hw/ssi: imx_spi: Fix various bugs in the imx_spi model
v8 cover hw/ssi: imx_spi: Fix various bugs in the imx_spi model
From: Bin Meng <[email protected]>
This v8 series is based on the following 2 versions:
- v5 series sent from Bin
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
- v7 series sent from Philippe
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612
This series fixes a bunch of bugs in current implementation of the imx spi controller, including the following issues:
- remove imx_spi_update_irq() in imx_spi_reset()
- chip select signal was not lower down when spi controller is disabled
- round up the tx burst length to be multiple of 8
- transfer incorrect data when the burst length is larger than 32 bit
- spi controller tx and rx fifo endianness is incorrect
- remove pointless variable (s->burst_length) initialization (Philippe)
- rework imx_spi_reset() to keep CONREG register value (Philippe)
- rework imx_spi_read() to handle block disabled (Philippe)
- rework imx_spi_write() to handle block disabled (Philippe)
Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7 (interrupt mode).
Changes in v8:
- keep the controller disable logic in the ECSPI_CONREG case in imx_spi_write()
Changes in v7:
- remove the RFC tag
Changes in v6:
- new patch: [RFC] remove pointless variable initialization
- new patch: [RFC] rework imx_spi_reset() to keep CONREG register value
- new patch: [RFC] rework imx_spi_read() to handle block disabled
- new patch: [RFC] rework imx_spi_write() to handle block disabled
Changes in v5:
- rename imx_spi_hard_reset() to imx_spi_soft_reset()
- round up the burst length to be multiple of 8
Changes in v4:
- adujst the patch 2,3 order
- rename imx_spi_soft_reset() to imx_spi_hard_reset() to avoid confusion
- s/normal/common/ in the commit message
- log the burst length value in the log message
Changes in v3:
- new patch: remove imx_spi_update_irq() in imx_spi_reset()
- Move the chip selects disable out of imx_spi_reset()
- new patch: log unimplemented burst length
- Simplify the tx fifo endianness handling
Changes in v2:
- Fix the "Fixes" tag in the commit message
- Use ternary operator as Philippe suggested
Bin Meng (5):
hw/ssi: imx_spi: Use a macro for number of chip selects supported hw/ssi: imx_spi: Remove imx_spi_update_irq() in imx_spi_reset() hw/ssi: imx_spi: Round up the burst length to be multiple of 8 hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic hw/ssi: imx_spi: Correct tx and rx fifo endianness
Philippe Mathieu-Daudé (4):
hw/ssi: imx_spi: Remove pointless variable initialization hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value hw/ssi: imx_spi: Rework imx_spi_read() to handle block disabled hw/ssi: imx_spi: Rework imx_spi_write() to handle block disabled
Xuzhou Cheng (1):
hw/ssi: imx_spi: Disable chip selects when controller is disabled
include/hw/ssi/imx_spi.h | 5 +- hw/ssi/imx_spi.c | 138 +++++++++++++++++++++++++++++------------------ 2 files changed, 90 insertions(+), 53 deletions(-)
Signature · 1 line
2.7.4
PATCH v8 1/10 hw/ssi: imx_spi: Use a macro for number of chip selects supported
From: Bin Meng <[email protected]>
Avoid using a magic number (4) everywhere for the number of chip selects supported.
---
(no changes since v1)
include/hw/ssi/imx_spi.h | 5 ++++- hw/ssi/imx_spi.c | 4 ++-- 2 files changed, 6 insertions(+), 3 deletions(-)
include/hw/ssi/imx_spi.h +4 -1
@@ -77,6 +77,9 @@#define EXTRACT (value, name) extract32(value, name##_SHIFT, name##_LENGTH) +/* number of chip selects supported */ +#define ECSPI_NUM_CS 4 +#define TYPE_IMX_SPI "imx.spi"OBJECT_DECLARE_SIMPLE_TYPE (IMXSPIState,IMX_SPI )
@@ -89,7 +92,7 @@ struct IMXSPIState {qemu_irq irq; -qemu_irq cs_lines[4 ]; +qemu_irq cs_lines[ECSPI_NUM_CS ];SSIBus *bus;
hw/ssi/imx_spi.c +2 -2
@@ -361,7 +361,7 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,/* We are in master mode */ -for (i =0 ; i <4 ; i++) { +for (i =0 ; i <ECSPI_NUM_CS ; i++) {qemu_set_irq (s->cs_lines[i], i ==imx_spi_selected_channel (s) ?0 :1 ); }
@@ -424,7 +424,7 @@ static void imx_spi_realize(DeviceState *dev, Error **errp)sysbus_init_mmio (SYS_BUS_DEVICE (dev), &s->iomem);sysbus_init_irq (SYS_BUS_DEVICE (dev), &s->irq); -for (i =0 ; i <4 ; ++i) { +for (i =0 ; i <ECSPI_NUM_CS ; ++i) {sysbus_init_irq (SYS_BUS_DEVICE (dev), &s->cs_lines[i]); }
Signature · 1 line
2.7.4
PATCH v8 2/10 hw/ssi: imx_spi: Remove imx_spi_update_irq() in imx_spi_reset()
In reply to Bin MengFrom: Bin Meng <[email protected]>
Usually the approach is that the device on the other end of the line is going to reset its state anyway, so there's no need to actively signal an irq line change during the reset hook.
Move imx_spi_update_irq() out of imx_spi_reset(), to a new function imx_spi_soft_reset() that is called when the controller is disabled.
---
(no changes since v5)
Changes in v5:
- rename imx_spi_hard_reset() to imx_spi_soft_reset()
Changes in v4:
- adujst the patch 2,3 order
- rename imx_spi_soft_reset() to imx_spi_hard_reset() to avoid confusion
Changes in v3:
- new patch: remove imx_spi_update_irq() in imx_spi_reset()
hw/ssi/imx_spi.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)
hw/ssi/imx_spi.c +10 -4
@@ -241,11 +241,16 @@ static void imx_spi_reset(DeviceState *dev)imx_spi_rxfifo_reset (s);imx_spi_txfifo_reset (s); -imx_spi_update_irq (s); - s->burst_length =0 ; } +static void imx_spi_soft_reset (IMXSPIState *s) +{ +imx_spi_reset (DEVICE (s)); + +imx_spi_update_irq (s); +} +static uint64_t imx_spi_read (void *opaque,hwaddr offset,unsigned size) {uint32_t value =0 ;
@@ -351,8 +356,9 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value, s->regs[ECSPI_CONREG ] = value;if (!imx_spi_is_enabled (s)) { -/* device is disabled, so this is a reset */ -imx_spi_reset (DEVICE (s)); +/* device is disabled, so this is a soft reset */ +imx_spi_soft_reset (s); +return ; }
Signature · 1 line
2.7.4
Re: [PATCH v8 02/10] hw/ssi: imx_spi: Remove imx_spi_update_irq() in imx_spi_reset()
Quoted 7 lines · Bin Meng
From: Bin Meng <[email protected]>
Usually the approach is that the device on the other end of the line is going to reset its state anyway, so there's no need to actively signal an irq line change during the reset hook.
Quoted 6 lines
Move
imx_spi_update_irq()out ofimx_spi_reset(), to a new functionimx_spi_soft_reset()that is called when the controller is disabled.Signed-off-by Bin Meng <[email protected]>---
thanks
-- PMM
PATCH v8 3/10 hw/ssi: imx_spi: Remove pointless variable initialization
In reply to Bin MengFrom: Philippe Mathieu-Daudé <[email protected]>
'burst_length' is cleared in imx_spi_reset(), which is called after imx_spi_realize(). Remove the initialization to simplify.
Message-Id: <[email protected]>
---
(no changes since v7)
Changes in v7:
- remove the RFC tag
Changes in v6:
- new patch: [RFC] remove pointless variable initialization
hw/ssi/imx_spi.c | 2 -- 1 file changed, 2 deletions(-)
hw/ssi/imx_spi.c +0 -2
@@ -434,8 +434,6 @@ static void imx_spi_realize(DeviceState *dev, Error **errp)sysbus_init_irq (SYS_BUS_DEVICE (dev), &s->cs_lines[i]); } - s->burst_length =0 ; -fifo32_create (&s->tx_fifo,ECSPI_FIFO_SIZE );fifo32_create (&s->rx_fifo,ECSPI_FIFO_SIZE ); }
Signature · 1 line
2.7.4
PATCH v8 4/10 hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
In reply to Bin MengFrom: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the ECSPI_CONREG. It is initialized to zero when the instance is created.
Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)
Message-Id: <[email protected]>
---
(no changes since v7)
Changes in v7:
- remove the RFC tag
Changes in v6:
- new patch: [RFC] rework imx_spi_reset() to keep CONREG register value
hw/ssi/imx_spi.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-)
hw/ssi/imx_spi.c +14 -3
@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG : +/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }imx_spi_rxfifo_reset (s);imx_spi_txfifo_reset (s);
Signature · 1 line
2.7.4
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 28 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the
ECSPI_CONREG. It is initialized to zero when the instance is created.Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)hw/ssi/imx_spi.c +4
-3@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG :
hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 250
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }
This retains the CONREG register value for both:
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
thanks
-- PMM
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 42 lines · Peter Maydell
Quoted 28 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the
ECSPI_CONREG. It is initialized to zero when the instance is created.Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)hw/ssi/imx_spi.c +4
-3@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG :hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 250
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }This retains the CONREG register value for both:
Quoted 8 lines
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
The POR value of CONREG is zero, which should be the default value, no?
Regards,
Bin
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 53 lines · Bin Meng
Quoted 42 lines · Peter Maydell
Quoted 28 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the
ECSPI_CONREG. It is initialized to zero when the instance is created.Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)hw/ssi/imx_spi.c +4
-3@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG :hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 250
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }This retains the CONREG register value for both:
Quoted 8 lines
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
Quoted 1 line
The POR value of CONREG is zero, which should be the default value, no?
But you're not setting it to zero here, you're leaving it with whatever value it had before. (That's correct for soft reset, but wrong for power-on.)
thanks
-- PMM
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 54 lines · Peter Maydell
Quoted 53 lines · Bin Meng
Quoted 50 lines · Peter Maydell
Quoted 28 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the
ECSPI_CONREG. It is initialized to zero when the instance is created.Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)hw/ssi/imx_spi.c +4
-3@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG :hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 250
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }This retains the CONREG register value for both:
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
Quoted 5 lines
Quoted 1 line
The POR value of CONREG is zero, which should be the default value, no?
But you're not setting it to zero here, you're leaving it with whatever value it had before. (That's correct for soft reset, but wrong for power-on.)
zero value on power-on is what I tried to describe as "It is initialized to zero when the instance is created."
Most of the codebase assumes QOM provides a zero-initialized instance state. Do you think it should be explicit?
Regards,
Phil.
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 15 lines · Philippe Mathieu-Daudé
Quoted 13 lines · Peter Maydell
Quoted 12 lines · Bin Meng
Quoted 9 lines · Peter Maydell
This retains the CONREG register value for both:
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
The POR value of CONREG is zero, which should be the default value, no?
But you're not setting it to zero here, you're leaving it with whatever value it had before. (That's correct for soft reset, but wrong for power-on.)
zero value on power-on is what I tried to describe as "It is initialized to zero when the instance is created."
Yes, but QOM device reset does not happen just once at startup and not thereafter. Consider:
* user starts QEMU
* QOM devices are created and realized
* QOM device reset happens
-- CONREG is zero here because QOM structs are zero-initialized
* guest runs
* guest modifies CONREG from its initial value
* system reset is requested (perhaps by user, perhaps by guest writing some register or another)
* QOM device reset happens
-- CONREG is not zero here, so reset must clear it
thanks
-- PMM
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 28 lines · Peter Maydell
Quoted 15 lines · Philippe Mathieu-Daudé
Quoted 13 lines · Peter Maydell
Quoted 12 lines · Bin Meng
Quoted 9 lines · Peter Maydell
This retains the CONREG register value for both:
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.Quoted 8 lines
Quoted 5 lines
Quoted 1 line
The POR value of CONREG is zero, which should be the default value, no?
But you're not setting it to zero here, you're leaving it with whatever value it had before. (That's correct for soft reset, but wrong for power-on.)
zero value on power-on is what I tried to describe as "It is initialized to zero when the instance is created."
Yes, but QOM device reset does not happen just once at startup and not thereafter. Consider:
Quoted 10 lines
* user starts QEMU
* QOM devices are created and realized
* QOM device reset happens
-- CONREG is zero here because QOM structs are zero-initialized
* guest runs
* guest modifies CONREG from its initial value
* system reset is requested (perhaps by user, perhaps by guest writing some register or another)
* QOM device reset happens
-- CONREG is not zero here, so reset must clear it
Oh I totally missed that :S
Bin, I'd correct this as:
- extract imx_spi_soft_reset(IMXSPIState *s) from imx_spi_reset()
- zero-initialize CONREG in imx_spi_reset().
static void imx_spi_soft_reset(IMXSPIState *s)
{
...
}
static void imx_spi_reset(DeviceState *dev)
{
IMXSPIState *s = IMX_SPI(dev);
s->regs[ECSPI_CONREG] = 0;
imx_spi_soft_reset(s);
}What do you think?
Phil.
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 13 lines · Philippe Mathieu-Daudé
Oh I totally missed that :S
Bin, I'd correct this as:
- extract
imx_spi_soft_reset(IMXSPIState *s)fromimx_spi_reset()
- zero-initialize CONREG inimx_spi_reset().static void
imx_spi_soft_reset(IMXSPIState *s)
{
...
}static void
imx_spi_reset(DeviceState *dev)
Quoted 8 lines
{ IMXSPIState *s = IMX_SPI(dev); s->regs[ECSPI_CONREG] = 0; imx_spi_soft_reset(s); }What do you think?
That doesn't give you anywhere to put the imx_spi_update_irq() call, which must happen only on soft reset and not on DeviceState reset. You could do one of:
* have a 'common reset' function that does most of this, plus an imx_spi_reset which clears CONREG and calls common reset and an imx_spi_soft_reset which calls common reset and imx_spi_update_irq()
* have imx_spi_soft_reset save the old CONREG in a local variable before calling imx_spi_reset and then restore it to s->regs
thanks
-- PMM
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 23 lines · Peter Maydell
Quoted 13 lines · Philippe Mathieu-Daudé
Oh I totally missed that :S
Bin, I'd correct this as:
- extract
imx_spi_soft_reset(IMXSPIState *s)fromimx_spi_reset()
- zero-initialize CONREG inimx_spi_reset().static void
imx_spi_soft_reset(IMXSPIState *s)
{
...
}static void
imx_spi_reset(DeviceState *dev)Quoted 8 lines
{ IMXSPIState *s = IMX_SPI(dev); s->regs[ECSPI_CONREG] = 0; imx_spi_soft_reset(s); }What do you think?
Quoted 10 lines
That doesn't give you anywhere to put the
imx_spi_update_irq()call, which must happen only on soft reset and not on DeviceState reset. You could do one of:
* have a 'common reset' function that does most of this, plus animx_spi_resetwhich clears CONREG and calls common reset and animx_spi_soft_resetwhich calls common reset andimx_spi_update_irq()
* haveimx_spi_soft_resetsave the old CONREG in a local variable before callingimx_spi_resetand then restore it tos->regs
Long term maintenance I'd prefer the 'common reset' approach (but this is probably subjective to my view on the hardware, since this is software, the 2nd approach is also valid but harder to represent thinking of hardware).
Bin, can you send a v9? (using the approach you prefer)
Thanks,
Phil.
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
Quoted 36 lines · Philippe Mathieu-Daudé
Quoted 23 lines · Peter Maydell
Quoted 13 lines · Philippe Mathieu-Daudé
Oh I totally missed that :S
Bin, I'd correct this as:
- extract
imx_spi_soft_reset(IMXSPIState *s)fromimx_spi_reset()
- zero-initialize CONREG inimx_spi_reset().static void
imx_spi_soft_reset(IMXSPIState *s)
{
...
}static void
imx_spi_reset(DeviceState *dev)Quoted 8 lines
{ IMXSPIState *s = IMX_SPI(dev); s->regs[ECSPI_CONREG] = 0; imx_spi_soft_reset(s); }What do you think?
Quoted 10 lines
That doesn't give you anywhere to put the
imx_spi_update_irq()call, which must happen only on soft reset and not on DeviceState reset. You could do one of:
* have a 'common reset' function that does most of this, plus animx_spi_resetwhich clears CONREG and calls common reset and animx_spi_soft_resetwhich calls common reset andimx_spi_update_irq()
* haveimx_spi_soft_resetsave the old CONREG in a local variable before callingimx_spi_resetand then restore it tos->regs
Quoted 7 lines
Long term maintenance I'd prefer the 'common reset' approach (but this is probably subjective to my view on the hardware, since this is software, the 2nd approach is also valid but harder to represent thinking of hardware).
Bin, can you send a v9? (using the approach you prefer)
Yes, I will send a v9.
Thanks Peter for the explanation on the QOM device reset scenarios.
Regards,
Bin
Re: [PATCH v8 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
In reply to Peter MaydellQuoted 55 lines · Peter Maydell
Quoted 53 lines · Bin Meng
Quoted 50 lines · Peter Maydell
Quoted 28 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, all registers are reset with the exception of the
ECSPI_CONREG. It is initialized to zero when the instance is created.Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)hw/ssi/imx_spi.c +4
-3@@ -231,12 +231,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)static void imx_spi_reset (DeviceState *dev) {IMXSPIState *s =IMX_SPI (dev); +int i;DPRINTF ("\n" ); -memset (s->regs,0 ,sizeof (s->regs)); - - s->regs[ECSPI_STATREG ] =0x00000003 ; +for (i =0 ; i <ARRAY_SIZE (s->regs); i++) { +switch (i) { +case ECSPI_CONREG :hw/ssi/imx_spi.c static void imx_spi_flush_txfifo(IMXSPIState *s) · line 250
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+/* CONREG is not updated on reset */ +break ; +case ECSPI_STATREG : + s->regs[i] =0x00000003 ; +break ; +default : + s->regs[i] =0 ; +break ; + } + }This retains the CONREG register value for both:
* 'soft' reset caused by write to device register to disable the block
-- this is corrcet as per the datasheet quote
* 'power on' reset via TYPE_DEVICE's reset method
-- but in this case we should reset CONREG, because the Device reset method is like a complete device powercycle and should return the device state to what it was when QEMU was first started.
Quoted 5 lines
Quoted 1 line
The POR value of CONREG is zero, which should be the default value, no?
But you're not setting it to zero here, you're leaving it with whatever value it had before. (That's correct for soft reset, but wrong for power-on.)
I think that's ensured by object_initialize_with_type().
Regards,
Bin
PATCH v8 5/10 hw/ssi: imx_spi: Rework imx_spi_read() to handle block disabled
In reply to Bin MengFrom: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, it stay it is 'internal reset logic' (internal clocks are gated off). Reading any register returns its reset value. Only update this value if the device is enabled.
Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)
Message-Id: <[email protected]>
---
(no changes since v7)
Changes in v7:
- remove the RFC tag
Changes in v6:
- new patch: [RFC] rework imx_spi_read() to handle block disabled
hw/ssi/imx_spi.c | 60 +++++++++++++++++++++++++++----------------------------- 1 file changed, 29 insertions(+), 31 deletions(-)
hw/ssi/imx_spi.c +29 -31
@@ -274,42 +274,40 @@ static uint64_t imx_spi_read(void *opaque, hwaddr offset, unsigned size)return 0 ; } -switch (index) { -case ECSPI_RXDATA : -if (!imx_spi_is_enabled (s)) { - value =0 ; - }else if (fifo32_is_empty (&s->rx_fifo)) { -/* value is undefined */ - value =0xdeadbeef ; - }else { -/* read from the RX FIFO */ - value =fifo32_pop (&s->rx_fifo); - } - -break ; -case ECSPI_TXDATA : -qemu_log_mask (LOG_GUEST_ERROR ,"[%s]%s: Trying to read from TX FIFO\n" , -TYPE_IMX_SPI , __func__); - -/* Reading from TXDATA gives 0 */ - -break ; -case ECSPI_MSGDATA : -qemu_log_mask (LOG_GUEST_ERROR ,"[%s]%s: Trying to read from MSG FIFO\n" , -TYPE_IMX_SPI , __func__); + value = s->regs[index]; + +if (imx_spi_is_enabled (s)) { +switch (index) { +case ECSPI_RXDATA : +if (fifo32_is_empty (&s->rx_fifo)) { +/* value is undefined */ + value =0xdeadbeef ; + }else { +/* read from the RX FIFO */ + value =fifo32_pop (&s->rx_fifo); + } +break ; +case ECSPI_TXDATA : +qemu_log_mask (LOG_GUEST_ERROR , +"[%s]%s: Trying to read from TX FIFO\n" , +TYPE_IMX_SPI , __func__); -/* Reading from MSGDATA gives 0 */ +/* Reading from TXDATA gives 0 */ +break ; +case ECSPI_MSGDATA : +qemu_log_mask (LOG_GUEST_ERROR , +"[%s]%s: Trying to read from MSG FIFO\n" , +TYPE_IMX_SPI , __func__); +/* Reading from MSGDATA gives 0 */ +break ; +default : +break ; + } -break ; -default : - value = s->regs[index]; -break ; +imx_spi_update_irq (s); } -DPRINTF ("reg[%s] => 0x%" PRIx32"\n" ,imx_spi_reg_name (index), value); -imx_spi_update_irq (s); -return (uint64_t )value; }
Signature · 1 line
2.7.4
PATCH v8 6/10 hw/ssi: imx_spi: Rework imx_spi_write() to handle block disabled
In reply to Bin MengFrom: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, only the ECSPI_CONREG register can be modified. Setting the EN bit enabled the device, clearing it "disables the block and resets the internal logic with the exception of the ECSPI_CONREG" register.
Ignore all other registers write except ECSPI_CONREG when the block is disabled.
Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)
Message-Id: <[email protected]>
---
Changes in v8:
- keep the controller disable logic in the ECSPI_CONREG case in imx_spi_write()
Changes in v7:
- remove the RFC tag
Changes in v6:
- new patch: [RFC] rework imx_spi_write() to handle block disabled
hw/ssi/imx_spi.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-)
hw/ssi/imx_spi.c +9 -4
@@ -327,6 +327,14 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,DPRINTF ("reg[%s] <= 0x%" PRIx32"\n" ,imx_spi_reg_name (index), (uint32_t )value); +if (!imx_spi_is_enabled (s)) { +/* Block is disabled */ +if (index !=ECSPI_CONREG ) { +/* Ignore access */ +return ; + } + } + change_mask = s->regs[index] ^ value;switch (index) {
@@ -335,10 +343,7 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value, TYPE_IMX_SPI, __func__);break ;case ECSPI_TXDATA : -if (!imx_spi_is_enabled (s)) { -/* Ignore writes if device is disabled */ -break ; - }else if (fifo32_is_full (&s->tx_fifo)) { +if (fifo32_is_full (&s->tx_fifo)) {/* Ignore writes if queue is full */ break ; }
Signature · 1 line
2.7.4
Re: [PATCH v8 06/10] hw/ssi: imx_spi: Rework imx_spi_write() to handle block disabled
Quoted 11 lines · Bin Meng
From: Philippe Mathieu-Daudé <[email protected]>
When the block is disabled, only the
ECSPI_CONREGregister can be modified. Setting the EN bit enabled the device, clearing it "disables the block and resets the internal logic with the exception of theECSPI_CONREG" register.Ignore all other registers write except
ECSPI_CONREGwhen the block is disabled.
Quoted 6 lines
Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
chapter 21.7.3: Control Register (ECSPIx_CONREG)Signed-off-by Philippe Mathieu-Daudé <[email protected]>Message-Id: <[email protected]>
Signed-off-by Bin Meng <[email protected]>
thanks
-- PMM
PATCH v8 7/10 hw/ssi: imx_spi: Disable chip selects when controller is disabled
In reply to Bin MengFrom: Xuzhou Cheng <[email protected]>
When a write to ECSPI_CONREG register to disable the SPI controller, imx_spi_reset() is called to reset the controller, but chip select lines should have been disabled, otherwise the state machine of any devices (e.g.: SPI flashes) connected to the SPI master is stuck to its last state and responds incorrectly to any follow-up commands.
---
(no changes since v3)
Changes in v3:
- Move the chip selects disable out of imx_spi_reset()
Changes in v2:
- Fix the "Fixes" tag in the commit message
hw/ssi/imx_spi.c | 6 ++++++ 1 file changed, 6 insertions(+)
hw/ssi/imx_spi.c +6 -0
@@ -257,9 +257,15 @@ static void imx_spi_reset(DeviceState *dev)static void imx_spi_soft_reset (IMXSPIState *s) { +int i; +imx_spi_reset (DEVICE (s));imx_spi_update_irq (s); + +for (i =0 ; i <ECSPI_NUM_CS ; i++) { +qemu_set_irq (s->cs_lines[i],1 ); + } }static uint64_t imx_spi_read (void *opaque,hwaddr offset,unsigned size)
Signature · 1 line
2.7.4
PATCH v8 8/10 hw/ssi: imx_spi: Round up the burst length to be multiple of 8
In reply to Bin MengFrom: Bin Meng <[email protected]>
Current implementation of the imx spi controller expects the burst length to be multiple of 8, which is the most common use case.
In case the burst length is not what we expect, log it to give user a chance to notice it, and round it up to be multiple of 8.
---
(no changes since v5)
Changes in v5:
- round up the burst length to be multiple of 8
Changes in v4:
- s/normal/common/ in the commit message
- log the burst length value in the log message
Changes in v3:
- new patch: log unimplemented burst length
hw/ssi/imx_spi.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-)
hw/ssi/imx_spi.c +14 -1
@@ -128,7 +128,20 @@ static uint8_t imx_spi_selected_channel(IMXSPIState *s)static uint32_t imx_spi_burst_length (IMXSPIState *s) { -return EXTRACT (s->regs[ECSPI_CONREG ],ECSPI_CONREG_BURST_LENGTH ) +1 ; +uint32_t burst; + + burst =EXTRACT (s->regs[ECSPI_CONREG ],ECSPI_CONREG_BURST_LENGTH ) +1 ; +if (burst %8 ) { +qemu_log_mask (LOG_UNIMP , +"[%s]%s: burst length (%d) not multiple of 8!\n" , +TYPE_IMX_SPI , __func__, burst); + burst =ROUND_UP (burst,8 ); +qemu_log_mask (LOG_UNIMP , +"[%s]%s: burst length rounded up to %d; this may not work.\n" , +TYPE_IMX_SPI , __func__, burst); + } + +return burst; }static bool imx_spi_is_enabled (IMXSPIState *s)
Signature · 1 line
2.7.4
Re: [PATCH v8 08/10] hw/ssi: imx_spi: Round up the burst length to be multiple of 8
Quoted 17 lines · Bin Meng
From: Bin Meng <[email protected]>
Current implementation of the imx spi controller expects the burst length to be multiple of 8, which is the most common use case.
In case the burst length is not what we expect, log it to give user a chance to notice it, and round it up to be multiple of 8.
Signed-off-by Bin Meng <[email protected]>hw/ssi/imx_spi.c +1
-1@@ -128,7 +128,20 @@ static uint8_t imx_spi_selected_channel(IMXSPIState *s)static uint32_t imx_spi_burst_length (IMXSPIState *s) { -return EXTRACT (s->regs[ECSPI_CONREG ],ECSPI_CONREG_BURST_LENGTH ) +1 ; +uint32_t burst;
hw/ssi/imx_spi.c static uint8_t imx_spi_selected_channel(IMXSPIState *s) · line 141
Quoted 10 lines
hw/ssi/imx_spi.c +10
-0+ + burst =EXTRACT (s->regs[ECSPI_CONREG ],ECSPI_CONREG_BURST_LENGTH ) +1 ; +if (burst %8 ) { +qemu_log_mask (LOG_UNIMP , +"[%s]%s: burst length (%d) not multiple of 8!\n" , +TYPE_IMX_SPI , __func__, burst); + burst =ROUND_UP (burst,8 ); +qemu_log_mask (LOG_UNIMP , +"[%s]%s: burst length rounded up to %d; this may not work.\n" , +TYPE_IMX_SPI , __func__, burst);
It's friendlier to the user to do the LOG_UNIMP when the unsupported CONREG value is written, rather than here where it is used. That way the warning happens once, rather than every time the device transmits data.
Also, you could squash the warning down into one line, something like:
"[%s]%s: burst length %d not supported: rounding up to next multiple of 8\n"
rather than logging twice.
thanks
-- PMM
PATCH v8 9/10 hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic
In reply to Bin MengFrom: 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
---
(no changes since v2)
Changes in v2:
- Use ternary operator as Philippe suggested
hw/ssi/imx_spi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
hw/ssi/imx_spi.c +1 -1
@@ -191,7 +191,7 @@ 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 ) ? :32 ; rx =0 ;
Signature · 1 line
2.7.4
PATCH v8 10/10 hw/ssi: imx_spi: Correct tx and rx fifo endianness
In reply to Bin MengFrom: 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
---
(no changes since v3)
Changes in v3:
- Simplify the tx fifo endianness handling
hw/ssi/imx_spi.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-)
hw/ssi/imx_spi.c +2 -5
@@ -175,7 +175,6 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)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);
@@ -196,7 +195,7 @@ static void imx_spi_flush_txfifo(IMXSPIState *s) rx =0 ;while (tx_burst >0 ) { -uint8_t byte = tx &0xff ; +uint8_t byte = tx >> (tx_burst -8 );DPRINTF ("writing 0x%02x\n" , (uint32_t )byte);
@@ -205,13 +204,11 @@ 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 v8 10/10] hw/ssi: imx_spi: Correct tx and rx fifo endianness
Quoted 23 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.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
Quoted 2 lines
Fixes c906a3a01582 ("i.MX: Add the Freescale SPI Controller")Signed-off-by Bin Meng <[email protected]>
thanks
-- PMM
Quoted 23 lines · Bin Meng
From: Bin Meng <[email protected]>
This v8 series is based on the following 2 versions:
- v5 series sent from Bin
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
- v7 series sent from Philippe
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612This series fixes a bunch of bugs in current implementation of the imx spi controller, including the following issues:
- remove
imx_spi_update_irq()inimx_spi_reset()
- chip select signal was not lower down when spi controller is disabled
- round up the tx burst length to be multiple of 8
- transfer incorrect data when the burst length is larger than 32 bit
- spi controller tx and rx fifo endianness is incorrect
- remove pointless variable (s->burst_length) initialization (Philippe)
- reworkimx_spi_reset()to keep CONREG register value (Philippe)
- reworkimx_spi_read()to handle block disabled (Philippe)
- reworkimx_spi_write()to handle block disabled (Philippe)
Quoted 6 lines
Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7 (interrupt mode).
Changes in v8:
- keep the controller disable logic in theECSPI_CONREGcase inimx_spi_write()
Ping?
Quoted 25 lines · Bin Meng
Quoted 23 lines · Bin Meng
From: Bin Meng <[email protected]>
This v8 series is based on the following 2 versions:
- v5 series sent from Bin
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
- v7 series sent from Philippe
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612This series fixes a bunch of bugs in current implementation of the imx spi controller, including the following issues:
- remove
imx_spi_update_irq()inimx_spi_reset()
- chip select signal was not lower down when spi controller is disabled
- round up the tx burst length to be multiple of 8
- transfer incorrect data when the burst length is larger than 32 bit
- spi controller tx and rx fifo endianness is incorrect
- remove pointless variable (s->burst_length) initialization (Philippe)
- reworkimx_spi_reset()to keep CONREG register value (Philippe)
- reworkimx_spi_read()to handle block disabled (Philippe)
- reworkimx_spi_write()to handle block disabled (Philippe)
Quoted 8 lines
Quoted 6 lines
Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7 (interrupt mode).
Changes in v8:
- keep the controller disable logic in theECSPI_CONREGcase inimx_spi_write()Ping?
Could we get this applied soon if no more comments?
Regards,
Bin
Quoted 30 lines · Bin Meng
Quoted 28 lines · Bin Meng
Quoted 26 lines · Bin Meng
From: Bin Meng <[email protected]>
This v8 series is based on the following 2 versions:
- v5 series sent from Bin
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
- v7 series sent from Philippe
http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612This series fixes a bunch of bugs in current implementation of the imx spi controller, including the following issues:
- remove
imx_spi_update_irq()inimx_spi_reset()
- chip select signal was not lower down when spi controller is disabled
- round up the tx burst length to be multiple of 8
- transfer incorrect data when the burst length is larger than 32 bit
- spi controller tx and rx fifo endianness is incorrect
- remove pointless variable (s->burst_length) initialization (Philippe)
- reworkimx_spi_reset()to keep CONREG register value (Philippe)
- reworkimx_spi_read()to handle block disabled (Philippe)
- reworkimx_spi_write()to handle block disabled (Philippe)Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7 (interrupt mode).
Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
Changes in v8:
- keep the controller disable logic in theECSPI_CONREGcase inimx_spi_write()Ping?
Could we get this applied soon if no more comments?
Sorry, I think I missed this re-send. I've reviewed or left comments on the patches that were still unreviewed.
thanks
-- PMM