n", ret); goto out; } ath6kl_dbg(ATH6KL_DBG_BOOT, "4-bit async irq mode enabled\n"); } /* give us some time to enable, in ms */ func->enable_timeout = 100; ret = sdio_set_block_size(func, HIF_MBOX_BLOCK_SIZE); if (ret) { ath6kl_err("Set sdio block size %d failed: %d)\n", HIF_MBOX_BLOCK_SIZE, ret); goto out; } out: sdio_release_host(func); return ret; } static int ath6kl_set_sdio_pm_caps(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct sdio_func *func = ar_sdio->func; mmc_pm_flag_t flags; int ret; flags = sdio_get_host_pm_caps(func); ath6kl_dbg(ATH6KL_DBG_SUSPEND, "sdio suspend pm_caps 0x%x\n", flags); if (!(flags & MMC_PM_WAKE_SDIO_IRQ) || !(flags & MMC_PM_KEEP_POWER)) return -EINVAL; ret = sdio_set_host_pm_flags(func, MMC_PM_KEEP_POWER); if (ret) { ath6kl_err("set sdio keep pwr flag failed: %d\n", ret); return ret; } /* sdio irq wakes up host */ ret = sdio_set_host_pm_flags(func, MMC_PM_WAKE_SDIO_IRQ); if (ret) ath6kl_err("set sdio wake irq flag failed: %d\n", ret); return ret; } static int ath6kl_sdio_suspend(struct ath6kl *ar, struct cfg80211_wowlan *wow) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct sdio_func *func = ar_sdio->func; mmc_pm_flag_t flags; bool try_deepsleep = false; int ret; if (ar->suspend_mode == WLAN_POWER_STATE_WOW || (!ar->suspend_mode && wow)) { ret = ath6kl_set_sdio_pm_caps(ar); if (ret) goto cut_pwr; ret = ath6kl_cfg80211_suspend(ar, ATH6KL_CFG_SUSPEND_WOW, wow); if (ret && ret != -ENOTCONN) ath6kl_err("wow suspend failed: %d\n", ret); if (ret && (!ar->wow_suspend_mode || ar->wow_suspend_mode == WLAN_POWER_STATE_DEEP_SLEEP)) try_deepsleep = true; else if (ret && ar->wow_suspend_mode == WLAN_POWER_STATE_CUT_PWR) goto cut_pwr; if (!ret) return 0; } if (ar->suspend_mode == WLAN_POWER_STATE_DEEP_SLEEP || !ar->suspend_mode || try_deepsleep) { flags = sdio_get_host_pm_caps(func); if (!(flags & MMC_PM_KEEP_POWER)) goto cut_pwr; ret = sdio_set_host_pm_flags(func, MMC_PM_KEEP_POWER); if (ret) goto cut_pwr; /* * Workaround to support Deep Sleep with MSM, set the host pm * flag as MMC_PM_WAKE_SDIO_IRQ to allow SDCC deiver to disable * the sdc2_clock and internally allows MSM to enter * TCXO shutdown properly. */ if ((flags & MMC_PM_WAKE_SDIO_IRQ)) { ret = sdio_set_host_pm_flags(func, MMC_PM_WAKE_SDIO_IRQ); if (ret) goto cut_pwr; } ret = ath6kl_cfg80211_suspend(ar, ATH6KL_CFG_SUSPEND_DEEPSLEEP, NULL); if (ret) goto cut_pwr; return 0; } cut_pwr: if (func->card && func->card->host) func->card->host->pm_flags &= ~MMC_PM_KEEP_POWER; return ath6kl_cfg80211_suspend(ar, ATH6KL_CFG_SUSPEND_CUTPOWER, NULL); } static int ath6kl_sdio_resume(struct ath6kl *ar) { switch (ar->state) { case ATH6KL_STATE_OFF: case ATH6KL_STATE_CUTPOWER: ath6kl_dbg(ATH6KL_DBG_SUSPEND, "sdio resume configuring sdio\n"); /* need to set sdio settings after power is cut from sdio */ ath6kl_sdio_config(ar); break; case ATH6KL_STATE_ON: break; case ATH6KL_STATE_DEEPSLEEP: break; case ATH6KL_STATE_WOW: break; case ATH6KL_STATE_SUSPENDING: break; case ATH6KL_STATE_RESUMING: break; case ATH6KL_STATE_RECOVERY: break; } ath6kl_cfg80211_resume(ar); return 0; } /* set the window address register (using 4-byte register access ). */ static int ath6kl_set_addrwin_reg(struct ath6kl *ar, u32 reg_addr, u32 addr) { int status; u8 addr_val[4]; s32 i; /* * Write bytes 1,2,3 of the register to set the upper address bytes, * the LSB is written last to initiate the access cycle */ for (i = 1; i <= 3; i++) { /* * Fill the buffer with the address byte value we want to * hit 4 times. */ memset(addr_val, ((u8 *)&addr)[i], 4); /* * Hit each byte of the register address with a 4-byte * write operation to the same address, this is a harmless * operation. */ status = ath6kl_sdio_read_write_sync(ar, reg_addr + i, addr_val, 4, HIF_WR_SYNC_BYTE_FIX); if (status) break; } if (status) { ath6kl_err("%s: failed to write initial bytes of 0x%x to window reg: 0x%X\n", __func__, addr, reg_addr); return status; } /* * Write the address register again, this time write the whole * 4-byte value. The effect here is that the LSB write causes the * cycle to start, the extra 3 byte write to bytes 1,2,3 has no * effect since we are writing the same values again */ status = ath6kl_sdio_read_write_sync(ar, reg_addr, (u8 *)(&addr), 4, HIF_WR_SYNC_BYTE_INC); if (status) { ath6kl_err("%s: failed to write 0x%x to window reg: 0x%X\n", __func__, addr, reg_addr); return status; } return 0; } static int ath6kl_sdio_diag_read32(struct ath6kl *ar, u32 address, u32 *data) { int status; /* set window register to start read cycle */ status = ath6kl_set_addrwin_reg(ar, WINDOW_READ_ADDR_ADDRESS, address); if (status) return status; /* read the data */ status = ath6kl_sdio_read_write_sync(ar, WINDOW_DATA_ADDRESS, (u8 *)data, sizeof(u32), HIF_RD_SYNC_BYTE_INC); if (status) { ath6kl_err("%s: failed to read from window data addr\n", __func__); return status; } return status; } static int ath6kl_sdio_diag_write32(struct ath6kl *ar, u32 address, __le32 data) { int status; u32 val = (__force u32) data; /* set write data */ status = ath6kl_sdio_read_write_sync(ar, WINDOW_DATA_ADDRESS, (u8 *) &val, sizeof(u32), HIF_WR_SYNC_BYTE_INC); if (status) { ath6kl_err("%s: failed to write 0x%x to window data addr\n", __func__, data); return status; } /* set window register, which starts the write cycle */ return ath6kl_set_addrwin_reg(ar, WINDOW_WRITE_ADDR_ADDRESS, address); } static int ath6kl_sdio_bmi_credits(struct ath6kl *ar) { u32 addr; unsigned long timeout; int ret; ar->bmi.cmd_credits = 0; /* Read the counter register to get the command credits */ addr = COUNT_DEC_ADDRESS + (HTC_MAILBOX_NUM_MAX + ENDPOINT1) * 4; timeout = jiffies + msecs_to_jiffies(BMI_COMMUNICATION_TIMEOUT); while (time_before(jiffies, timeout) && !ar->bmi.cmd_credits) { /* * Hit the credit counter with a 4-byte access, the first byte * read will hit the counter and cause a decrement, while the * remaining 3 bytes has no effect. The rationale behind this * is to make all HIF accesses 4-byte aligned. */ ret = ath6kl_sdio_read_write_sync(ar, addr, (u8 *)&ar->bmi.cmd_credits, 4, HIF_RD_SYNC_BYTE_INC); if (ret) { ath6kl_err("Unable to decrement the command credit count register: %d\n", ret); return ret; } /* The counter is only 8 bits. * Ignore anything in the upper 3 bytes */ ar->bmi.cmd_credits &= 0xFF; } if (!ar->bmi.cmd_credits) { ath6kl_err("bmi communication timeout\n"); return -ETIMEDOUT; } return 0; } static int ath6kl_bmi_get_rx_lkahd(struct ath6kl *ar) { unsigned long timeout; u32 rx_word = 0; int ret = 0; timeout = jiffies + msecs_to_jiffies(BMI_COMMUNICATION_TIMEOUT); while ((time_before(jiffies, timeout)) && !rx_word) { ret = ath6kl_sdio_read_write_sync(ar, RX_LOOKAHEAD_VALID_ADDRESS, (u8 *)&rx_word, sizeof(rx_word), HIF_RD_SYNC_BYTE_INC); if (ret) { ath6kl_err("unable to read RX_LOOKAHEAD_VALID\n"); return ret; } /* all we really want is one bit */ rx_word &= (1 << ENDPOINT1); } if (!rx_word) { ath6kl_err("bmi_recv_buf FIFO empty\n"); return -EINVAL; } return ret; } static int ath6kl_sdio_bmi_write(struct ath6kl *ar, u8 *buf, u32 len) { int ret; u32 addr; ret = ath6kl_sdio_bmi_credits(ar); if (ret) return ret; addr = ar->mbox_info.htc_addr; ret = ath6kl_sdio_read_write_sync(ar, addr, buf, len, HIF_WR_SYNC_BYTE_INC); if (ret) { ath6kl_err("unable to send the bmi data to the device\n"); return ret; } return 0; } static int ath6kl_sdio_bmi_read(struct ath6kl *ar, u8 *buf, u32 len) { int ret; u32 addr; /* * During normal bootup, small reads may be required. * Rather than issue an HIF Read and then wait as the Target * adds successive bytes to the FIFO, we wait here until * we know that response data is available. * * This allows us to cleanly timeout on an unexpected * Target failure rather than risk problems at the HIF level. * In particular, this avoids SDIO timeouts and possibly garbage * data on some host controllers. And on an interconnect * such as Compact Flash (as well as some SDIO masters) which * does not provide any indication on data timeout, it avoids * a potential hang or garbage response. * * Synchronization is more difficult for reads larger than the * size of the MBOX FIFO (128B), because the Target is unable * to push the 129th byte of data until AFTER the Host posts an * HIF Read and removes some FIFO data. So for large reads the * Host proceeds to post an HIF Read BEFORE all the data is * actually available to read. Fortunately, large BMI reads do * not occur in practice -- they're supported for debug/development. * * So Host/Target BMI synchronization is divided into these cases: * CASE 1: length < 4 * Should not happen * * CASE 2: 4 <= length <= 128 * Wait for first 4 bytes to be in FIFO * If CONSERVATIVE_BMI_READ is enabled, also wait for * a BMI command credit, which indicates that the ENTIRE * response is available in the the FIFO * * CASE 3: length > 128 * Wait for the first 4 bytes to be in FIFO * * For most uses, a small timeout should be sufficient and we will * usually see a response quickly; but there may be some unusual * (debug) cases of BMI_EXECUTE where we want an larger timeout. * For now, we use an unbounded busy loop while waiting for * BMI_EXECUTE. * * If BMI_EXECUTE ever needs to support longer-latency execution, * especially in production, this code needs to be enhanced to sleep * and yield. Also note that BMI_COMMUNICATION_TIMEOUT is currently * a function of Host processor speed. */ if (len >= 4) { /* NB: Currently, always true */ ret = ath6kl_bmi_get_rx_lkahd(ar); if (ret) return ret; } addr = ar->mbox_info.htc_addr; ret = ath6kl_sdio_read_write_sync(ar, addr, buf, len, HIF_RD_SYNC_BYTE_INC); if (ret) { ath6kl_err("Unable to read the bmi data from the device: %d\n", ret); return ret; } return 0; } static void ath6kl_sdio_stop(struct ath6kl *ar) { struct ath6kl_sdio *ar_sdio = ath6kl_sdio_priv(ar); struct bus_request *req, *tmp_req; void *context; /* FIXME: make sure that wq is not queued again */ cancel_work_sync(&ar_sdio->wr_async_work); spin_lock_bh(&ar_sdio->wr_async_lock); list_for_each_entry_safe(req, tmp_req, &ar_sdio->wr_asyncq, list) { list_del(&req->list); if (req->scat_req) { /* this is a scatter gather request */ req->scat_req->status = -ECANCELED; req->scat_req->complete(ar_sdio->ar->htc_target, req->scat_req); } else { context = req->packet; ath6kl_sdio_free_bus_req(ar_sdio, req); ath6kl_hif_rw_comp_handler(context, -ECANCELED); } } spin_unlock_bh(&ar_sdio->wr_async_lock); WARN_ON(get_queue_depth(&ar_sdio->scat_req) != 4); } static const struct ath6kl_hif_ops ath6kl_sdio_ops = { .read_write_sync = ath6kl_sdio_read_write_sync, .write_async = ath6kl_sdio_write_async, .irq_enable = ath6kl_sdio_irq_enable, .irq_disable = ath6kl_sdio_irq_disable, .scatter_req_get = ath6kl_sdio_scatter_req_get, .scatter_req_add = ath6kl_sdio_scatter_req_add, .enable_scatter = ath6kl_sdio_enable_scatter, .scat_req_rw = ath6kl_sdio_async_rw_scatter, .cleanup_scatter = ath6kl_sdio_cleanup_scatter, .suspend = ath6kl_sdio_suspend, .resume = ath6kl_sdio_resume, .diag_read32 = ath6kl_sdio_diag_read32, .diag_write32 = ath6kl_sdio_diag_write32, .bmi_read = ath6kl_sdio_bmi_read, .bmi_write = ath6kl_sdio_bmi_write, .power_on = ath6kl_sdio_power_on, .power_off = ath6kl_sdio_power_off, .stop = ath6kl_sdio_stop, }; #ifdef CONFIG_PM_SLEEP /* * Empty handlers so that mmc subsystem doesn't remove us entirely during * suspend. We instead follow cfg80211 suspend/resume handlers. */ static int ath6kl_sdio_pm_suspend(struct device *device) { ath6kl_dbg(ATH6KL_DBG_SUSPEND, "sdio pm suspend\n"); return 0; } static int ath6kl_sdio_pm_resume(struct device *device) { ath6kl_dbg(ATH6KL_DBG_SUSPEND, "sdio pm resume\n"); return 0; } static SIMPLE_DEV_PM_OPS(ath6kl_sdio_pm_ops, ath6kl_sdio_pm_suspend, ath6kl_sdio_pm_resume); #define ATH6KL_SDIO_PM_OPS (&ath6kl_sdio_pm_ops) #else #define ATH6KL_SDIO_PM_OPS NULL #endif /* CONFIG_PM_SLEEP */ static int ath6kl_sdio_probe(struct sdio_func *func, const struct sdio_device_id *id) { int ret; struct ath6kl_sdio *ar_sdio; struct ath6kl *ar; int count; ath6kl_dbg(ATH6KL_DBG_BOOT, "sdio new func %d vendor 0x%x device 0x%x block 0x%x/0x%x\n", func->num, func->vendor, func->device, func->max_blksize, func->cur_blksize); ar_sdio = kzalloc(sizeof(struct ath6kl_sdio), GFP_KERNEL); if (!ar_sdio) return -ENOMEM; ar_sdio->dma_buffer = kzalloc(HIF_DMA_BUFFER_SIZE, GFP_KERNEL); if (!ar_sdio->dma_buffer) { ret = -ENOMEM; goto err_hif; } ar_sdio->func = func; sdio_set_drvdata(func, ar_sdio); ar_sdio->id = id; ar_sdio->is_disabled = true; spin_lock_init(&ar_sdio->lock); spin_lock_init(&ar_sdio->scat_lock); spin_lock_init(&ar_sdio->wr_async_lock); mutex_init(&ar_sdio->dma_buffer_mutex); INIT_LIST_HEAD(&ar_sdio->scat_req); INIT_LIST_HEAD(&ar_sdio->bus_req_freeq); INIT_LIST_HEAD(&ar_sdio->wr_asyncq); INIT_WORK(&ar_sdio->wr_async_work, ath6kl_sdio_write_async_work); init_waitqueue_head(&ar_sdio->irq_wq); for (count = 0; count < BUS_REQUEST_MAX_NUM; count++) ath6kl_sdio_free_bus_req(ar_sdio, &ar_sdio->bus_req[count]); ar = ath6kl_core_create(&ar_sdio->func->dev); if (!ar) { ath6kl_err("Failed to alloc ath6kl core\n"); ret = -ENOMEM; goto err_dma; } ar_sdio->ar = ar; ar->hif_type = ATH6KL_HIF_TYPE_SDIO; ar->hif_priv = ar_sdio; ar->hif_ops = &ath6kl_sdio_ops; ar->bmi.max_data_size = 256; ath6kl_sdio_set_mbox_info(ar); ret = ath6kl_sdio_config(ar); if (ret) { ath6kl_err("Failed to config sdio: %d\n", ret); goto err_core_alloc; } ret = ath6kl_core_init(ar, ATH6KL_HTC_TYPE_MBOX); if (ret) { ath6kl_err("Failed to init ath6kl core\n"); goto err_core_alloc; } return ret; err_core_alloc: ath6kl_core_destroy(ar_sdio->ar); err_dma: kfree(ar_sdio->dma_buffer); err_hif: kfree(ar_sdio); return ret; } static void ath6kl_sdio_remove(struct sdio_func *func) { struct ath6kl_sdio *ar_sdio; ath6kl_dbg(ATH6KL_DBG_BOOT, "sdio removed func %d vendor 0x%x device 0x%x\n", func->num, func->vendor, func->device); ar_sdio = sdio_get_drvdata(func); ath6kl_stop_txrx(ar_sdio->ar); cancel_work_sync(&ar_sdio->wr_async_work); ath6kl_core_cleanup(ar_sdio->ar); ath6kl_core_destroy(ar_sdio->ar); kfree(ar_sdio->dma_buffer); kfree(ar_sdio); } static const struct sdio_device_id ath6kl_sdio_devices[] = { {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6003_BASE | 0x0))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6003_BASE | 0x1))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6004_BASE | 0x0))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6004_BASE | 0x1))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6004_BASE | 0x2))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6004_BASE | 0x18))}, {SDIO_DEVICE(MANUFACTURER_CODE, (MANUFACTURER_ID_AR6004_BASE | 0x19))}, {}, }; MODULE_DEVICE_TABLE(sdio, ath6kl_sdio_devices); static struct sdio_driver ath6kl_sdio_driver = { .name = "ath6kl_sdio", .id_table = ath6kl_sdio_devices, .probe = ath6kl_sdio_probe, .remove = ath6kl_sdio_remove, .drv.pm = ATH6KL_SDIO_PM_OPS, }; static int __init ath6kl_sdio_init(void) { int ret; ret = sdio_register_driver(&ath6kl_sdio_driver); if (ret) ath6kl_err("sdio driver registration failed: %d\n", ret); return ret; } static void __exit ath6kl_sdio_exit(void) { sdio_unregister_driver(&ath6kl_sdio_driver); } module_init(ath6kl_sdio_init); module_exit(ath6kl_sdio_exit); MODULE_AUTHOR("Atheros Communications, Inc."); MODULE_DESCRIPTION("Driver support for Atheros AR600x SDIO devices"); MODULE_LICENSE("Dual BSD/GPL"); MODULE_FIRMWARE(AR6003_HW_2_0_FW_DIR "/" AR6003_HW_2_0_OTP_FILE); MODULE_FIRMWARE(AR6003_HW_2_0_FW_DIR "/" AR6003_HW_2_0_FIRMWARE_FILE); MODULE_FIRMWARE(AR6003_HW_2_0_FW_DIR "/" AR6003_HW_2_0_PATCH_FILE); MODULE_FIRMWARE(AR6003_HW_2_0_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6003_HW_2_0_DEFAULT_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6003_HW_2_1_1_FW_DIR "/" AR6003_HW_2_1_1_OTP_FILE); MODULE_FIRMWARE(AR6003_HW_2_1_1_FW_DIR "/" AR6003_HW_2_1_1_FIRMWARE_FILE); MODULE_FIRMWARE(AR6003_HW_2_1_1_FW_DIR "/" AR6003_HW_2_1_1_PATCH_FILE); MODULE_FIRMWARE(AR6003_HW_2_1_1_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6003_HW_2_1_1_DEFAULT_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_0_FW_DIR "/" AR6004_HW_1_0_FIRMWARE_FILE); MODULE_FIRMWARE(AR6004_HW_1_0_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_0_DEFAULT_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_1_FW_DIR "/" AR6004_HW_1_1_FIRMWARE_FILE); MODULE_FIRMWARE(AR6004_HW_1_1_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_1_DEFAULT_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_2_FW_DIR "/" AR6004_HW_1_2_FIRMWARE_FILE); MODULE_FIRMWARE(AR6004_HW_1_2_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_2_DEFAULT_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_3_FW_DIR "/" AR6004_HW_1_3_FIRMWARE_FILE); MODULE_FIRMWARE(AR6004_HW_1_3_BOARD_DATA_FILE); MODULE_FIRMWARE(AR6004_HW_1_3_DEFAULT_BOARD_DATA_FILE);
Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.
Created with Cregit http://github.com/cregit/cregit
Version 2.0-RC1