blob: c84db454bfd4d70e79f7f35cc2a98d35c62b0eed [file] [log] [blame]
// SPDX-License-Identifier: GPL-2.0+
/*
* Multicore Navigator driver for TI Keystone 2 devices.
*
* (C) Copyright 2012-2014
* Texas Instruments Incorporated, <www.ti.com>
*/
#include <asm/io.h>
#include <asm/ti-common/keystone_nav.h>
#include <linux/delay.h>
#include <linux/string.h>
struct qm_config qm_memmap = {
.stat_cfg = KS2_QM_QUEUE_STATUS_BASE,
.queue = (void *)KS2_QM_MANAGER_QUEUES_BASE,
.mngr_vbusm = KS2_QM_BASE_ADDRESS,
.i_lram = KS2_QM_LINK_RAM_BASE,
.proxy = (void *)KS2_QM_MANAGER_Q_PROXY_BASE,
.status_ram = KS2_QM_STATUS_RAM_BASE,
.mngr_cfg = (void *)KS2_QM_CONF_BASE,
.intd_cfg = KS2_QM_INTD_CONF_BASE,
.desc_mem = (void *)KS2_QM_DESC_SETUP_BASE,
.region_num = KS2_QM_REGION_NUM,
.pdsp_cmd = KS2_QM_PDSP1_CMD_BASE,
.pdsp_ctl = KS2_QM_PDSP1_CTRL_BASE,
.pdsp_iram = KS2_QM_PDSP1_IRAM_BASE,
.qpool_num = KS2_QM_QPOOL_NUM,
};
/*
* We are going to use only one type of descriptors - host packet
* descriptors. We staticaly allocate memory for them here
*/
struct qm_host_desc desc_pool[HDESC_NUM] __aligned(sizeof(struct qm_host_desc));
static struct qm_config *qm_cfg;
inline int num_of_desc_to_reg(int num_descr)
{
int j, num;
for (j = 0, num = 32; j < 15; j++, num *= 2) {
if (num_descr <= num)
return j;
}
return 15;
}
int _qm_init(struct qm_config *cfg)
{
u32 j;
qm_cfg = cfg;
qm_cfg->mngr_cfg->link_ram_base0 = qm_cfg->i_lram;
qm_cfg->mngr_cfg->link_ram_size0 = HDESC_NUM * 8 - 1;
qm_cfg->mngr_cfg->link_ram_base1 = 0;
qm_cfg->mngr_cfg->link_ram_size1 = 0;
qm_cfg->mngr_cfg->link_ram_base2 = 0;
qm_cfg->desc_mem[0].base_addr = (u32)desc_pool;
qm_cfg->desc_mem[0].start_idx = 0;
qm_cfg->desc_mem[0].desc_reg_size =
(((sizeof(struct qm_host_desc) >> 4) - 1) << 16) |
num_of_desc_to_reg(HDESC_NUM);
memset(desc_pool, 0, sizeof(desc_pool));
for (j = 0; j < HDESC_NUM; j++)
qm_push(&desc_pool[j], qm_cfg->qpool_num);
return QM_OK;
}
int qm_init(void)
{
return _qm_init(&qm_memmap);
}
void qm_close(void)
{
u32 j;
queue_close(qm_cfg->qpool_num);
qm_cfg->mngr_cfg->link_ram_base0 = 0;
qm_cfg->mngr_cfg->link_ram_size0 = 0;
qm_cfg->mngr_cfg->link_ram_base1 = 0;
qm_cfg->mngr_cfg->link_ram_size1 = 0;
qm_cfg->mngr_cfg->link_ram_base2 = 0;
for (j = 0; j < qm_cfg->region_num; j++) {
qm_cfg->desc_mem[j].base_addr = 0;
qm_cfg->desc_mem[j].start_idx = 0;
qm_cfg->desc_mem[j].desc_reg_size = 0;
}
qm_cfg = NULL;
}
void qm_push(struct qm_host_desc *hd, u32 qnum)
{
u32 regd;
cpu_to_bus((u32 *)hd, sizeof(struct qm_host_desc)/4);
regd = (u32)hd | ((sizeof(struct qm_host_desc) >> 4) - 1);
writel(regd, &qm_cfg->queue[qnum].ptr_size_thresh);
}
void qm_buff_push(struct qm_host_desc *hd, u32 qnum,
void *buff_ptr, u32 buff_len)
{
hd->orig_buff_len = buff_len;
hd->buff_len = buff_len;
hd->orig_buff_ptr = (u32)buff_ptr;
hd->buff_ptr = (u32)buff_ptr;
qm_push(hd, qnum);
}
struct qm_host_desc *qm_pop(u32 qnum)
{
u32 uhd;
uhd = readl(&qm_cfg->queue[qnum].ptr_size_thresh) & ~0xf;
if (uhd)
cpu_to_bus((u32 *)uhd, sizeof(struct qm_host_desc)/4);
return (struct qm_host_desc *)uhd;
}
struct qm_host_desc *qm_pop_from_free_pool(void)
{
return qm_pop(qm_cfg->qpool_num);
}
void queue_close(u32 qnum)
{
struct qm_host_desc *hd;
while ((hd = qm_pop(qnum)))
;
}
/**
* DMA API
*/
static int ksnav_rx_disable(struct pktdma_cfg *pktdma)
{
u32 j, v, k;
for (j = 0; j < pktdma->rx_ch_num; j++) {
v = readl(&pktdma->rx_ch[j].cfg_a);
if (!(v & CPDMA_CHAN_A_ENABLE))
continue;
writel(v | CPDMA_CHAN_A_TDOWN, &pktdma->rx_ch[j].cfg_a);
for (k = 0; k < TDOWN_TIMEOUT_COUNT; k++) {
udelay(100);
v = readl(&pktdma->rx_ch[j].cfg_a);
if (!(v & CPDMA_CHAN_A_ENABLE))
continue;
}
/* TODO: teardown error on if TDOWN_TIMEOUT_COUNT is reached */
}
/* Clear all of the flow registers */
for (j = 0; j < pktdma->rx_flow_num; j++) {
writel(0, &pktdma->rx_flows[j].control);
writel(0, &pktdma->rx_flows[j].tags);
writel(0, &pktdma->rx_flows[j].tag_sel);
writel(0, &pktdma->rx_flows[j].fdq_sel[0]);
writel(0, &pktdma->rx_flows[j].fdq_sel[1]);
writel(0, &pktdma->rx_flows[j].thresh[0]);
writel(0, &pktdma->rx_flows[j].thresh[1]);
writel(0, &pktdma->rx_flows[j].thresh[2]);
}
return QM_OK;
}
static int ksnav_tx_disable(struct pktdma_cfg *pktdma)
{
u32 j, v, k;
for (j = 0; j < pktdma->tx_ch_num; j++) {
v = readl(&pktdma->tx_ch[j].cfg_a);
if (!(v & CPDMA_CHAN_A_ENABLE))
continue;
writel(v | CPDMA_CHAN_A_TDOWN, &pktdma->tx_ch[j].cfg_a);
for (k = 0; k < TDOWN_TIMEOUT_COUNT; k++) {
udelay(100);
v = readl(&pktdma->tx_ch[j].cfg_a);
if (!(v & CPDMA_CHAN_A_ENABLE))
continue;
}
/* TODO: teardown error on if TDOWN_TIMEOUT_COUNT is reached */
}
return QM_OK;
}
int ksnav_init(struct pktdma_cfg *pktdma, struct rx_buff_desc *rx_buffers)
{
u32 j, v;
struct qm_host_desc *hd;
u8 *rx_ptr;
if (pktdma == NULL || rx_buffers == NULL ||
rx_buffers->buff_ptr == NULL || qm_cfg == NULL)
return QM_ERR;
pktdma->rx_flow = rx_buffers->rx_flow;
/* init rx queue */
rx_ptr = rx_buffers->buff_ptr;
for (j = 0; j < rx_buffers->num_buffs; j++) {
hd = qm_pop(qm_cfg->qpool_num);
if (hd == NULL)
return QM_ERR;
qm_buff_push(hd, pktdma->rx_free_q,
rx_ptr, rx_buffers->buff_len);
rx_ptr += rx_buffers->buff_len;
}
ksnav_rx_disable(pktdma);
/* configure rx channels */
v = CPDMA_REG_VAL_MAKE_RX_FLOW_A(1, 1, 0, 0, 0, 0, 0, pktdma->rx_rcv_q);
writel(v, &pktdma->rx_flows[pktdma->rx_flow].control);
writel(0, &pktdma->rx_flows[pktdma->rx_flow].tags);
writel(0, &pktdma->rx_flows[pktdma->rx_flow].tag_sel);
v = CPDMA_REG_VAL_MAKE_RX_FLOW_D(0, pktdma->rx_free_q, 0,
pktdma->rx_free_q);
writel(v, &pktdma->rx_flows[pktdma->rx_flow].fdq_sel[0]);
writel(v, &pktdma->rx_flows[pktdma->rx_flow].fdq_sel[1]);
writel(0, &pktdma->rx_flows[pktdma->rx_flow].thresh[0]);
writel(0, &pktdma->rx_flows[pktdma->rx_flow].thresh[1]);
writel(0, &pktdma->rx_flows[pktdma->rx_flow].thresh[2]);
for (j = 0; j < pktdma->rx_ch_num; j++)
writel(CPDMA_CHAN_A_ENABLE, &pktdma->rx_ch[j].cfg_a);
/* configure tx channels */
/* Disable loopback in the tx direction */
writel(0, &pktdma->global->emulation_control);
/* Set QM base address, only for K2x devices */
writel(KS2_QM_BASE_ADDRESS, &pktdma->global->qm_base_addr[0]);
/* Enable all channels. The current state isn't important */
for (j = 0; j < pktdma->tx_ch_num; j++) {
writel(0, &pktdma->tx_ch[j].cfg_b);
writel(CPDMA_CHAN_A_ENABLE, &pktdma->tx_ch[j].cfg_a);
}
return QM_OK;
}
int ksnav_close(struct pktdma_cfg *pktdma)
{
if (!pktdma)
return QM_ERR;
ksnav_tx_disable(pktdma);
ksnav_rx_disable(pktdma);
queue_close(pktdma->rx_free_q);
queue_close(pktdma->rx_rcv_q);
queue_close(pktdma->tx_snd_q);
return QM_OK;
}
int ksnav_send(struct pktdma_cfg *pktdma, u32 *pkt, int num_bytes, u32 swinfo2)
{
struct qm_host_desc *hd;
hd = qm_pop(qm_cfg->qpool_num);
if (hd == NULL)
return QM_ERR;
hd->desc_info = num_bytes;
hd->swinfo[2] = swinfo2;
hd->packet_info = qm_cfg->qpool_num;
qm_buff_push(hd, pktdma->tx_snd_q, pkt, num_bytes);
return QM_OK;
}
void *ksnav_recv(struct pktdma_cfg *pktdma, u32 **pkt, int *num_bytes)
{
struct qm_host_desc *hd;
hd = qm_pop(pktdma->rx_rcv_q);
if (!hd)
return NULL;
*pkt = (u32 *)hd->buff_ptr;
*num_bytes = hd->desc_info & 0x3fffff;
return hd;
}
void ksnav_release_rxhd(struct pktdma_cfg *pktdma, void *hd)
{
struct qm_host_desc *_hd = (struct qm_host_desc *)hd;
_hd->buff_len = _hd->orig_buff_len;
_hd->buff_ptr = _hd->orig_buff_ptr;
qm_push(_hd, pktdma->rx_free_q);
}