hw/net/lan9118: Fix overflow in TX FIFO
v2 cover hw/net/lan9118: Fix overflow in TX FIFO
Fix for https://gitlab.com/qemu-project/qemu/-/issues/2267
Since v1:
- Renamed definition as MIL_TXFIFO_SIZE
- Addressed Peter review comments in patches 1 & 2 (add comment, return TXE INT)
- Trivial patches while digesting Peter's analysis [*]
More work expected during 9.1.
Philippe Mathieu-Daudé (11):
hw/net/lan9118: Replace magic '2048' value by MIL_TXFIFO_SIZE
definition
hw/net/lan9118: Fix overflow in MIL TX FIFO hw/net/lan9118: Remove duplicated assignment hw/net/lan9118: Replace magic '5' value by TX_FIF_SZ_RESET definition hw/net/lan9118: Add definitions for FIFO allocated sizes hw/net/lan9118: Use TX_DATA_FIFO_BYTES definition hw/net/lan9118: Rename tx_fifo_size -> tx_fifo_bytes hw/net/lan9118: Use RX_STATUS_FIFO_BYTES definition hw/net/lan9118: Rename rx_status_fifo_size -> rx_status_fifo_wordcount hw/net/lan9118: Use RX_DATA_FIFO_BYTES definition hw/net/lan9118: Rename rx_fifo_size -> rx_fifo_wordcount
hw/net/lan9118.c | 104 +++++++++++++++++++++++++++++++---------------- 1 file changed, 70 insertions(+), 34 deletions(-)
Signature · 1 line
2.41.0
PATCH v2 1/11 hw/net/lan9118: Replace magic '2048' value by MIL_TXFIFO_SIZE definition
The magic 2048 is explained in the LAN9211 datasheet (DS00002414A) in chapter 1.4, "10/100 Ethernet MAC":
The MAC Interface Layer (MIL), within the MAC, contains a 2K Byte transmit and a 128 Byte receive FIFO which is separate from the TX and RX FIFOs. [...]
---
(Not including Peter R-b from v1 due to semantic change)
hw/net/lan9118.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
hw/net/lan9118.c +9 -3
@@ -150,6 +150,12 @@ do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)#define GPT_TIMER_EN 0x20000000 +/* +* The MAC Interface Layer (MIL), within the MAC, contains a 2K Byte transmit +* and a 128 Byte receive FIFO which is separate from the TX and RX FIFOs. +*/ +#define MIL_TXFIFO_SIZE 2048 +enum tx_state {TX_IDLE ,TX_B ,
@@ -166,7 +172,7 @@ typedef struct {int32_t pad;int32_t fifo_used;int32_t len; -uint8_t data[2048 ]; +uint8_t data[MIL_TXFIFO_SIZE ]; } LAN9118Packet;static const VMStateDescription vmstate_lan9118_packet = {
@@ -182,7 +188,7 @@ static const VMStateDescription vmstate_lan9118_packet = {VMSTATE_INT32 (pad, LAN9118Packet),VMSTATE_INT32 (fifo_used, LAN9118Packet),VMSTATE_INT32 (len, LAN9118Packet), -VMSTATE_UINT8_ARRAY (data, LAN9118Packet,2048 ), +VMSTATE_UINT8_ARRAY (data, LAN9118Packet,MIL_TXFIFO_SIZE ), VMSTATE_END_OF_LIST() } };
@@ -544,7 +550,7 @@ static ssize_t lan9118_receive(NetClientState *nc, const uint8_t *buf,return -1 ; } -if (size >=2048 || size <14 ) { +if (size >=MIL_TXFIFO_SIZE || size <14 ) {return -1 ; }
Signature · 1 line
2.41.0
Re: [PATCH-for-9.0 v2 01/11] hw/net/lan9118: Replace magic '2048' value by MIL_TXFIFO_SIZE definition
Quoted 11 lines · Philippe Mathieu-Daudé
The magic 2048 is explained in the LAN9211 datasheet (DS00002414A) in chapter 1.4, "10/100 Ethernet MAC":
The MAC Interface Layer (MIL), within the MAC, contains a 2K Byte transmit and a 128 Byte receive FIFO which is separate from the TX and RX FIFOs. [...]
Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
(Not including Peter R-b from v1 due to semantic change)
Though the use of the constant in lan9118_receive() reveals that our implementation is using the same buffer for both tx and rx...
thanks
-- PMM
PATCH v2 2/11 hw/net/lan9118: Fix overflow in MIL TX FIFO
In reply to Philippe Mathieu-DaudéWhen the MAC Interface Layer (MIL) transmit FIFO is full, truncate the packet, and raise the Transmitter Error (TXE) flag.
Broken since model introduction in commit 2a42499017 ("LAN9118 emulation").
When using the reproducer from
https://gitlab.com/qemu-project/qemu/-/issues/2267 we get:
hw/net/lan9118.c:798:17: runtime error:
index 2048 out of bounds for type 'uint8_t[2048]' (aka 'unsigned char[2048]')
#0 0x563ec9a057b1 in tx_fifo_push hw/net/lan9118.c:798:43
#1 0x563ec99fbb28 in lan9118_writel hw/net/lan9118.c:1042:9
#2 0x563ec99f2de2 in lan9118_16bit_mode_write hw/net/lan9118.c:1205:9 #3 0x563ecbf78013 in memory_region_write_accessor system/memory.c:497:5 #4 0x563ecbf776f5 in access_with_adjusted_size system/memory.c:573:18 #5 0x563ecbf75643 in memory_region_dispatch_write system/memory.c:1521:16 #6 0x563ecc01bade in flatview_write_continue_step system/physmem.c:2713:18 #7 0x563ecc01b374 in flatview_write_continue system/physmem.c:2743:19 #8 0x563ecbff1c9b in flatview_write system/physmem.c:2774:12 #9 0x563ecbff1768 in address_space_write system/physmem.c:2894:18
...
[*] LAN9118 DS00002266B.pdf, Table 5.3.3 "INTERRUPT STATUS REGISTER"
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2267
--- hw/net/lan9118.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-)
hw/net/lan9118.c +15 -1
@@ -799,8 +799,22 @@ static void tx_fifo_push(lan9118_state *s, uint32_t val)/* Documentation is somewhat unclear on the ordering of bytes in FIFO words. Empirical results show it to be little-endian. */ -/* TODO: FIFO overflow checking. */ while (n--) { +if (s->txp->len ==MIL_TXFIFO_SIZE ) { +/* +* No more space in the FIFO. The datasheet is not +* precise about this case. We choose what is easiest +* to model: the packet is truncated, and TXE is raised. +* +* Note, it could be a fragmented packet, but we currently +* do not handle that (see earlier TX_B case). +*/ +qemu_log_mask (LOG_GUEST_ERROR , +"MIL TX FIFO overrun, discarding %u byte%s\n" , + n, n >1 ?"s" :"" ); + s->int_sts |=TXE_INT ; +break ; + } s->txp->data[s->txp->len] = val &0xff ; s->txp->len++; val >>=8 ;
Signature · 1 line
2.41.0
Re: [PATCH-for-9.0 v2 02/11] hw/net/lan9118: Fix overflow in MIL TX FIFO
Quoted 25 lines · Philippe Mathieu-Daudé
When the MAC Interface Layer (MIL) transmit FIFO is full, truncate the packet, and raise the Transmitter Error (TXE) flag.
Broken since model introduction in commit
2a42499017("LAN9118 emulation").When using the reproducer from
https://gitlab.com/qemu-project/qemu/-/issues/2267 we get:
hw/net/lan9118.c:798:17: runtime error:
index 2048 out of bounds for type 'uint8_t[2048]' (aka 'unsigned char[2048]')#0 0x563ec9a057b1 in tx_fifo_push hw/net/lan9118.c:798:43 #1 0x563ec99fbb28 in lan9118_writel hw/net/lan9118.c:1042:9 #2 0x563ec99f2de2 in lan9118_16bit_mode_write hw/net/lan9118.c:1205:9 #3 0x563ecbf78013 in memory_region_write_accessor system/memory.c:497:5 #4 0x563ecbf776f5 in access_with_adjusted_size system/memory.c:573:18 #5 0x563ecbf75643 in memory_region_dispatch_write system/memory.c:1521:16 #6 0x563ecc01bade in flatview_write_continue_step system/physmem.c:2713:18 #7 0x563ecc01b374 in flatview_write_continue system/physmem.c:2743:19 #8 0x563ecbff1c9b in flatview_write system/physmem.c:2774:12 #9 0x563ecbff1768 in address_space_write system/physmem.c:2894:18 ...
Quoted 8 lines
[*] LAN9118 DS00002266B.pdf, Table 5.3.3 "INTERRUPT STATUS REGISTER"
Reported-by Will LesterReported-by Chuhong Yuan <[email protected]>Suggested-by Peter Maydell <[email protected]>Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2267
Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
thanks
-- PMM
PATCH v2 3/11 hw/net/lan9118: Remove duplicated assignment
In reply to Philippe Mathieu-Daudés->txp->fifo_used is zeroed in the next 3 lines.
--- hw/net/lan9118.c | 1 - 1 file changed, 1 deletion(-)
hw/net/lan9118.c +0 -1
@@ -438,7 +438,6 @@ static void lan9118_reset(DeviceState *d) s->hw_cfg = s->mode_16bit ?0x00050000 :0x00050004 ; s->pmt_ctrl &=0x45 ; s->gpio_cfg =0 ; - s->txp->fifo_used =0 ; s->txp->state =TX_IDLE ; s->txp->cmd_a =0xffffffffu ; s->txp->cmd_b =0xffffffffu ;
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 03/11] hw/net/lan9118: Remove duplicated assignment
Quoted 13 lines · Philippe Mathieu-Daudé
s->txp->fifo_usedis zeroed in the next 3 lines.Signed-off-by Philippe Mathieu-Daudé <[email protected]>--- hw/net/lan9118.c | 1 - 1 file changed, 1 deletion(-)hw/net/lan9118.c +0
-0
hw/net/lan9118.c static void lan9118_reset(DeviceState *d) · line 443
Quoted 7 lines
hw/net/lan9118.c +0
-1@@ -438,7 +438,6 @@ static void lan9118_reset(DeviceState *d) s->hw_cfg = s->mode_16bit ?0x00050000 :0x00050004 ; s->pmt_ctrl &=0x45 ; s->gpio_cfg =0 ; - s->txp->fifo_used =0 ; s->txp->state =TX_IDLE ; s->txp->cmd_a =0xffffffffu ; s->txp->cmd_b =0xffffffffu ;
thanks
-- PMM
PATCH v2 4/11 hw/net/lan9118: Replace magic '5' value by TX_FIF_SZ_RESET definition
In reply to Philippe Mathieu-DaudéTX_FIF_SZ is described in chapter 5.3.9,
"HW_CFG — HARDWARE CONFIGURATION REGISTER".
--- hw/net/lan9118.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
hw/net/lan9118.c +3 -1
@@ -71,6 +71,8 @@ do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)#define CSR_E2P_CMD 0xb0#define CSR_E2P_DATA 0xb4 +#define TX_FIF_SZ_RESET 5 +#define E2P_CMD_MAC_ADDR_LOADED 0x100/* IRQ_CFG */
@@ -435,7 +437,7 @@ static void lan9118_reset(DeviceState *d) s->fifo_int =0x48000000 ; s->rx_cfg =0 ; s->tx_cfg =0 ; - s->hw_cfg = s->mode_16bit ?0x00050000 :0x00050004 ; + s->hw_cfg = (TX_FIF_SZ_RESET <<16 ) | (s->mode_16bit <<2 ); s->pmt_ctrl &=0x45 ; s->gpio_cfg =0 ; s->txp->state =TX_IDLE ;
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 04/11] hw/net/lan9118: Replace magic '5' value by TX_FIF_SZ_RESET definition
Quoted 6 lines · Philippe Mathieu-Daudé
TX_FIF_SZis described in chapter 5.3.9,
"HW_CFG— HARDWARE CONFIGURATION REGISTER".Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
thanks
-- PMM
PATCH v2 5/11 hw/net/lan9118: Add definitions for FIFO allocated sizes
In reply to Philippe Mathieu-DaudéAdd definitions for the TX_FIF_SZ=5 case, per TABLE 5-3 "VALID TX/RX FIFO ALLOCATIONS".
--- hw/net/lan9118.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
hw/net/lan9118.c +11 -0
@@ -158,6 +158,17 @@ do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)*/ #define MIL_TXFIFO_SIZE 2048 +/* +* TX and RX FIFO space is configurable through the TX FIFO Size (TX_FIF_SZ) +* field in the hardware configuration (CSR HW_CFG) register. These are the +* default configuration settings for TX_FIF_SZ = 5 +* (see TABLE 5-3: VALID TX/RX FIFO ALLOCATIONS). +*/ +#define TX_DATA_FIFO_BYTES 4608/* 1152 words */ +#define TX_STATUS_FIFO_BYTES 512/* 128 words */ +#define RX_DATA_FIFO_BYTES 10560/* 2640 words */ +#define RX_STATUS_FIFO_BYTES 704/* 176 words */ +enum tx_state {TX_IDLE ,TX_B ,
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 05/11] hw/net/lan9118: Add definitions for FIFO allocated sizes
Quoted 17 lines · Philippe Mathieu-Daudé
Add definitions for the TX_FIF_SZ=5 case, per TABLE 5-3 "VALID TX/RX FIFO ALLOCATIONS".
Signed-off-by Philippe Mathieu-Daudé <[email protected]>--- hw/net/lan9118.c | 11 +++++++++++ 1 file changed, 11 insertions(+)hw/net/lan9118.c +0
-0@@ -158,6 +158,17 @@ do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0)*/ #define MIL_TXFIFO_SIZE 2048
hw/net/lan9118.c do { printf("lan9118: " fmt , ## __VA_ARGS__); } while (0) · line 170
Quoted 10 lines
hw/net/lan9118.c +10
-0+/* +* TX and RX FIFO space is configurable through the TX FIFO Size (TX_FIF_SZ) +* field in the hardware configuration (CSR HW_CFG) register. These are the +* default configuration settings for TX_FIF_SZ = 5 +* (see TABLE 5-3: VALID TX/RX FIFO ALLOCATIONS). +*/ +#define TX_DATA_FIFO_BYTES 4608/* 1152 words */ +#define TX_STATUS_FIFO_BYTES 512/* 128 words */ +#define RX_DATA_FIFO_BYTES 10560/* 2640 words */ +#define RX_STATUS_FIFO_BYTES 704/* 176 words */
We could make these do the actual calculations, rather than hardcoding the results:
#define TX_STATUS_FIFO_BYTES 512
#define TX_TOTAL_FIFO_BYTES(TX_FIF_SZ) (1024 * (TX_FIF_SZ)) #define RX_TOTAL_FIFO_BYTES(TX_FIF_SZ) (16384 - TX_TOTAL_FIFO_BYTES(TX_FIF_SZ)) #define TX_DATA_FIFO_BYTES(TX_FIF_SZ) (TX_TOTAL_FIFO_BYTES(TX_FIF_SZ)
- TX_STATUS_FIFO_BYTES)
#define RX_STATUS_FIFO_BYTES (RX_TOTAL_FIFO_BYTES(TX_FIF_SZ) >> 4)
#define RX_DATA_FIFO_BYTES(TX_FIF_SZ) \
(RX_TOTAL_FIFO_BYTES(TX_FIF_SZ) - RX_STATUS_FIFO_BYTES(TX_FIF_SZ))thanks
-- PMM
PATCH v2 6/11 hw/net/lan9118: Use TX_DATA_FIFO_BYTES definition
In reply to Philippe Mathieu-Daudé--- hw/net/lan9118.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/net/lan9118.c +1 -1
@@ -456,7 +456,7 @@ static void lan9118_reset(DeviceState *d) s->txp->cmd_b =0xffffffffu ; s->txp->len =0 ; s->txp->fifo_used =0 ; - s->tx_fifo_size =4608 ; + s->tx_fifo_size =TX_DATA_FIFO_BYTES ; s->tx_status_fifo_used =0 ; s->rx_status_fifo_size =704 ; s->rx_fifo_size =2640 ;
Signature · 1 line
2.41.0
PATCH v2 7/11 hw/net/lan9118: Rename tx_fifo_size -> tx_fifo_bytes
In reply to Philippe Mathieu-Daudétx_fifo_size is a byte count, rename it to avoid confusion.
--- hw/net/lan9118.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
hw/net/lan9118.c +5 -5
@@ -251,7 +251,7 @@ struct lan9118_state {int32_t eeprom_writable;uint8_t eeprom[128 ]; -int32_t tx_fifo_size; +int32_t tx_fifo_bytes;LAN9118Packet *txp;LAN9118Packet tx_packet;
@@ -322,7 +322,7 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_UINT32 (phy_int_mask, lan9118_state),VMSTATE_INT32 (eeprom_writable, lan9118_state),VMSTATE_UINT8_ARRAY (eeprom, lan9118_state,128 ), -VMSTATE_INT32 (tx_fifo_size, lan9118_state), +VMSTATE_INT32 (tx_fifo_bytes, lan9118_state),/* txp always points at tx_packet so need not be saved */ VMSTATE_STRUCT (tx_packet, lan9118_state,0 , vmstate_lan9118_packet, LAN9118Packet),
@@ -456,7 +456,7 @@ static void lan9118_reset(DeviceState *d) s->txp->cmd_b =0xffffffffu ; s->txp->len =0 ; s->txp->fifo_used =0 ; - s->tx_fifo_size =TX_DATA_FIFO_BYTES ; + s->tx_fifo_bytes =TX_DATA_FIFO_BYTES ; s->tx_status_fifo_used =0 ; s->rx_status_fifo_size =704 ; s->rx_fifo_size =2640 ;
@@ -757,7 +757,7 @@ static void tx_fifo_push(lan9118_state *s, uint32_t val) {int n; -if (s->txp->fifo_used == s->tx_fifo_size) { +if (s->txp->fifo_used == s->tx_fifo_bytes) { s->int_sts |=TDFO_INT ;return ; }
@@ -1285,7 +1285,7 @@ static uint64_t lan9118_readl(void *opaque, hwaddr offset,return (s->rx_status_fifo_used <<16 ) | (s->rx_fifo_used <<2 );case CSR_TX_FIFO_INF :return (s->tx_status_fifo_used <<16 ) - | (s->tx_fifo_size - s->txp->fifo_used); + | (s->tx_fifo_bytes - s->txp->fifo_used);case CSR_PMT_CTRL :return s->pmt_ctrl;case CSR_GPIO_CFG :
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 07/11] hw/net/lan9118: Rename tx_fifo_size -> tx_fifo_bytes
Quoted 2 lines · Philippe Mathieu-Daudé
tx_fifo_sizeis a byte count, rename it to avoid confusion.
But we don't consistently use it as a byte count.
In tx_fifo_push():
if (s->txp->fifo_used == s->tx_fifo_size) {
s->int_sts |= TDFO_INT;
return;
}where fifo_used is incremented for every word handled by the tx_fifo, not for every byte.
thanks
-- PMM
PATCH v2 8/11 hw/net/lan9118: Use RX_STATUS_FIFO_BYTES definition
In reply to Philippe Mathieu-Daudérx_status_fifo[] is an array of words,rx_status_fifo_size is a word count.
--- hw/net/lan9118.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)
hw/net/lan9118.c +5 -4
@@ -262,7 +262,7 @@ struct lan9118_state {int32_t rx_status_fifo_size;int32_t rx_status_fifo_used;int32_t rx_status_fifo_head; -uint32_t rx_status_fifo[896 ]; +uint32_t rx_status_fifo[RX_STATUS_FIFO_BYTES /4 ];int32_t rx_fifo_size;int32_t rx_fifo_used;int32_t rx_fifo_head;
@@ -332,7 +332,9 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_INT32 (rx_status_fifo_size, lan9118_state),VMSTATE_INT32 (rx_status_fifo_used, lan9118_state),VMSTATE_INT32 (rx_status_fifo_head, lan9118_state), -VMSTATE_UINT32_ARRAY (rx_status_fifo, lan9118_state,896 ), +VMSTATE_UINT32_ARRAY (rx_status_fifo, lan9118_state, +RX_STATUS_FIFO_BYTES /4 ), +VMSTATE_UNUSED (896 *4 -RX_STATUS_FIFO_BYTES ),VMSTATE_INT32 (rx_fifo_size, lan9118_state),VMSTATE_INT32 (rx_fifo_used, lan9118_state), VMSTATE_INT32(rx_fifo_head, lan9118_state),
@@ -458,10 +460,9 @@ static void lan9118_reset(DeviceState *d) s->txp->fifo_used =0 ; s->tx_fifo_bytes =TX_DATA_FIFO_BYTES ; s->tx_status_fifo_used =0 ; - s->rx_status_fifo_size =704 ; s->rx_fifo_size =2640 ; s->rx_fifo_used =0 ; - s->rx_status_fifo_size =176 ; + s->rx_status_fifo_size =RX_STATUS_FIFO_BYTES /4 ; s->rx_status_fifo_used =0 ; s->rxp_offset =0 ; s->rxp_size =0 ;
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 08/11] hw/net/lan9118: Use RX_STATUS_FIFO_BYTES definition
Quoted 23 lines · Philippe Mathieu-Daudé
rx_status_fifo[] is an array of words,
rx_status_fifo_sizeis a word count.Signed-off-by Philippe Mathieu-Daudé <[email protected]>--- hw/net/lan9118.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)hw/net/lan9118.c +1
-1@@ -262,7 +262,7 @@ struct lan9118_state {int32_t rx_status_fifo_size;int32_t rx_status_fifo_used;int32_t rx_status_fifo_head; -uint32_t rx_status_fifo[896 ]; +uint32_t rx_status_fifo[RX_STATUS_FIFO_BYTES /4 ];int32_t rx_fifo_size;int32_t rx_fifo_used;int32_t rx_fifo_head;
hw/net/lan9118.c static const VMStateDescription vmstate_lan9118 = { · line 340
Quoted 10 lines
hw/net/lan9118.c +3
-1@@ -332,7 +332,9 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_INT32 (rx_status_fifo_size, lan9118_state),VMSTATE_INT32 (rx_status_fifo_used, lan9118_state),VMSTATE_INT32 (rx_status_fifo_head, lan9118_state), -VMSTATE_UINT32_ARRAY (rx_status_fifo, lan9118_state,896 ), +VMSTATE_UINT32_ARRAY (rx_status_fifo, lan9118_state, +RX_STATUS_FIFO_BYTES /4 ), +VMSTATE_UNUSED (896 *4 -RX_STATUS_FIFO_BYTES ),VMSTATE_INT32 (rx_fifo_size, lan9118_state),VMSTATE_INT32 (rx_fifo_used, lan9118_state), VMSTATE_INT32(rx_fifo_head, lan9118_state),
Ideally in the state struct we should have the arrays be the size of the largest possible FIFO, not the size of the default-out-of-reset FIFO, to leave the way open for making the FIFO size be runtime configurable in future if we want that.
thanks
-- PMM
PATCH v2 9/11 hw/net/lan9118: Rename rx_status_fifo_size -> rx_status_fifo_wordcount
In reply to Philippe Mathieu-Daudérx_status_fifo_size is a word count, rename it to avoid confusion.
--- hw/net/lan9118.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
hw/net/lan9118.c +8 -8
@@ -259,7 +259,7 @@ struct lan9118_state {int32_t tx_status_fifo_head;uint32_t tx_status_fifo[512 ]; -int32_t rx_status_fifo_size; +int32_t rx_status_fifo_wordcount;int32_t rx_status_fifo_used;int32_t rx_status_fifo_head;uint32_t rx_status_fifo[RX_STATUS_FIFO_BYTES /4 ];
@@ -329,7 +329,7 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_INT32 (tx_status_fifo_used, lan9118_state),VMSTATE_INT32 (tx_status_fifo_head, lan9118_state),VMSTATE_UINT32_ARRAY (tx_status_fifo, lan9118_state,512 ), -VMSTATE_INT32 (rx_status_fifo_size, lan9118_state), +VMSTATE_INT32 (rx_status_fifo_wordcount, lan9118_state),VMSTATE_INT32 (rx_status_fifo_used, lan9118_state),VMSTATE_INT32 (rx_status_fifo_head, lan9118_state), VMSTATE_UINT32_ARRAY(rx_status_fifo, lan9118_state,
@@ -462,7 +462,7 @@ static void lan9118_reset(DeviceState *d) s->tx_status_fifo_used =0 ; s->rx_fifo_size =2640 ; s->rx_fifo_used =0 ; - s->rx_status_fifo_size =RX_STATUS_FIFO_BYTES /4 ; + s->rx_status_fifo_wordcount =RX_STATUS_FIFO_BYTES /4 ; s->rx_status_fifo_used =0 ; s->rxp_offset =0 ; s->rxp_size =0 ;
@@ -568,7 +568,7 @@ static ssize_t lan9118_receive(NetClientState *nc, const uint8_t *buf, }/* TODO: Implement FIFO overflow notification. */ -if (s->rx_status_fifo_used == s->rx_status_fifo_size) { +if (s->rx_status_fifo_used == s->rx_status_fifo_wordcount) {return -1 ; }
@@ -609,8 +609,8 @@ static ssize_t lan9118_receive(NetClientState *nc, const uint8_t *buf,rx_fifo_push (s, crc); } n = s->rx_status_fifo_head + s->rx_status_fifo_used; -if (n >= s->rx_status_fifo_size) { - n -= s->rx_status_fifo_size; +if (n >= s->rx_status_fifo_wordcount) { + n -= s->rx_status_fifo_wordcount; } s->rx_packet_size[s->rx_packet_size_tail] = fifo_len; s->rx_packet_size_tail = (s->rx_packet_size_tail +1023 ) &1023 ;
@@ -732,8 +732,8 @@ static uint32_t rx_status_fifo_pop(lan9118_state *s)if (s->rx_status_fifo_used !=0 ) { s->rx_status_fifo_used--; s->rx_status_fifo_head++; -if (s->rx_status_fifo_head >= s->rx_status_fifo_size) { - s->rx_status_fifo_head -= s->rx_status_fifo_size; +if (s->rx_status_fifo_head >= s->rx_status_fifo_wordcount) { + s->rx_status_fifo_head -= s->rx_status_fifo_wordcount; }/* ??? What value should be returned when the FIFO is empty? */ DPRINTF ("RX status pop 0x%08x\n" , val);
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 09/11] hw/net/lan9118: Rename rx_status_fifo_size -> rx_status_fifo_wordcount
Quoted 14 lines · Philippe Mathieu-Daudé
rx_status_fifo_sizeis a word count, rename it to avoid confusion.Suggested-by Peter Maydell <[email protected]>Signed-off-by Philippe Mathieu-Daudé <[email protected]>--- hw/net/lan9118.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)hw/net/lan9118.c +0
-0
hw/net/lan9118.c struct lan9118_state { · line 264
Quoted 7 lines
hw/net/lan9118.c +1
-1@@ -259,7 +259,7 @@ struct lan9118_state {int32_t tx_status_fifo_head;uint32_t tx_status_fifo[512 ]; -int32_t rx_status_fifo_size; +int32_t rx_status_fifo_wordcount;int32_t rx_status_fifo_used;int32_t rx_status_fifo_head;
True, but rx_status_fifo_used and rx_status_fifo_head are also word counts. Should we try to indicate units in all these names, or is that getting unwieldy?
-- PMM
PATCH v2 10/11 hw/net/lan9118: Use RX_DATA_FIFO_BYTES definition
In reply to Philippe Mathieu-Daudérx_fifo[] is an array of words,rx_fifo_size is a word count.
--- hw/net/lan9118.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
hw/net/lan9118.c +5 -3
@@ -266,7 +266,7 @@ struct lan9118_state {int32_t rx_fifo_size;int32_t rx_fifo_used;int32_t rx_fifo_head; -uint32_t rx_fifo[3360 ]; +uint32_t rx_fifo[RX_DATA_FIFO_BYTES /4 ];int32_t rx_packet_size_head;int32_t rx_packet_size_tail;int32_t rx_packet_size[1024 ];
@@ -338,7 +338,9 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_INT32 (rx_fifo_size, lan9118_state),VMSTATE_INT32 (rx_fifo_used, lan9118_state),VMSTATE_INT32 (rx_fifo_head, lan9118_state), -VMSTATE_UINT32_ARRAY (rx_fifo, lan9118_state,3360 ), +VMSTATE_UINT32_ARRAY (rx_fifo, lan9118_state, +RX_DATA_FIFO_BYTES /4 ), +VMSTATE_UNUSED (3360 *4 -RX_DATA_FIFO_BYTES ),VMSTATE_INT32 (rx_packet_size_head, lan9118_state),VMSTATE_INT32 (rx_packet_size_tail, lan9118_state), VMSTATE_INT32_ARRAY(rx_packet_size, lan9118_state, 1024),
@@ -460,7 +462,7 @@ static void lan9118_reset(DeviceState *d) s->txp->fifo_used =0 ; s->tx_fifo_bytes =TX_DATA_FIFO_BYTES ; s->tx_status_fifo_used =0 ; - s->rx_fifo_size =2640 ; + s->rx_fifo_size =RX_DATA_FIFO_BYTES /4 ; s->rx_fifo_used =0 ; s->rx_status_fifo_wordcount =RX_STATUS_FIFO_BYTES /4 ; s->rx_status_fifo_used =0 ;
Signature · 1 line
2.41.0
Re: [PATCH-for-9.1 v2 10/11] hw/net/lan9118: Use RX_DATA_FIFO_BYTES definition
Quoted 3 lines · Philippe Mathieu-Daudé
rx_fifo[] is an array of words,
rx_fifo_sizeis a word count.
True, but that's not why rx_fifo[] has been sized to 3360. It's 3360 because that is the worst-case RX data FIFO size in words (if TX_FIF_SZ is 2 then the RX data FIFO is 13440 bytes, which is 3360 words). So in this case the array size is correct.
thanks
-- PMM
PATCH v2 11/11 hw/net/lan9118: Rename rx_fifo_size -> rx_fifo_wordcount
In reply to Philippe Mathieu-Daudérx_fifo_size is a word count, rename it to avoid confusion.
--- hw/net/lan9118.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-)
hw/net/lan9118.c +11 -10
@@ -263,7 +263,7 @@ struct lan9118_state {int32_t rx_status_fifo_used;int32_t rx_status_fifo_head;uint32_t rx_status_fifo[RX_STATUS_FIFO_BYTES /4 ]; -int32_t rx_fifo_size; +int32_t rx_fifo_wordcount;int32_t rx_fifo_used;int32_t rx_fifo_head;uint32_t rx_fifo[RX_DATA_FIFO_BYTES /4 ];
@@ -335,7 +335,7 @@ static const VMStateDescription vmstate_lan9118 = {VMSTATE_UINT32_ARRAY (rx_status_fifo, lan9118_state,RX_STATUS_FIFO_BYTES /4 ),VMSTATE_UNUSED (896 *4 -RX_STATUS_FIFO_BYTES ), -VMSTATE_INT32 (rx_fifo_size, lan9118_state), +VMSTATE_INT32 (rx_fifo_wordcount, lan9118_state),VMSTATE_INT32 (rx_fifo_used, lan9118_state),VMSTATE_INT32 (rx_fifo_head, lan9118_state), VMSTATE_UINT32_ARRAY(rx_fifo, lan9118_state,
@@ -462,7 +462,7 @@ static void lan9118_reset(DeviceState *d) s->txp->fifo_used =0 ; s->tx_fifo_bytes =TX_DATA_FIFO_BYTES ; s->tx_status_fifo_used =0 ; - s->rx_fifo_size =RX_DATA_FIFO_BYTES /4 ; + s->rx_fifo_wordcount =RX_DATA_FIFO_BYTES /4 ; s->rx_fifo_used =0 ; s->rx_status_fifo_wordcount =RX_STATUS_FIFO_BYTES /4 ; s->rx_status_fifo_used =0 ;
@@ -504,8 +504,9 @@ static void rx_fifo_push(lan9118_state *s, uint32_t val) {int fifo_pos; fifo_pos = s->rx_fifo_head + s->rx_fifo_used; -if (fifo_pos >= s->rx_fifo_size) - fifo_pos -= s->rx_fifo_size; +if (fifo_pos >= s->rx_fifo_wordcount) { + fifo_pos -= s->rx_fifo_wordcount; + } s->rx_fifo[fifo_pos] = val; s->rx_fifo_used++; }
@@ -584,7 +585,7 @@ static ssize_t lan9118_receive(NetClientState *nc, const uint8_t *buf, fifo_len = (size + n +3 ) >>2 ;/* Add a word for the CRC. */ fifo_len++; -if (s->rx_fifo_size - s->rx_fifo_used < fifo_len) { +if (s->rx_fifo_wordcount - s->rx_fifo_used < fifo_len) {return -1 ; }
@@ -672,8 +673,8 @@ static uint32_t rx_fifo_pop(lan9118_state *s) }else if (s->rxp_size >0 ) { s->rxp_size--; val = s->rx_fifo[s->rx_fifo_head++]; -if (s->rx_fifo_head >= s->rx_fifo_size) { - s->rx_fifo_head -= s->rx_fifo_size; +if (s->rx_fifo_head >= s->rx_fifo_wordcount) { + s->rx_fifo_head -= s->rx_fifo_wordcount; } s->rx_fifo_used--; }else if (s->rxp_pad >0 ) {
@@ -1135,8 +1136,8 @@ static void lan9118_writel(void *opaque, hwaddr offset, s->rxp_offset =0 ; } s->rx_fifo_head += s->rxp_size; -if (s->rx_fifo_head >= s->rx_fifo_size) { - s->rx_fifo_head -= s->rx_fifo_size; +if (s->rx_fifo_head >= s->rx_fifo_wordcount) { + s->rx_fifo_head -= s->rx_fifo_wordcount; } }break ;
Signature · 1 line
2.41.0
Quoted 10 lines · Philippe Mathieu-Daudé
Fix for https://gitlab.com/qemu-project/qemu/-/issues/2267
Since v1:
- Renamed definition asMIL_TXFIFO_SIZE
- Addressed Peter review comments in patches 1 & 2 (add comment, return TXE INT)
- Trivial patches while digesting Peter's analysis [*]More work expected during 9.1.
Quoted 7 lines
Philippe Mathieu-Daudé (11):
hw/net/lan9118: Replace magic '2048' value byMIL_TXFIFO_SIZE
definition
hw/net/lan9118: Fix overflow in MIL TX FIFO hw/net/lan9118: Remove duplicated assignment
Patches 1 & 2 queued.