spi: Fix FSI-attached controller and AT25 drivers
Subsystems (5)
Eddie James
maintainerv2 cover spi: Fix FSI-attached controller and AT25 drivers
This series implements a number of fixes for the FSI-attached SPI controller driver and the AT25 eeprom driver.
Changes since v1:
- Add patch 6 for checking the mux status before transfers
- Minor fixes to at25 driver formatting and such.
Brad Bishop (4):
spi: fsi: Handle 9 to 15 byte transfers lengths
spi: fsi: Fix clock running too fast
spi: fsi: Fix use of the bneq+ sequencer instruction
eeprom: at25: Split reads into chunks and cap write size
Eddie James (3):
dt-bindings: fsi: fsi2spi: Document new restricted property
spi: fsi: Implement restricted size for certain controllers
spi: fsi: Check mux status before transfers
.../devicetree/bindings/fsi/ibm,fsi2spi.yaml | 10 ++ drivers/misc/eeprom/at25.c | 94 +++++++----- drivers/spi/spi-fsi.c | 139 ++++++++++++++---- 3 files changed, 172 insertions(+), 71 deletions(-)
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 1/7 spi: fsi: Handle 9 to 15 byte transfers lengths
From: Brad Bishop <[email protected]>
The trailing <len> - 8 bytes of transfer data in this size range is no longer ignored.
--- drivers/spi/spi-fsi.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/spi/spi-fsi.c +3 -3
@@ -258,15 +258,15 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx,if (loops >1 ) {fsi_spi_sequence_add (seq,SPI_FSI_SEQUENCE_BRANCH (idx)); -if (rem) -fsi_spi_sequence_add (seq, rem); - rc =fsi_spi_write_reg (ctx,SPI_FSI_COUNTER_CFG ,SPI_FSI_COUNTER_CFG_LOOPS (loops -1 ));if (rc)return rc; } +if (rem) +fsi_spi_sequence_add (seq, rem); +return 0 ; }
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 2/7 spi: fsi: Fix clock running too fast
In reply to Eddie JamesFrom: Brad Bishop <[email protected]>
Use a clock divider tuned to a 200MHz FSI clock. Use of the previous divider at 200MHz results in corrupt data from endpoint devices. Ideally the clock divider would be calculated from the FSI clock, but that would require some significant work on the FSI driver.
--- drivers/spi/spi-fsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/spi/spi-fsi.c +1 -1
@@ -350,7 +350,7 @@ static int fsi_spi_transfer_init(struct fsi_spi *ctx)u64 status =0ULL ;u64 wanted_clock_cfg =SPI_FSI_CLOCK_CFG_ECC_DISABLE |SPI_FSI_CLOCK_CFG_SCK_NO_DEL | -FIELD_PREP (SPI_FSI_CLOCK_CFG_SCK_DIV ,4 ); +FIELD_PREP (SPI_FSI_CLOCK_CFG_SCK_DIV ,19 ); end = jiffies +msecs_to_jiffies (SPI_FSI_INIT_TIMEOUT_MS ); do {
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 3/7 spi: fsi: Fix use of the bneq+ sequencer instruction
In reply to Eddie JamesFrom: Brad Bishop <[email protected]>
All of the switches in N2_count_control in the counter configuration are required to make the branch if not equal and increment command work. Set them when using bneq+.
A side effect of this mode requires a dummy write to TDR when both transmitting and receiving otherwise the controller won't start shifting receive data.
It is likely not possible to avoid TDR underrun errors in this mode and they are harmless, so do not check for them.
--- drivers/spi/spi-fsi.c | 28 +++++++++++++++++++++++++--- 1 file changed, 25 insertions(+), 3 deletions(-)
drivers/spi/spi-fsi.c +25 -3
@@ -29,6 +29,10 @@#define SPI_FSI_ERROR 0x0#define SPI_FSI_COUNTER_CFG 0x1#define SPI_FSI_COUNTER_CFG_LOOPS (x) (((u64)(x) & 0xffULL) << 32) +#define SPI_FSI_COUNTER_CFG_N2_RX BIT_ULL(8) +#define SPI_FSI_COUNTER_CFG_N2_TX BIT_ULL(9) +#define SPI_FSI_COUNTER_CFG_N2_IMPLICIT BIT_ULL(10) +#define SPI_FSI_COUNTER_CFG_N2_RELOAD BIT_ULL(11)#define SPI_FSI_CFG1 0x2#define SPI_FSI_CLOCK_CFG 0x3#define SPI_FSI_CLOCK_CFG_MM_ENABLE BIT_ULL(32)
@@ -61,7 +65,7 @@#define SPI_FSI_STATUS_RDR_OVERRUN BIT_ULL(62)#define SPI_FSI_STATUS_RDR_FULL BIT_ULL(63)#define SPI_FSI_STATUS_ANY_ERROR \ - (SPI_FSI_STATUS_ERROR | SPI_FSI_STATUS_TDR_UNDERRUN | \ + (SPI_FSI_STATUS_ERROR | \ SPI_FSI_STATUS_TDR_OVERRUN | SPI_FSI_STATUS_RDR_UNDERRUN | \ SPI_FSI_STATUS_RDR_OVERRUN)#define SPI_FSI_PORT_CTRL 0x9
@@ -238,6 +242,7 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx,int rc;u8 len =min (transfer->len,8U );u8 rem = transfer->len % len; +u64 cfg =0ULL ; loops = transfer->len / len;
@@ -258,8 +263,14 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx,if (loops >1 ) {fsi_spi_sequence_add (seq,SPI_FSI_SEQUENCE_BRANCH (idx)); - rc =fsi_spi_write_reg (ctx,SPI_FSI_COUNTER_CFG , -SPI_FSI_COUNTER_CFG_LOOPS (loops -1 )); + cfg =SPI_FSI_COUNTER_CFG_LOOPS (loops -1 ); +if (transfer->rx_buf) + cfg |=SPI_FSI_COUNTER_CFG_N2_RX | +SPI_FSI_COUNTER_CFG_N2_TX | +SPI_FSI_COUNTER_CFG_N2_IMPLICIT | +SPI_FSI_COUNTER_CFG_N2_RELOAD ; + + rc =fsi_spi_write_reg (ctx,SPI_FSI_COUNTER_CFG , cfg);if (rc)return rc; }
@@ -275,6 +286,7 @@ static int fsi_spi_transfer_data(struct fsi_spi *ctx, {int rc =0 ;u64 status =0ULL ; +u64 cfg =0ULL ;if (transfer->tx_buf) {int nb;
@@ -312,6 +324,16 @@ static int fsi_spi_transfer_data(struct fsi_spi *ctx,u64 in =0ULL ;u8 *rx = transfer->rx_buf; + rc =fsi_spi_read_reg (ctx,SPI_FSI_COUNTER_CFG , &cfg); +if (rc) +return rc; + +if (cfg &SPI_FSI_COUNTER_CFG_N2_IMPLICIT ) { + rc =fsi_spi_write_reg (ctx,SPI_FSI_DATA_TX ,0 ); +if (rc) +return rc; + } + while (transfer->len > recv) { do { rc = fsi_spi_read_reg(ctx, SPI_FSI_STATUS,
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 4/7 dt-bindings: fsi: fsi2spi: Document new restricted property
In reply to Eddie JamesAdd documentation for the "fsi2spi,restricted" property which indicates a controller shouldn't sequence loops and therefore has a smaller transfer size.
--- Documentation/devicetree/bindings/fsi/ibm,fsi2spi.yaml | 10 ++++++++++ 1 file changed, 10 insertions(+)
Documentation/devicetree/bindings/fsi/ibm,fsi2spi.yaml +10 -0
@@ -24,6 +24,16 @@ properties: items: - description: FSI slave address +patternProperties: + "^spi(@.*|-[0-9a-f])*$": + type: object + + properties: + fsi2spi,restricted: + description: indicates the controller should not use looping in the + sequencer and therefore has a smaller maximum transfer size + type: boolean + required: - compatible - reg
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 5/7 spi: fsi: Implement restricted size for certain controllers
In reply to Eddie JamesSome of the FSI-attached SPI controllers cannot use the loop command in programming the sequencer due to security requirements. Add a boolean devicetree property that describes this condition and restrict the size for these controllers. Also, add more transfers directly in the sequence up to the length of the sequence register.
--- drivers/spi/spi-fsi.c | 65 +++++++++++++++++++++++++++++++++++-------- 1 file changed, 53 insertions(+), 12 deletions(-)
drivers/spi/spi-fsi.c +53 -12
@@ -24,7 +24,8 @@#define SPI_FSI_BASE 0x70000#define SPI_FSI_INIT_TIMEOUT_MS 1000 -#define SPI_FSI_MAX_TRANSFER_SIZE 2048 +#define SPI_FSI_MAX_XFR_SIZE 2048 +#define SPI_FSI_MAX_XFR_SIZE_RESTRICTED 32#define SPI_FSI_ERROR 0x0#define SPI_FSI_COUNTER_CFG 0x1
@@ -74,6 +75,8 @@ struct fsi_spi {struct device *dev;/* SPI controller device */ struct fsi_device *fsi;/* FSI2SPI CFAM engine device */ u32 base; +size_t max_xfr_size; +bool restricted; };struct fsi_spi_sequence {
@@ -209,8 +212,12 @@ static int fsi_spi_reset(struct fsi_spi *ctx)if (rc)return rc; -return fsi_spi_write_reg (ctx,SPI_FSI_CLOCK_CFG , -SPI_FSI_CLOCK_CFG_RESET2 ); + rc =fsi_spi_write_reg (ctx,SPI_FSI_CLOCK_CFG , +SPI_FSI_CLOCK_CFG_RESET2 ); +if (rc) +return rc; + +return fsi_spi_write_reg (ctx,SPI_FSI_STATUS ,0ULL ); }static int fsi_spi_sequence_add (struct fsi_spi_sequence *seq,u8 val)
@@ -218,8 +225,8 @@ static int fsi_spi_sequence_add(struct fsi_spi_sequence *seq, u8 val)/* * Add the next byte of instruction to the 8-byte sequence register. * Then decrement the counter so that the next instruction will go in -* the right place. Return the number of "slots" left in the sequence -* register. +* the right place. Return the index of the slot we just filled in the +* sequence register. */ seq->data |= (u64 )val << seq->bit; seq->bit -=8 ;
@@ -237,9 +244,11 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx,struct fsi_spi_sequence *seq, struct spi_transfer *transfer) { + bool docfg = false;int loops;int idx;int rc; +u8 val =0 ;u8 len =min (transfer->len,8U );u8 rem = transfer->len % len;u64 cfg =0ULL ;
@@ -247,22 +256,42 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx, loops = transfer->len / len;if (transfer->tx_buf) { - idx =fsi_spi_sequence_add (seq, -SPI_FSI_SEQUENCE_SHIFT_OUT (len)); + val =SPI_FSI_SEQUENCE_SHIFT_OUT (len); + idx =fsi_spi_sequence_add (seq, val); +if (rem) rem =SPI_FSI_SEQUENCE_SHIFT_OUT (rem); }else if (transfer->rx_buf) { - idx =fsi_spi_sequence_add (seq, -SPI_FSI_SEQUENCE_SHIFT_IN (len)); + val =SPI_FSI_SEQUENCE_SHIFT_IN (len); + idx =fsi_spi_sequence_add (seq, val); +if (rem) rem =SPI_FSI_SEQUENCE_SHIFT_IN (rem); }else {return -EINVAL ; } +if (ctx->restricted) { +const int eidx = rem ?5 :6 ; + +while (loops >1 && idx <= eidx) { + idx =fsi_spi_sequence_add (seq, val); + loops--; + docfg = true; + } + +if (loops >1 ) { +dev_warn (ctx->dev,"No sequencer slots; aborting.\n" ); +return -EINVAL ; + } + } +if (loops >1 ) {fsi_spi_sequence_add (seq,SPI_FSI_SEQUENCE_BRANCH (idx)); + docfg = true; + } +if (docfg) { cfg =SPI_FSI_COUNTER_CFG_LOOPS (loops -1 );if (transfer->rx_buf) cfg |= SPI_FSI_COUNTER_CFG_N2_RX |
@@ -273,6 +302,8 @@ static int fsi_spi_sequence_transfer(struct fsi_spi *ctx, rc =fsi_spi_write_reg (ctx,SPI_FSI_COUNTER_CFG , cfg);if (rc)return rc; + }else { +fsi_spi_write_reg (ctx,SPI_FSI_COUNTER_CFG ,0ULL ); }if (rem)
@@ -429,7 +460,7 @@ static int fsi_spi_transfer_one_message(struct spi_controller *ctlr,/* Sequencer must do shift out (tx) first. */ if (!transfer->tx_buf || - transfer->len >SPI_FSI_MAX_TRANSFER_SIZE ) { + transfer->len > (ctx->max_xfr_size +8 )) { rc = -EINVAL ; goto error; }
@@ -453,7 +484,7 @@ static int fsi_spi_transfer_one_message(struct spi_controller *ctlr,/* Sequencer can only do shift in (rx) after tx. */ if (next->rx_buf) { -if (next->len >SPI_FSI_MAX_TRANSFER_SIZE ) { +if (next->len > ctx->max_xfr_size) { rc = -EINVAL ; goto error; }
@@ -498,7 +529,9 @@ static int fsi_spi_transfer_one_message(struct spi_controller *ctlr,static size_t fsi_spi_max_transfer_size (struct spi_device *spi) { -return SPI_FSI_MAX_TRANSFER_SIZE ; +struct fsi_spi *ctx =spi_controller_get_devdata (spi->controller); + +return ctx->max_xfr_size; }static int fsi_spi_probe (struct device *dev)
@@ -546,6 +579,14 @@ static int fsi_spi_probe(struct device *dev) ctx->fsi = fsi; ctx->base = base +SPI_FSI_BASE ; +if (of_property_read_bool (np,"fsi2spi,restricted" )) { + ctx->restricted = true; + ctx->max_xfr_size =SPI_FSI_MAX_XFR_SIZE_RESTRICTED ; + }else { + ctx->restricted = false; + ctx->max_xfr_size =SPI_FSI_MAX_XFR_SIZE ; + } + rc =devm_spi_register_controller (dev, ctlr);if (rc)spi_controller_put (ctlr);
Signature · 1 line
2.24.0
Eddie James
maintainerPATCH v2 6/7 spi: fsi: Check mux status before transfers
In reply to Eddie JamesThe SPI controllers are not accessible if the mux isn't set. Therefore, check the mux status before starting a transfer and fail out if it isn't set.
--- drivers/spi/spi-fsi.c | 40 +++++++++++++++++++++++++++------------- 1 file changed, 27 insertions(+), 13 deletions(-)
drivers/spi/spi-fsi.c +27 -13
@@ -12,6 +12,7 @@#define FSI_ENGID_SPI 0x23#define FSI_MBOX_ROOT_CTRL_8 0x2860 +#define FSI_MBOX_ROOT_CTRL_8_SPI 0xf0000000#define FSI2SPI_DATA0 0x00#define FSI2SPI_DATA1 0x04
@@ -84,6 +85,26 @@ struct fsi_spi_sequence {u64 data; }; +static int fsi_spi_check_mux (struct fsi_device *fsi,struct device *dev) +{ +int rc; +u32 root_ctrl_8; +__be32 root_ctrl_8_be; + + rc =fsi_slave_read (fsi->slave,FSI_MBOX_ROOT_CTRL_8 , &root_ctrl_8_be, +sizeof (root_ctrl_8_be)); +if (rc) +return rc; + + root_ctrl_8 =be32_to_cpu (root_ctrl_8_be); +dev_dbg (dev,"Root control register 8: %08x\n" , root_ctrl_8); +if ((root_ctrl_8 &FSI_MBOX_ROOT_CTRL_8_SPI ) == +FSI_MBOX_ROOT_CTRL_8_SPI ) +return 0 ; + +return -ENOLINK ; +} +static int fsi_spi_check_status (struct fsi_spi *ctx) {int rc;
@@ -449,11 +470,15 @@ static int fsi_spi_transfer_init(struct fsi_spi *ctx)static int fsi_spi_transfer_one_message (struct spi_controller *ctlr,struct spi_message *mesg) { -int rc =0 ; +int rc;u8 seq_slave =SPI_FSI_SEQUENCE_SEL_SLAVE (mesg->spi->chip_select +1 );struct spi_transfer *transfer;struct fsi_spi *ctx =spi_controller_get_devdata (ctlr); + rc =fsi_spi_check_mux (ctx->fsi, ctx->dev); +if (rc) +return rc; +list_for_each_entry (transfer, &mesg->transfers, transfer_list) {struct fsi_spi_sequence seq;struct spi_transfer *next =NULL ;
@@ -537,24 +562,13 @@ static size_t fsi_spi_max_transfer_size(struct spi_device *spi)static int fsi_spi_probe (struct device *dev) {int rc; -u32 root_ctrl_8;struct device_node *np;int num_controllers_registered =0 ;struct fsi_device *fsi =to_fsi_dev (dev); -/* -* Check the SPI mux before attempting to probe. If the mux isn't set -* then the SPI controllers can't access their slave devices. -*/ - rc =fsi_slave_read (fsi->slave,FSI_MBOX_ROOT_CTRL_8 , &root_ctrl_8, -sizeof (root_ctrl_8)); + rc =fsi_spi_check_mux (fsi, dev);if (rc) -return rc; - -if (!root_ctrl_8) { -dev_dbg (dev,"SPI mux not set, aborting probe.\n" );return -ENODEV ; - }for_each_available_child_of_node (dev->of_node, np) {u32 base;
Signature · 1 line
2.24.0
Re: [PATCH linux dev-5.7 v2 6/7] spi: fsi: Check mux status before transfers
Quoted 15 lines · Eddie James
The SPI controllers are not accessible if the mux isn't set. Therefore, check the mux status before starting a transfer and fail out if it isn't set.
Signed-off-by Eddie James <[email protected]>--- drivers/spi/spi-fsi.c | 40 +++++++++++++++++++++++++++------------- 1 file changed, 27 insertions(+), 13 deletions(-)drivers/spi/spi-fsi.c +0
-0
drivers/spi/spi-fsi.c line 15
Quoted 4 lines
drivers/spi/spi-fsi.c +1
-0@@ -12,6 +12,7 @@#define FSI_ENGID_SPI 0x23#define FSI_MBOX_ROOT_CTRL_8 0x2860 +#define FSI_MBOX_ROOT_CTRL_8_SPI 0xf0000000
Could be called SPI_MUX or something similar.
Quoted 15 lines
drivers/spi/spi-fsi.c +8
-0#define FSI2SPI_DATA0 0x00#define FSI2SPI_DATA1 0x04@@ -84,6 +85,26 @@ struct fsi_spi_sequence {u64 data; }; +static int fsi_spi_check_mux (struct fsi_device *fsi,struct device *dev) +{ +int rc; +u32 root_ctrl_8; +__be32 root_ctrl_8_be; + + rc =fsi_slave_read (fsi->slave,FSI_MBOX_ROOT_CTRL_8 , &root_ctrl_8_be, +sizeof (root_ctrl_8_be));
drivers/spi/spi-fsi.c struct fsi_spi_sequence { · line 105
Quoted 10 lines
drivers/spi/spi-fsi.c +10
-0+if (rc) +return rc; + + root_ctrl_8 =be32_to_cpu (root_ctrl_8_be); +dev_dbg (dev,"Root control register 8: %08x\n" , root_ctrl_8); +if ((root_ctrl_8 &FSI_MBOX_ROOT_CTRL_8_SPI ) == +FSI_MBOX_ROOT_CTRL_8_SPI ) +return 0 ; + +return -ENOLINK ;
We bubble this error up to device drivers. Are they ok with ENOLINK?
Quoted 33 lines
drivers/spi/spi-fsi.c +7
-4+} +static int fsi_spi_check_status (struct fsi_spi *ctx) {int rc;@@ -449,11 +470,15 @@ static int fsi_spi_transfer_init(struct fsi_spi *ctx)static int fsi_spi_transfer_one_message (struct spi_controller *ctlr,struct spi_message *mesg) { -int rc =0 ; +int rc;u8 seq_slave =SPI_FSI_SEQUENCE_SEL_SLAVE (mesg->spi->chip_select +1 );struct spi_transfer *transfer;struct fsi_spi *ctx =spi_controller_get_devdata (ctlr); + rc =fsi_spi_check_mux (ctx->fsi, ctx->dev); +if (rc) +return rc; +list_for_each_entry (transfer, &mesg->transfers, transfer_list) {struct fsi_spi_sequence seq;struct spi_transfer *next =NULL ;@@ -537,24 +562,13 @@ static size_t fsi_spi_max_transfer_size(struct spi_device *spi)static int fsi_spi_probe (struct device *dev) {int rc; -u32 root_ctrl_8;struct device_node *np;int num_controllers_registered =0 ;struct fsi_device *fsi =to_fsi_dev (dev); -/* -* Check the SPI mux before attempting to probe. If the mux isn't set
drivers/spi/spi-fsi.c static size_t fsi_spi_max_transfer_size(struct spi_device *spi) · line 555
Quoted 10 lines
drivers/spi/spi-fsi.c +1
-8-* then the SPI controllers can't access their slave devices. -*/ - rc =fsi_slave_read (fsi->slave,FSI_MBOX_ROOT_CTRL_8 , &root_ctrl_8, -sizeof (root_ctrl_8)); + rc =fsi_spi_check_mux (fsi, dev);if (rc) -return rc; - -if (!root_ctrl_8) { -dev_dbg (dev,"SPI mux not set, aborting probe.\n" );
Should we make this a dev_warn so it's clear why the driver did not "work"?
Quoted 9 lines
drivers/spi/spi-fsi.c +0
-1return -ENODEV ; - }for_each_available_child_of_node (dev->of_node, np) {u32 base;--
2.24.0
Eddie James
maintainerPATCH v2 7/7 eeprom: at25: Split reads into chunks and cap write size
In reply to Eddie JamesFrom: Brad Bishop <[email protected]>
Make use of spi_max_transfer_size to avoid requesting transfers that are too large for some spi controllers.
--- drivers/misc/eeprom/at25.c | 94 ++++++++++++++++++++++---------------- 1 file changed, 54 insertions(+), 40 deletions(-)
drivers/misc/eeprom/at25.c +54 -40
@@ -64,12 +64,17 @@ static int at25_ee_read(void *priv, unsigned int offset, {struct at25_data *at25 = priv;char *buf = val; +size_t max_chunk =spi_max_transfer_size (at25->spi); +size_t num_msgs =DIV_ROUND_UP (count, max_chunk); +size_t nr_bytes =0 ;u8 command[EE_MAXADDRLEN +1 ];u8 *cp;ssize_t status;struct spi_transfer t[2 ];struct spi_message m;u8 instr; +unsigned int msg_offset; +size_t msg_count;if (unlikely (offset >= at25->chip.byte_len))return -EINVAL ;
@@ -78,57 +83,64 @@ static int at25_ee_read(void *priv, unsigned int offset,if (unlikely (!count))return -EINVAL ; - cp = command; - - instr =AT25_READ ; -if (at25->chip.flags &EE_INSTR_BIT3_IS_ADDR ) -if (offset >= (1U << (at25->addrlen *8 ))) - instr |=AT25_INSTR_BIT3 ; - *cp++ = instr; - -/* 8/16/24-bit address is written MSB first */ -switch (at25->addrlen) { -default :/* case 3 */ - *cp++ = offset >>16 ; -/* fall through */ -case 2 : - *cp++ = offset >>8 ; -/* fall through */ -case 1 : -case 0 :/* can't happen: for better codegen */ - *cp++ = offset >>0 ; - } + msg_offset = (unsigned int )offset; + msg_count =min (count, max_chunk); +while (num_msgs) { + cp = command; -spi_message_init (&m); -memset (t,0 ,sizeof (t)); + instr =AT25_READ ; +if (at25->chip.flags &EE_INSTR_BIT3_IS_ADDR ) +if (msg_offset >= (1U << (at25->addrlen *8 ))) + instr |=AT25_INSTR_BIT3 ; + *cp++ = instr; - t[0 ].tx_buf = command; - t[0 ].len = at25->addrlen +1 ; -spi_message_add_tail (&t[0 ], &m); +/* 8/16/24-bit address is written MSB first */ +switch (at25->addrlen) { +default :/* case 3 */ + *cp++ = msg_offset >>16 ; +/* fall through */ +case 2 : + *cp++ = msg_offset >>8 ; +/* fall through */ +case 1 : +case 0 :/* can't happen: for better codegen */ + *cp++ = msg_offset >>0 ; + } - t[1 ].rx_buf = buf; - t[1 ].len = count; -spi_message_add_tail (&t[1 ], &m); +spi_message_init (&m); +memset (t,0 ,sizeof (t)); -mutex_lock (&at25->lock); + t[0 ].tx_buf = command; + t[0 ].len = at25->addrlen +1 ; +spi_message_add_tail (&t[0 ], &m); -/* Read it all at once. -* -* REVISIT that's potentially a problem with large chips, if -* other devices on the bus need to be accessed regularly or -* this chip is clocked very slowly -*/ - status =spi_sync (at25->spi, &m); -dev_dbg (&at25->spi->dev,"read %zu bytes at %d --> %zd\n" , - count, offset, status); + t[1 ].rx_buf = buf + nr_bytes; + t[1 ].len = msg_count; +spi_message_add_tail (&t[1 ], &m); -mutex_unlock (&at25->lock); -return status; +mutex_lock (&at25->lock); + + status =spi_sync (at25->spi, &m); + +mutex_unlock (&at25->lock); + +if (status) +return status; + + --num_msgs; + msg_offset += msg_count; + nr_bytes += msg_count; + } + +dev_dbg (&at25->spi->dev,"read %zu bytes at %d\n" , + count, offset); +return 0 ; }static int at25_ee_write (void *priv,unsigned int off,void *val,size_t count) {struct at25_data *at25 = priv; +size_t maxsz =spi_max_transfer_size (at25->spi);const char *buf = val;int status =0 ;unsigned buf_size;
@@ -191,6 +203,8 @@ static int at25_ee_write(void *priv, unsigned int off, void *val, size_t count) segment = buf_size - (offset % buf_size);if (segment > count) segment = count; +if (segment > maxsz) + segment = maxsz;memcpy (cp, buf, segment); status =spi_write (at25->spi, bounce, segment + at25->addrlen +1 );
Signature · 1 line
2.24.0
Quoted 7 lines · Eddie James
This series implements a number of fixes for the FSI-attached SPI controller driver and the AT25 eeprom driver.
Changes since v1:
- Add patch 6 for checking the mux status before transfers
- Minor fixes to at25 driver formatting and such.
I've merged this series into dev-5.8. Please consider my review comments and send the series upstream.
Cheers,
Joel
Quoted 20 lines
Brad Bishop (4):
spi: fsi: Handle 9 to 15 byte transfers lengths
spi: fsi: Fix clock running too fast
spi: fsi: Fix use of the bneq+ sequencer instruction
eeprom: at25: Split reads into chunks and cap write sizeEddie James (3):
dt-bindings: fsi: fsi2spi: Document new restricted property
spi: fsi: Implement restricted size for certain controllers
spi: fsi: Check mux status before transfers.../devicetree/bindings/fsi/ibm,fsi2spi.yaml | 10 ++ drivers/misc/eeprom/at25.c | 94 +++++++----- drivers/spi/spi-fsi.c | 139 ++++++++++++++---- 3 files changed, 172 insertions(+), 71 deletions(-)--
2.24.0
Eddie James
maintainerQuoted 9 lines · Joel Stanley
Quoted 6 lines · Eddie James
This series implements a number of fixes for the FSI-attached SPI controller driver and the AT25 eeprom driver.
Changes since v1:
- Add patch 6 for checking the mux status before transfers
- Minor fixes to at25 driver formatting and such.I've merged this series into dev-5.8. Please consider my review comments and send the series upstream.
Thanks! Will do.
Eddie
Quoted 24 lines
Cheers,
Joel
Quoted 19 lines
Brad Bishop (4):
spi: fsi: Handle 9 to 15 byte transfers lengths
spi: fsi: Fix clock running too fast
spi: fsi: Fix use of the bneq+ sequencer instruction
eeprom: at25: Split reads into chunks and cap write sizeEddie James (3):
dt-bindings: fsi: fsi2spi: Document new restricted property
spi: fsi: Implement restricted size for certain controllers
spi: fsi: Check mux status before transfers.../devicetree/bindings/fsi/ibm,fsi2spi.yaml | 10 ++ drivers/misc/eeprom/at25.c | 94 +++++++----- drivers/spi/spi-fsi.c | 139 ++++++++++++++---- 3 files changed, 172 insertions(+), 71 deletions(-)--
2.24.0