[U-Boot] [PATCH] zynq: fix I2C in SPL
The commit 'f48ef0d81aa837a33020f8d61abb3929ba613774' did break I2C support because requesting the clock for the I2C ip-block isn't supported during SPL.
To fixup this we add support requesting clocks for: - i2c0 - i2c1
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
---
drivers/clk/clk_zynq.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/clk/clk_zynq.c b/drivers/clk/clk_zynq.c index 482f093..9a4e949 100644 --- a/drivers/clk/clk_zynq.c +++ b/drivers/clk/clk_zynq.c @@ -434,6 +434,9 @@ static ulong zynq_clk_get_rate(struct clk *clk) case lqspi_clk ... pcap_clk: case sdio0_clk ... spi1_clk: return zynq_clk_get_peripheral_rate(priv, id, 0); + case i2c0_aper_clk ... i2c1_aper_clk: + return zynq_clk_get_cpu_rate(priv, cpu_1x_clk); + default: return -ENXIO; }
Hi,
first of all I would rephrase subject a little bit.
Something like: "ARM: zynq: Add missing i2c get_rate for fixing i2c SPL"
On 14. 02. 19 6:50, Hannes Schmelzer wrote:
The commit 'f48ef0d81aa837a33020f8d61abb3929ba613774' did break I2C support because requesting the clock for the I2C ip-block isn't supported during SPL.
To fixup this we add support requesting clocks for:
- i2c0
- i2c1
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
drivers/clk/clk_zynq.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/clk/clk_zynq.c b/drivers/clk/clk_zynq.c index 482f093..9a4e949 100644 --- a/drivers/clk/clk_zynq.c +++ b/drivers/clk/clk_zynq.c @@ -434,6 +434,9 @@ static ulong zynq_clk_get_rate(struct clk *clk) case lqspi_clk ... pcap_clk: case sdio0_clk ... spi1_clk: return zynq_clk_get_peripheral_rate(priv, id, 0);
- case i2c0_aper_clk ... i2c1_aper_clk:
return zynq_clk_get_cpu_rate(priv, cpu_1x_clk);
please remove this empty line.
default: return -ENXIO; }
I am also curious why you are enabling I2C in SPL. What are you doing with that?
Thanks, Michal
Hi,
Hi Michal,
first of all I would rephrase subject a little bit.
Something like: "ARM: zynq: Add missing i2c get_rate for fixing i2c SPL"
OK.
On 14. 02. 19 6:50, Hannes Schmelzer wrote:
The commit 'f48ef0d81aa837a33020f8d61abb3929ba613774' did break I2C support because requesting the clock for the I2C ip-block isn't supported during SPL.
To fixup this we add support requesting clocks for:
- i2c0
- i2c1
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
drivers/clk/clk_zynq.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/clk/clk_zynq.c b/drivers/clk/clk_zynq.c index 482f093..9a4e949 100644 --- a/drivers/clk/clk_zynq.c +++ b/drivers/clk/clk_zynq.c @@ -434,6 +434,9 @@ static ulong zynq_clk_get_rate(struct clk *clk) case lqspi_clk ... pcap_clk: case sdio0_clk ... spi1_clk: return zynq_clk_get_peripheral_rate(priv, id, 0);
- case i2c0_aper_clk ... i2c1_aper_clk:
return zynq_clk_get_cpu_rate(priv, cpu_1x_clk);please remove this empty line.
OK. I'll send V2 patch.
default: return -ENXIO; }I am also curious why you are enabling I2C in SPL. What are you doing with that?
I have to turn on very early some voltage rails on my board, where i have to communicate with i2c to a resetcontroller.
Thanks, Michal
cheers, Hannes
On 14. 02. 19 8:52, Hannes Schmelzer wrote:
Hi,
Hi Michal,
first of all I would rephrase subject a little bit.
Something like: "ARM: zynq: Add missing i2c get_rate for fixing i2c SPL"
OK.
On 14. 02. 19 6:50, Hannes Schmelzer wrote:
The commit 'f48ef0d81aa837a33020f8d61abb3929ba613774' did break I2C support because requesting the clock for the I2C ip-block isn't supported during SPL.
To fixup this we add support requesting clocks for:
- i2c0
- i2c1
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
drivers/clk/clk_zynq.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/clk/clk_zynq.c b/drivers/clk/clk_zynq.c index 482f093..9a4e949 100644 --- a/drivers/clk/clk_zynq.c +++ b/drivers/clk/clk_zynq.c @@ -434,6 +434,9 @@ static ulong zynq_clk_get_rate(struct clk *clk) case lqspi_clk ... pcap_clk: case sdio0_clk ... spi1_clk: return zynq_clk_get_peripheral_rate(priv, id, 0);
- case i2c0_aper_clk ... i2c1_aper_clk:
return zynq_clk_get_cpu_rate(priv, cpu_1x_clk);please remove this empty line.
OK. I'll send V2 patch.
default: return -ENXIO; }I am also curious why you are enabling I2C in SPL. What are you doing with that?
I have to turn on very early some voltage rails on my board, where i have to communicate with i2c to a resetcontroller.
ok. How does that code look like? Just dm_i2c_write?
I have never tried i2c in spl.
Thanks, Michal
On 2/14/19 1:26 PM, Michal Simek wrote:
On 14. 02. 19 8:52, Hannes Schmelzer wrote:
Hi,
Hi Michal,
first of all I would rephrase subject a little bit.
Something like: "ARM: zynq: Add missing i2c get_rate for fixing i2c SPL"
OK.
On 14. 02. 19 6:50, Hannes Schmelzer wrote:
The commit 'f48ef0d81aa837a33020f8d61abb3929ba613774' did break I2C support because requesting the clock for the I2C ip-block isn't supported during SPL.
To fixup this we add support requesting clocks for:
- i2c0
- i2c1
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
drivers/clk/clk_zynq.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/clk/clk_zynq.c b/drivers/clk/clk_zynq.c index 482f093..9a4e949 100644 --- a/drivers/clk/clk_zynq.c +++ b/drivers/clk/clk_zynq.c @@ -434,6 +434,9 @@ static ulong zynq_clk_get_rate(struct clk *clk) case lqspi_clk ... pcap_clk: case sdio0_clk ... spi1_clk: return zynq_clk_get_peripheral_rate(priv, id, 0);
- case i2c0_aper_clk ... i2c1_aper_clk:
return zynq_clk_get_cpu_rate(priv, cpu_1x_clk);please remove this empty line.
OK. I'll send V2 patch.
default: return -ENXIO; }I am also curious why you are enabling I2C in SPL. What are you doing with that?
I have to turn on very early some voltage rails on my board, where i have to communicate with i2c to a resetcontroller.
ok. How does that code look like? Just dm_i2c_write?
I have never tried i2c in spl.
hi michal,
here for example:
int board_init(void) /* spl stage */ { u16 u16buf; u8 u8buf; int rc; struct udevice *dev;
/* peripheral RESET on PSOC reset-controller */ rc = i2c_get_chip_for_busnum(0, RSTCTRL_ADDR_PSOC, 1, &dev); if (rc == 0) { u16buf = RSTCTRL_CTRLSPEC_nPCIRST; rc = dm_i2c_write(dev, RSTCTRL_CTRLSPEC, (uint8_t *)&u16buf, sizeof(u16buf)); } else { printf("failed to get i2c chip for PSOC resetcontroller!\n"); }
return 0; }
later on my plan is to push my board mainline once development has been finished.
Thanks, Michal
cheers, Hannes
participants (4)
-
Hannes Schmelzer -
Hannes Schmelzer
-
Hannes Schmelzer -
Michal Simek