hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic
PATCH 1/2 hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic
From: Bin Meng <[email protected]>
For the ECSPIx_CONREG register BURST_LENGTH field, the manual says:
0x020 A SPI burst contains the 1 LSB in first word and all 32 bits in second word. 0x021 A SPI burst contains the 2 LSB in first word and all 32 bits in second word.
Current logic uses either s->burst_length or 32, whichever smaller, to determine how many bits it should read from the tx fifo each time. For example, for a 48 bit burst length, current logic transfers the first 32 bit from the first word in the tx fifo, followed by a 16 bit from the second word in the tx fifo, which is wrong. The correct logic should be: transfer the first 16 bit from the first word in the tx fifo, followed by a 32 bit from the second word in the tx fifo.
With this change, SPI flash can be successfully probed by U-Boot on imx6 sabrelite board.
=> sf probe
SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiB
---
hw/ssi/imx_spi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
hw/ssi/imx_spi.c +4 -1
@@ -178,7 +178,10 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("data tx:0x%08x\n" , tx); - tx_burst =MIN (s->burst_length,32 ); + tx_burst = s->burst_length %32 ; +if (tx_burst ==0 ) { + tx_burst =32 ; + } rx =0 ;
Signature · 1 line
2.7.4
PATCH 2/2 hw/ssi: imx_spi: Correct tx and rx fifo endianness
From: Bin Meng <[email protected]>
The endianness of data exchange between tx and rx fifo is incorrect. Earlier bytes are supposed to show up on MSB and later bytes on LSB,
ie: in big endian. The manual does not explicitly say this, but the U-Boot and Linux driver codes have a swap on the data transferred to tx fifo and from rx fifo.
With this change, U-Boot read from / write to SPI flash tests pass.
=> sf test 1ff000 1000
SPI flash test:
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 Mbps
Test passed
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 Mbps
---
hw/ssi/imx_spi.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)
hw/ssi/imx_spi.c +12 -4
@@ -156,13 +156,14 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) {uint32_t tx;uint32_t rx; +uint32_t data; +uint8_t byte;DPRINTF ("Begin: TX Fifo Size = %d, RX Fifo Size = %d\n" ,fifo32_num_used (&s->tx_fifo),fifo32_num_used (&s->rx_fifo));while (!fifo32_is_empty (&s->tx_fifo)) {int tx_burst =0 ; -int index =0 ;if (s->burst_length <=0 ) { s->burst_length =imx_spi_burst_length (s);
@@ -183,10 +184,18 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) tx_burst =32 ; } + data =0 ; +for (int i =0 ; i < tx_burst /8 ; i++) { + byte = tx &0xff ; + tx = tx >>8 ; + data = (data <<8 ) | byte; + } + tx = data; + rx =0 ;while (tx_burst >0 ) { -uint8_t byte = tx &0xff ; + byte = tx &0xff ;DPRINTF ("writing 0x%02x\n" , (uint32_t )byte);
@@ -196,12 +205,11 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("0x%02x read\n" , (uint32_t )byte); tx = tx >>8 ; - rx |= (byte << (index *8 )); + rx = (rx <<8 ) | byte;/* Remove 8 bits from the actual burst */ tx_burst -=8 ; s->burst_length -=8 ; - index++; }DPRINTF ("data rx:0x%08x\n" , rx);
Signature · 1 line
2.7.4
Re: [PATCH 2/2] hw/ssi: imx_spi: Correct tx and rx fifo endianness
Quoted 8 lines · Bin Meng
From: Bin Meng <[email protected]>
The endianness of data exchange between tx and rx fifo is incorrect. Earlier bytes are supposed to show up on MSB and later bytes on LSB,
ie: in big endian. The manual does not explicitly say this, but the U-Boot and Linux driver codes have a swap on the data transferred to tx fifo and from rx fifo.
To check my understanding, if we have a burst length of 16 bits, say, when we do the fifo32_pop() of a 32 bit word, where in that word and which way round are the 2 bytes we are going to transfer ?
Quoted 44 lines
With this change, U-Boot read from / write to SPI flash tests pass.
=> sf test 1ff000 1000
SPI flash test:
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 Mbps
Test passed
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 MbpsFixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
hw/ssi/imx_spi.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)hw/ssi/imx_spi.c +2
-1@@ -156,13 +156,14 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) {uint32_t tx;uint32_t rx; +uint32_t data; +uint8_t byte;DPRINTF ("Begin: TX Fifo Size = %d, RX Fifo Size = %d\n" ,fifo32_num_used (&s->tx_fifo),fifo32_num_used (&s->rx_fifo));while (!fifo32_is_empty (&s->tx_fifo)) {int tx_burst =0 ; -int index =0 ;if (s->burst_length <=0 ) { s->burst_length =imx_spi_burst_length (s);@@ -183,10 +184,18 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) tx_burst =32 ;
hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 194
Quoted 10 lines
hw/ssi/imx_spi.c +8
-0} + data =0 ; +for (int i =0 ; i < tx_burst /8 ; i++) { + byte = tx &0xff ; + tx = tx >>8 ; + data = (data <<8 ) | byte; + } + tx = data; +
Why carefully reverse the order of bytes in the word and then take a byte at a time from the bottom of the word in the loop below, when you could change the loop to take bytes from the top of the word instead ?
Quoted 23 lines
hw/ssi/imx_spi.c +2
-4rx =0 ;while (tx_burst >0 ) { -uint8_t byte = tx &0xff ; + byte = tx &0xff ;DPRINTF ("writing 0x%02x\n" , (uint32_t )byte);@@ -196,12 +205,11 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("0x%02x read\n" , (uint32_t )byte); tx = tx >>8 ; - rx |= (byte << (index *8 )); + rx = (rx <<8 ) | byte;/* Remove 8 bits from the actual burst */ tx_burst -=8 ; s->burst_length -=8 ; - index++; }DPRINTF ("data rx:0x%08x\n" , rx); --
Quoted 1 line
2.7.4
thanks
-- PMM
Re: [PATCH 2/2] hw/ssi: imx_spi: Correct tx and rx fifo endianness
Hi Peter,
Quoted 11 lines · Peter Maydell
Quoted 8 lines · Bin Meng
From: Bin Meng <[email protected]>
The endianness of data exchange between tx and rx fifo is incorrect. Earlier bytes are supposed to show up on MSB and later bytes on LSB,
ie: in big endian. The manual does not explicitly say this, but the U-Boot and Linux driver codes have a swap on the data transferred to tx fifo and from rx fifo.
Quoted 3 lines
To check my understanding, if we have a burst length of 16 bits, say, when we do the
fifo32_pop()of a 32 bit word, where in that word and which way round are the 2 bytes we are going to transfer ?
Say the fifo was written with a value of 0x00001234 when the burst length is 16 bits, 0x12 will be transferred first then followed by 0x34.
Quoted 47 lines
Quoted 47 lines
With this change, U-Boot read from / write to SPI flash tests pass.
=> sf test 1ff000 1000
SPI flash test:
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 Mbps
Test passed
0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
2 write: 235 ticks, 17 KiB/s 0.136 Mbps
3 read: 2 ticks, 2000 KiB/s 16.000 MbpsFixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
hw/ssi/imx_spi.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)hw/ssi/imx_spi.c +2
-1@@ -156,13 +156,14 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) {uint32_t tx;uint32_t rx; +uint32_t data; +uint8_t byte;DPRINTF ("Begin: TX Fifo Size = %d, RX Fifo Size = %d\n" ,fifo32_num_used (&s->tx_fifo),fifo32_num_used (&s->rx_fifo));while (!fifo32_is_empty (&s->tx_fifo)) {int tx_burst =0 ; -int index =0 ;if (s->burst_length <=0 ) { s->burst_length =imx_spi_burst_length (s);@@ -183,10 +184,18 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) tx_burst =32 ; }
Quoted 13 lines
hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 194
Quoted 8 lines
hw/ssi/imx_spi.c +8
-0+ data =0 ; +for (int i =0 ; i < tx_burst /8 ; i++) { + byte = tx &0xff ; + tx = tx >>8 ; + data = (data <<8 ) | byte; + } + tx = data; +Why carefully reverse the order of bytes in the word and then take a byte at a time from the bottom of the word in the loop below, when you could change the loop to take bytes from the top of the word instead ?
Ah, yes, this can be rewritten to simplify a little.
Quoted 16 lines
Quoted 16 lines
hw/ssi/imx_spi.c +2
-2rx =0 ;while (tx_burst >0 ) { -uint8_t byte = tx &0xff ; + byte = tx &0xff ;DPRINTF ("writing 0x%02x\n" , (uint32_t )byte);@@ -196,12 +205,11 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("0x%02x read\n" , (uint32_t )byte); tx = tx >>8 ; - rx |= (byte << (index *8 )); + rx = (rx <<8 ) | byte;
Quoted 8 lines
hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s)
Quoted 8 lines
hw/ssi/imx_spi.c +0
-2/* Remove 8 bits from the actual burst */ tx_burst -=8 ; s->burst_length -=8 ; - index++; }DPRINTF ("data rx:0x%08x\n" , rx); --
Regards,
Bin
Quoted 22 lines · Bin Meng
From: Bin Meng <[email protected]>
For the
ECSPIx_CONREGregisterBURST_LENGTHfield, the manual says:
0x020A SPI burst contains the 1 LSB in first word and all 32 bits in second word.0x021A SPI burst contains the 2 LSB in first word and all 32 bits in second word.Current logic uses either
s->burst_lengthor 32, whichever smaller, to determine how many bits it should read from the tx fifo each time. For example, for a 48 bit burst length, current logic transfers the first 32 bit from the first word in the tx fifo, followed by a 16 bit from the second word in the tx fifo, which is wrong. The correct logic should be: transfer the first 16 bit from the first word in the tx fifo, followed by a 32 bit from the second word in the tx fifo.With this change, SPI flash can be successfully probed by U-Boot on imx6 sabrelite board.
=> sf probe
SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiB
Quoted 7 lines
Fixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
hw/ssi/imx_spi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
Ping?
Quoted 28 lines · Bin Meng
Quoted 26 lines · Bin Meng
From: Bin Meng <[email protected]>
For the
ECSPIx_CONREGregisterBURST_LENGTHfield, the manual says:
0x020A SPI burst contains the 1 LSB in first word and all 32 bits in second word.0x021A SPI burst contains the 2 LSB in first word and all 32 bits in second word.Current logic uses either
s->burst_lengthor 32, whichever smaller, to determine how many bits it should read from the tx fifo each time. For example, for a 48 bit burst length, current logic transfers the first 32 bit from the first word in the tx fifo, followed by a 16 bit from the second word in the tx fifo, which is wrong. The correct logic should be: transfer the first 16 bit from the first word in the tx fifo, followed by a 32 bit from the second word in the tx fifo.With this change, SPI flash can be successfully probed by U-Boot on imx6 sabrelite board.
=> sf probe
SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiBFixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
Quoted 5 lines
Quoted 3 lines
hw/ssi/imx_spi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)Ping?
Ping?
Quoted 33 lines · Bin Meng
From: Bin Meng <[email protected]>
For the
ECSPIx_CONREGregisterBURST_LENGTHfield, the manual says:
0x020A SPI burst contains the 1 LSB in first word and all 32 bits in second word.0x021A SPI burst contains the 2 LSB in first word and all 32 bits in second word.Current logic uses either
s->burst_lengthor 32, whichever smaller, to determine how many bits it should read from the tx fifo each time. For example, for a 48 bit burst length, current logic transfers the first 32 bit from the first word in the tx fifo, followed by a 16 bit from the second word in the tx fifo, which is wrong. The correct logic should be: transfer the first 16 bit from the first word in the tx fifo, followed by a 32 bit from the second word in the tx fifo.With this change, SPI flash can be successfully probed by U-Boot on imx6 sabrelite board.
=> sf probe
SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiBFixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
hw/ssi/imx_spi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)hw/ssi/imx_spi.c +0
-0
hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 184
Quoted 8 lines
hw/ssi/imx_spi.c +4
-1@@ -178,7 +178,10 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("data tx:0x%08x\n" , tx); - tx_burst =MIN (s->burst_length,32 ); + tx_burst = s->burst_length %32 ; +if (tx_burst ==0 ) { + tx_burst =32 ; + }
Or alternatively using ternary operator:
tx_burst = (s->burst_length % 32) ?: 32;
Quoted 4 lines
hw/ssi/imx_spi.c +0
-0rx =0 ;
Quoted 46 lines · Philippe Mathieu-Daudé
Quoted 33 lines · Bin Meng
From: Bin Meng <[email protected]>
For the
ECSPIx_CONREGregisterBURST_LENGTHfield, the manual says:
0x020A SPI burst contains the 1 LSB in first word and all 32 bits in second word.0x021A SPI burst contains the 2 LSB in first word and all 32 bits in second word.Current logic uses either
s->burst_lengthor 32, whichever smaller, to determine how many bits it should read from the tx fifo each time. For example, for a 48 bit burst length, current logic transfers the first 32 bit from the first word in the tx fifo, followed by a 16 bit from the second word in the tx fifo, which is wrong. The correct logic should be: transfer the first 16 bit from the first word in the tx fifo, followed by a 32 bit from the second word in the tx fifo.With this change, SPI flash can be successfully probed by U-Boot on imx6 sabrelite board.
=> sf probe
SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiBFixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>---
hw/ssi/imx_spi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)hw/ssi/imx_spi.c +0
-0hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 184
Quoted 8 lines
hw/ssi/imx_spi.c +4
-1@@ -178,7 +178,10 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)DPRINTF ("data tx:0x%08x\n" , tx); - tx_burst =MIN (s->burst_length,32 ); + tx_burst = s->burst_length %32 ; +if (tx_burst ==0 ) { + tx_burst =32 ; + }Or alternatively using ternary operator:
Quoted 1 line
tx_burst = (s->burst_length % 32) ?: 32;
Updated this in v2 series:
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=222931
Quoted 3 lines
Reviewed-by Philippe Mathieu-Daudé <[email protected]>
Regards,
Bin