[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 4/8] hw/net:ftgmac100: update TX and RX packet buffers address to
From: |
Cédric Le Goater |
Subject: |
[PULL 4/8] hw/net:ftgmac100: update TX and RX packet buffers address to 64 bits |
Date: |
Tue, 9 Jul 2024 13:52:24 +0200 |
From: Jamin Lin <jamin_lin@aspeedtech.com>
ASPEED AST2700 SOC is a 64 bits quad core CPUs (Cortex-a35)
And the base address of dram is "0x4 00000000" which
is 64bits address.
It have "TXDES 2" and "RXDES 2" to save the high part
physical address of packet buffer.
Ex: TX packet buffer address [34:0]
The "TXDES 2" bits [18:16] which corresponds the bits [34:32]
of the 64 bits address of the TX packet buffer address
and "TXDES 3" bits [31:0] which corresponds the bits [31:0]
of the 64 bits address of the TX packet buffer address.
Update TX and RX packet buffers address type to
64 bits for dram 64 bits address DMA support.
Signed-off-by: Jamin Lin <jamin_lin@aspeedtech.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
---
hw/net/ftgmac100.c | 21 ++++++++++++++++++---
1 file changed, 18 insertions(+), 3 deletions(-)
diff --git a/hw/net/ftgmac100.c b/hw/net/ftgmac100.c
index 68956aeb94ae..80f9cd56d53f 100644
--- a/hw/net/ftgmac100.c
+++ b/hw/net/ftgmac100.c
@@ -175,6 +175,8 @@
#define FTGMAC100_TXDES1_TX2FIC (1 << 30)
#define FTGMAC100_TXDES1_TXIC (1 << 31)
+#define FTGMAC100_TXDES2_TXBUF_BADR_HI(x) (((x) >> 16) & 0x7)
+
/*
* Receive descriptor
*/
@@ -208,13 +210,15 @@
#define FTGMAC100_RXDES1_UDP_CHKSUM_ERR (1 << 26)
#define FTGMAC100_RXDES1_IP_CHKSUM_ERR (1 << 27)
+#define FTGMAC100_RXDES2_RXBUF_BADR_HI(x) (((x) >> 16) & 0x7)
+
/*
* Receive and transmit Buffer Descriptor
*/
typedef struct {
uint32_t des0;
uint32_t des1;
- uint32_t des2; /* not used by HW */
+ uint32_t des2; /* used by HW 64 bits DMA */
uint32_t des3;
} FTGMAC100Desc;
@@ -531,6 +535,7 @@ static void ftgmac100_do_tx(FTGMAC100State *s, uint64_t
tx_ring,
int frame_size = 0;
uint8_t *ptr = s->frame;
uint64_t addr = tx_descriptor;
+ uint64_t buf_addr = 0;
uint32_t flags = 0;
while (1) {
@@ -569,7 +574,12 @@ static void ftgmac100_do_tx(FTGMAC100State *s, uint64_t
tx_ring,
len = sizeof(s->frame) - frame_size;
}
- if (dma_memory_read(&address_space_memory, bd.des3,
+ buf_addr = bd.des3;
+ if (s->dma64) {
+ buf_addr = deposit64(buf_addr, 32, 32,
+ FTGMAC100_TXDES2_TXBUF_BADR_HI(bd.des2));
+ }
+ if (dma_memory_read(&address_space_memory, buf_addr,
ptr, len, MEMTXATTRS_UNSPECIFIED)) {
qemu_log_mask(LOG_GUEST_ERROR, "%s: failed to read packet @
0x%x\n",
__func__, bd.des3);
@@ -1022,7 +1032,7 @@ static ssize_t ftgmac100_receive(NetClientState *nc,
const uint8_t *buf,
uint32_t flags = 0;
uint64_t addr;
uint32_t crc;
- uint32_t buf_addr;
+ uint64_t buf_addr = 0;
uint8_t *crc_ptr;
uint32_t buf_len;
size_t size = len;
@@ -1087,7 +1097,12 @@ static ssize_t ftgmac100_receive(NetClientState *nc,
const uint8_t *buf,
if (size < 4) {
buf_len += size - 4;
}
+
buf_addr = bd.des3;
+ if (s->dma64) {
+ buf_addr = deposit64(buf_addr, 32, 32,
+ FTGMAC100_RXDES2_RXBUF_BADR_HI(bd.des2));
+ }
if (first && proto == ETH_P_VLAN && buf_len >= 18) {
bd.des1 = lduw_be_p(buf + 14) | FTGMAC100_RXDES1_VLANTAG_AVAIL;
--
2.45.2
- [PULL 0/8] aspeed queue, Cédric Le Goater, 2024/07/09
- [PULL 1/8] hw/net:ftgmac100: update memory region size to 64KB, Cédric Le Goater, 2024/07/09
- [PULL 2/8] hw/net:ftgmac100: update ring base address to 64 bits, Cédric Le Goater, 2024/07/09
- [PULL 5/8] aspeed/soc: set dma64 property for AST2700 ftgmac100, Cédric Le Goater, 2024/07/09
- [PULL 4/8] hw/net:ftgmac100: update TX and RX packet buffers address to 64 bits,
Cédric Le Goater <=
- [PULL 7/8] machine_aspeed.py: update to test ASPEED OpenBMC SDK v09.02 for AST2700, Cédric Le Goater, 2024/07/09
- [PULL 3/8] hw/net:ftgmac100: introduce TX and RX ring base address high registers to support 64 bits, Cédric Le Goater, 2024/07/09
- [PULL 6/8] hw/block: m25p80: support quad mode for w25q01jvq, Cédric Le Goater, 2024/07/09
- [PULL 8/8] machine_aspeed.py: update to test network for AST2700, Cédric Le Goater, 2024/07/09
- Re: [PULL 0/8] aspeed queue, Richard Henderson, 2024/07/09