[U-Boot] [PATCH] arm: socfpga: make SPL_TEXT_BASE overridable
To boot from fpga on socfpga gen5, we need to set CONFIG_SPL_TEXT_BASE to 0xC0000000 (hps2fpgaslaves base address).
Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been successful so far, let's make this value overridable in socfpga_common.h, so that we can have different board configs override this in socfpga_common.h.
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com ---
include/configs/socfpga_common.h | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/include/configs/socfpga_common.h b/include/configs/socfpga_common.h index 181af9b646..191204b27b 100644 --- a/include/configs/socfpga_common.h +++ b/include/configs/socfpga_common.h @@ -248,8 +248,10 @@ unsigned int cm_get_qspi_controller_clk_hz(void); * 0xFFEz_zzzz ...... Malloc area (grows up to top) * 0xFFE3_FFFF ...... End of SRAM (top) */ +#ifndef CONFIG_SPL_TEXT_BASE #define CONFIG_SPL_TEXT_BASE CONFIG_SYS_INIT_RAM_ADDR #define CONFIG_SPL_MAX_SIZE CONFIG_SYS_INIT_RAM_SIZE +#endif
#if defined(CONFIG_TARGET_SOCFPGA_ARRIA10) /* SPL memory allocation configuration, this is for FAT implementation */
On 3/15/19 8:44 PM, Simon Goldschmidt wrote:
To boot from fpga on socfpga gen5, we need to set CONFIG_SPL_TEXT_BASE to 0xC0000000 (hps2fpgaslaves base address).
Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been successful so far, let's make this value overridable in socfpga_common.h, so that we can have different board configs override this in socfpga_common.h.
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
Is this a fix for current release or new feature for next ?
include/configs/socfpga_common.h | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/include/configs/socfpga_common.h b/include/configs/socfpga_common.h index 181af9b646..191204b27b 100644 --- a/include/configs/socfpga_common.h +++ b/include/configs/socfpga_common.h @@ -248,8 +248,10 @@ unsigned int cm_get_qspi_controller_clk_hz(void);
- 0xFFEz_zzzz ...... Malloc area (grows up to top)
- 0xFFE3_FFFF ...... End of SRAM (top)
*/ +#ifndef CONFIG_SPL_TEXT_BASE #define CONFIG_SPL_TEXT_BASE CONFIG_SYS_INIT_RAM_ADDR #define CONFIG_SPL_MAX_SIZE CONFIG_SYS_INIT_RAM_SIZE +#endif
#if defined(CONFIG_TARGET_SOCFPGA_ARRIA10) /* SPL memory allocation configuration, this is for FAT implementation */
Am 15.03.2019 um 22:20 schrieb Marek Vasut:
On 3/15/19 8:44 PM, Simon Goldschmidt wrote:
To boot from fpga on socfpga gen5, we need to set CONFIG_SPL_TEXT_BASE to 0xC0000000 (hps2fpgaslaves base address).
Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been successful so far, let's make this value overridable in socfpga_common.h, so that we can have different board configs override this in socfpga_common.h.
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
Is this a fix for current release or new feature for next ?
Yes, it's a fix for the current release. It's the only thing missing to *finally* get mainline SPL booting from FPGA (on socfpga gen5).
As written in the log, the original plan was to move CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out (still), which is why we have to do it in a different way.
I could provide a new board config to boot from FPGA if you want, but that did not make much sense to me as the FPGA image would be missing...
Regards, Simon
include/configs/socfpga_common.h | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/include/configs/socfpga_common.h b/include/configs/socfpga_common.h index 181af9b646..191204b27b 100644 --- a/include/configs/socfpga_common.h +++ b/include/configs/socfpga_common.h @@ -248,8 +248,10 @@ unsigned int cm_get_qspi_controller_clk_hz(void);
- 0xFFEz_zzzz ...... Malloc area (grows up to top)
- 0xFFE3_FFFF ...... End of SRAM (top)
*/ +#ifndef CONFIG_SPL_TEXT_BASE #define CONFIG_SPL_TEXT_BASE CONFIG_SYS_INIT_RAM_ADDR #define CONFIG_SPL_MAX_SIZE CONFIG_SYS_INIT_RAM_SIZE +#endif
#if defined(CONFIG_TARGET_SOCFPGA_ARRIA10) /* SPL memory allocation configuration, this is for FAT implementation */
On 3/16/19 9:23 AM, Simon Goldschmidt wrote:
Am 15.03.2019 um 22:20 schrieb Marek Vasut:
On 3/15/19 8:44 PM, Simon Goldschmidt wrote:
To boot from fpga on socfpga gen5, we need to set CONFIG_SPL_TEXT_BASE to 0xC0000000 (hps2fpgaslaves base address).
Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been successful so far, let's make this value overridable in socfpga_common.h, so that we can have different board configs override this in socfpga_common.h.
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
Is this a fix for current release or new feature for next ?
Yes, it's a fix for the current release. It's the only thing missing to *finally* get mainline SPL booting from FPGA (on socfpga gen5).
As written in the log, the original plan was to move CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out (still), which is why we have to do it in a different way.
I could provide a new board config to boot from FPGA if you want, but that did not make much sense to me as the FPGA image would be missing...
Well, applied.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?
On Saturday, March 16, 2019, Marek Vasut marex@denx.de wrote:
On 3/16/19 9:23 AM, Simon Goldschmidt wrote:
Am 15.03.2019 um 22:20 schrieb Marek Vasut:
On 3/15/19 8:44 PM, Simon Goldschmidt wrote:
To boot from fpga on socfpga gen5, we need to set CONFIG_SPL_TEXT_BASE to 0xC0000000 (hps2fpgaslaves base address).
Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been successful so far, let's make this value overridable in socfpga_common.h, so that we can have different board configs override this in socfpga_common.h.
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
Is this a fix for current release or new feature for next ?
Yes, it's a fix for the current release. It's the only thing missing to *finally* get mainline SPL booting from FPGA (on socfpga gen5).
As written in the log, the original plan was to move CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out (still), which is why we have to do it in a different way.
I could provide a new board config to boot from FPGA if you want, but that did not make much sense to me as the FPGA image would be missing...
Well, applied.
Thanks.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?
I started that conversion (still in patchwork) but Tom said it produced errors and he kind of took it further. I'm not really on track what's still missing.
But of course I would still prefer the Kconfig version instead of this #ifndef...
Regards, Simon
On 3/16/19 1:36 PM, Simon Goldschmidt wrote:
On Saturday, March 16, 2019, Marek Vasut <marex@denx.de mailto:marex@denx.de> wrote:
On 3/16/19 9:23 AM, Simon Goldschmidt wrote: > Am 15.03.2019 um 22:20 schrieb Marek Vasut: >> On 3/15/19 8:44 PM, Simon Goldschmidt wrote: >>> To boot from fpga on socfpga gen5, we need to set >>> CONFIG_SPL_TEXT_BASE to >>> 0xC0000000 (hps2fpgaslaves base address). >>> >>> Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been >>> successful so >>> far, let's make this value overridable in socfpga_common.h, so that >>> we can >>> have different board configs override this in socfpga_common.h. >>> >>> Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com <mailto:simon.k.r.goldschmidt@gmail.com>> >> >> Is this a fix for current release or new feature for next ? > > Yes, it's a fix for the current release. It's the only thing missing to > *finally* get mainline SPL booting from FPGA (on socfpga gen5). > > As written in the log, the original plan was to move > CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out > (still), which is why we have to do it in a different way. > > I could provide a new board config to boot from FPGA if you want, but > that did not make much sense to me as the FPGA image would be missing... Well, applied.Thanks.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?I started that conversion (still in patchwork) but Tom said it produced errors and he kind of took it further. I'm not really on track what's still missing.
But of course I would still prefer the Kconfig version instead of this #ifndef...
Always :)
On 16.03.19 13:50, Marek Vasut wrote:
On 3/16/19 1:36 PM, Simon Goldschmidt wrote:
On Saturday, March 16, 2019, Marek Vasut <marex@denx.de mailto:marex@denx.de> wrote:
On 3/16/19 9:23 AM, Simon Goldschmidt wrote: > Am 15.03.2019 um 22:20 schrieb Marek Vasut: >> On 3/15/19 8:44 PM, Simon Goldschmidt wrote: >>> To boot from fpga on socfpga gen5, we need to set >>> CONFIG_SPL_TEXT_BASE to >>> 0xC0000000 (hps2fpgaslaves base address). >>> >>> Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been >>> successful so >>> far, let's make this value overridable in socfpga_common.h, so that >>> we can >>> have different board configs override this in socfpga_common.h. >>> >>> Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com <mailto:simon.k.r.goldschmidt@gmail.com>> >> >> Is this a fix for current release or new feature for next ? > > Yes, it's a fix for the current release. It's the only thing missing to > *finally* get mainline SPL booting from FPGA (on socfpga gen5). > > As written in the log, the original plan was to move > CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out > (still), which is why we have to do it in a different way. > > I could provide a new board config to boot from FPGA if you want, but > that did not make much sense to me as the FPGA image would be missing... Well, applied.Thanks.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?I started that conversion (still in patchwork) but Tom said it produced errors and he kind of took it further. I'm not really on track what's still missing.
But of course I would still prefer the Kconfig version instead of this #ifndef...
Always :)
What's the status here, Tom? Anything I can do? Should I try again now with my original approach of just using moveconfig.py?
Regards, Simon
On Sat, Mar 16, 2019 at 10:16:24PM +0100, Simon Goldschmidt wrote:
On 16.03.19 13:50, Marek Vasut wrote:
On 3/16/19 1:36 PM, Simon Goldschmidt wrote:
On Saturday, March 16, 2019, Marek Vasut <marex@denx.de mailto:marex@denx.de> wrote:
On 3/16/19 9:23 AM, Simon Goldschmidt wrote: > Am 15.03.2019 um 22:20 schrieb Marek Vasut: >> On 3/15/19 8:44 PM, Simon Goldschmidt wrote: >>> To boot from fpga on socfpga gen5, we need to set >>> CONFIG_SPL_TEXT_BASE to >>> 0xC0000000 (hps2fpgaslaves base address). >>> >>> Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been >>> successful so >>> far, let's make this value overridable in socfpga_common.h, so that >>> we can >>> have different board configs override this in socfpga_common.h. >>> >>> Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com <mailto:simon.k.r.goldschmidt@gmail.com>> >> >> Is this a fix for current release or new feature for next ? > > Yes, it's a fix for the current release. It's the only thing missing to > *finally* get mainline SPL booting from FPGA (on socfpga gen5). > > As written in the log, the original plan was to move > CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out > (still), which is why we have to do it in a different way. > > I could provide a new board config to boot from FPGA if you want, but > that did not make much sense to me as the FPGA image would be missing... Well, applied.Thanks.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?I started that conversion (still in patchwork) but Tom said it produced errors and he kind of took it further. I'm not really on track what's still missing.
But of course I would still prefer the Kconfig version instead of this #ifndef...
Always :)
What's the status here, Tom? Anything I can do? Should I try again now with my original approach of just using moveconfig.py?
My last try and migrating in turn ran into a reminder that I need to first redo the keystone stuff to not use CONFIG_SPL_TEXT_BASE the way it does in other cases and just use the value directly. Let me see if I can get that done..
On Sun, Mar 17, 2019 at 08:54:08AM -0400, Tom Rini wrote:
On Sat, Mar 16, 2019 at 10:16:24PM +0100, Simon Goldschmidt wrote:
On 16.03.19 13:50, Marek Vasut wrote:
On 3/16/19 1:36 PM, Simon Goldschmidt wrote:
On Saturday, March 16, 2019, Marek Vasut <marex@denx.de mailto:marex@denx.de> wrote:
On 3/16/19 9:23 AM, Simon Goldschmidt wrote: > Am 15.03.2019 um 22:20 schrieb Marek Vasut: >> On 3/15/19 8:44 PM, Simon Goldschmidt wrote: >>> To boot from fpga on socfpga gen5, we need to set >>> CONFIG_SPL_TEXT_BASE to >>> 0xC0000000 (hps2fpgaslaves base address). >>> >>> Since converting CONFIG_SPL_TEXT_BASE to Kconfig hasn't been >>> successful so >>> far, let's make this value overridable in socfpga_common.h, so that >>> we can >>> have different board configs override this in socfpga_common.h. >>> >>> Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com <mailto:simon.k.r.goldschmidt@gmail.com>> >> >> Is this a fix for current release or new feature for next ? > > Yes, it's a fix for the current release. It's the only thing missing to > *finally* get mainline SPL booting from FPGA (on socfpga gen5). > > As written in the log, the original plan was to move > CONFIG_SPL_TEXT_BASE to Kconfig completely, but that didn't work out > (still), which is why we have to do it in a different way. > > I could provide a new board config to boot from FPGA if you want, but > that did not make much sense to me as the FPGA image would be missing... Well, applied.Thanks.
That said, do you plan to work on the Kconfig conversion further ? What's the problem there ?I started that conversion (still in patchwork) but Tom said it produced errors and he kind of took it further. I'm not really on track what's still missing.
But of course I would still prefer the Kconfig version instead of this #ifndef...
Always :)
What's the status here, Tom? Anything I can do? Should I try again now with my original approach of just using moveconfig.py?
My last try and migrating in turn ran into a reminder that I need to first redo the keystone stuff to not use CONFIG_SPL_TEXT_BASE the way it does in other cases and just use the value directly. Let me see if I can get that done..
It may have finally dawned on me that the OMAP2PLUS symbol CONFIG_ISW_ENTRY_ADDR should also be exposed to Keystone and used there to store the "SPL_TEXT_BASE" type address in the case where we don't actually enable SPL on these platforms. Testing that now, and then I can test migrating CONFIG_SPL_TEXT_BASE on top of that.
participants (3)
-
Marek Vasut -
Simon Goldschmidt -
Tom Rini