arm: socfpga: fix SPL booting from fpga OnChip RAM

PATCH New Simon Goldschmidt 3 participants 8 messages Oct 6, 2018 – Oct 8, 2018

Subsystems (3)

Arm (32-bit) Arm sub-architectures Arm/socfpga architecture

lore

Share & tools

lore

Simon Goldschmidt

author

2018-10-06 20:33 UTC #0 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

This patch prevents disabling the FPGA bridges when SPL or U-Boot is executed from FPGA onchip RAM.

Signed-off-by Simon Goldschmidt <[email protected]>

---

There are other patches required to make boot from FPGA work correctly:
- CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000 ideally this is done by moving SPL_TEXT_BASE to Kconfig:
https://patchwork.ozlabs.org/patch/976918/
- spl_ram.c needs this patch by Michal Simek:
https://patchwork.ozlabs.org/patch/978694/
- For U-Boot to be run from FPGA onchip RAM, either CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch is required to relocate gd->env_addr:
https://patchwork.ozlabs.org/patch/975702/

---
 arch/arm/mach-socfpga/include/mach/base_addr_ac5.h | 1 +
 arch/arm/mach-socfpga/misc_gen5.c                  | 8 ++++++++
 arch/arm/mach-socfpga/spl_gen5.c                   | 4 ++++
 3 files changed, 13 insertions(+)

arch/arm/mach-socfpga/include/mach/base_addr_ac5.h +1 -0

index bb9e3faa29..2725e9fcc3 100644
@@ -6,6 +6,7 @@
 #ifndef _SOCFPGA_BASE_ADDRS_H_
 #define _SOCFPGA_BASE_ADDRS_H_
 
+#define SOCFPGA_FPGA_SLAVES_ADDRESS	0xc0000000
 #define SOCFPGA_STM_ADDRESS		0xfc000000
 #define SOCFPGA_DAP_ADDRESS		0xff000000
 #define SOCFPGA_EMAC0_ADDRESS		0xff700000

arch/arm/mach-socfpga/misc_gen5.c +8 -0

index 429c3d6cd5..fa66484783 100644
@@ -187,7 +187,15 @@ void socfpga_sdram_remap_zero(void)
 	setbits_le32(&scu_regs->sacr, 0xfff);
 
 	/* Configure the L2 controller to make SDRAM start at 0 */
+#if (defined(CONFIG_SPL_BUILD) && \
+	(CONFIG_SPL_TEXT_BASE == SOCFPGA_FPGA_SLAVES_ADDRESS)) || \
+	((CONFIG_SYS_TEXT_BASE >= SOCFPGA_FPGA_SLAVES_ADDRESS) && \
+	(CONFIG_SYS_TEXT_BASE < SOCFPGA_STM_ADDRESS))
+	/* remap.mpuzero, keep fpga bridge enabled */
+	writel(0x9, &nic301_regs->remap);
+#else
 	writel(0x1, &nic301_regs->remap);	/* remap.mpuzero */
+#endif
 	writel(0x1, &pl310->pl310_addr_filter_start);
 }

arch/arm/mach-socfpga/spl_gen5.c +4 -0

index be318cc0d9..41e7787f11 100644
@@ -92,8 +92,10 @@ void board_init_f(ulong dummy)
 
 	/* Put everything into reset but L4WD0. */
 	socfpga_per_reset_all();
+#if CONFIG_SPL_TEXT_BASE != SOCFPGA_FPGA_SLAVES_ADDRESS
 	/* Put FPGA bridges into reset too. */
 	socfpga_bridges_reset(1);
+#endif
 
 	socfpga_per_reset(SOCFPGA_RESET(SDR), 0);
 	socfpga_per_reset(SOCFPGA_RESET(UART0), 0);

@@ -163,5 +165,7 @@ void board_init_f(ulong dummy)
 		hang();
 	}
 
+#if CONFIG_SPL_TEXT_BASE != SOCFPGA_FPGA_SLAVES_ADDRESS
 	socfpga_bridges_reset(1);
+#endif
 }

Signature · 1 line

2.17.1

Tom Rini

2018-10-06 20:38 UTC #1 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 7 lines · Simon Goldschmidt

This patch prevents disabling the FPGA bridges when SPL or U-Boot is executed from FPGA onchip RAM.
Signed-off-by Simon Goldschmidt <[email protected]>
---

Quoted 11 lines

There are other patches required to make boot from FPGA work correctly:
- CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000 ideally this is done by moving SPL_TEXT_BASE to Kconfig:
https://patchwork.ozlabs.org/patch/976918/
- spl_ram.c needs this patch by Michal Simek:
https://patchwork.ozlabs.org/patch/978694/
- For U-Boot to be run from FPGA onchip RAM, either CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch is required to relocate gd->env_addr:
https://patchwork.ozlabs.org/patch/975702/

For the record, I'm currently reviewing the last two parts here. The first one will come soon (I do Kconfig migrations separate so I can see a lack of size change).

Signature · 8 lines

Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20181006/db5b6331/attachment.sig>

Simon Goldschmidt

author

2018-10-07 11:08 UTC #2 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 20 lines · Tom Rini

Quoted 11 lines · Simon Goldschmidt
This patch prevents disabling the FPGA bridges when SPL or U-Boot is executed from FPGA onchip RAM.
Signed-off-by Simon Goldschmidt <[email protected]>
---
There are other patches required to make boot from FPGA work correctly:
- CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000 ideally this is done by moving SPL_TEXT_BASE to Kconfig:
Quoted 7 lines
   https://patchwork.ozlabs.org/patch/976918/
- spl_ram.c needs this patch by Michal Simek:
   https://patchwork.ozlabs.org/patch/978694/
- For U-Boot to be run from FPGA onchip RAM, either
   CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch
   is required to relocate gd->env_addr:
   https://patchwork.ozlabs.org/patch/975702/

Quoted 3 lines

For the record, I'm currently reviewing the last two parts here. The first one will come soon (I do Kconfig migrations separate so I can see a lack of size change).

OK, after Tom merged the last to parts, I can confirm I can successfully boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still missing).

Thanks,

Simon

Marek Vasut

2018-10-07 12:36 UTC #3 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 26 lines · Simon Goldschmidt

Quoted 20 lines · Tom Rini
Quoted 7 lines · Simon Goldschmidt
This patch prevents disabling the FPGA bridges when SPL or U-Boot is executed from FPGA onchip RAM.
Signed-off-by Simon Goldschmidt <[email protected]>
---
Quoted 11 lines
There are other patches required to make boot from FPGA work correctly:
- CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000 ideally this is done by moving SPL_TEXT_BASE to Kconfig:
   https://patchwork.ozlabs.org/patch/976918/
- spl_ram.c needs this patch by Michal Simek:
   https://patchwork.ozlabs.org/patch/978694/
- For U-Boot to be run from FPGA onchip RAM, either CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch is required to relocate gd->env_addr:
   https://patchwork.ozlabs.org/patch/975702/
Quoted 3 lines
For the record, I'm currently reviewing the last two parts here. The first one will come soon (I do Kconfig migrations separate so I can see a lack of size change).

Quoted 4 lines

OK, after Tom merged the last to parts, I can confirm I can successfully boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still missing).

So is this patch still needed ?

Signature · 2 lines

Best regards,
Marek Vasut

Simon Goldschmidt

author

2018-10-08 04:04 UTC #4 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Marek Vasut <[email protected]> schrieb am Mo., 8. Okt. 2018, 02:20:

Quoted 27 lines · Simon Goldschmidt

Quoted 26 lines · Tom Rini
Quoted 23 lines · Simon Goldschmidt
Quoted 12 lines
This patch prevents disabling the FPGA bridges when SPL or U-Boot is executed from FPGA onchip RAM.
Signed-off-by Simon Goldschmidt <[email protected]>
---
There are other patches required to make boot from FPGA work correctly:
- CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000 ideally this is done by moving SPL_TEXT_BASE to Kconfig:
Quoted 7 lines
   https://patchwork.ozlabs.org/patch/976918/
- spl_ram.c needs this patch by Michal Simek:
   https://patchwork.ozlabs.org/patch/978694/
- For U-Boot to be run from FPGA onchip RAM, either
   CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch
   is required to relocate gd->env_addr:
   https://patchwork.ozlabs.org/patch/975702/
For the record, I'm currently reviewing the last two parts here. The first one will come soon (I do Kconfig migrations separate so I can see a lack of size change).

Quoted 7 lines

Quoted 4 lines
OK, after Tom merged the last to parts, I can confirm I can successfully boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still missing).
So is this patch still needed ?

Yes, sorry for being unclear there.

Simon

Marek Vasut

2018-10-08 09:03 UTC #5 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 44 lines · Simon Goldschmidt

Marek Vasut <marex at denx.de <mailto:[email protected]>> schrieb am Mo., 8. Okt. 2018, 02:20:

    On 10/07/2018 01:08 PM, Simon Goldschmidt wrote:
    > On 06.10.2018 22:38, Tom Rini wrote:
    >> On Sat, Oct 06, 2018 at 10:33:42PM +0200, Simon Goldschmidt wrote:
    >>
    >>> This patch prevents disabling the FPGA bridges when
    >>> SPL or U-Boot is executed from FPGA onchip RAM.
    >>>
    >>> Signed-off-by: Simon Goldschmidt
    <[email protected]
    <mailto:[email protected]>>
    >>> ---
    >>>
    >>>
    >>> There are other patches required to make boot from
    >>> FPGA work correctly:
    >>> - CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000
    >>>    ideally this is done by moving SPL_TEXT_BASE to Kconfig:
    >>>    https://patchwork.ozlabs.org/patch/976918/
    >>> - spl_ram.c needs this patch by Michal Simek:
    >>>    https://patchwork.ozlabs.org/patch/978694/
    >>> - For U-Boot to be run from FPGA onchip RAM, either
    >>>    CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch
    >>>    is required to relocate gd->env_addr:
    >>>    https://patchwork.ozlabs.org/patch/975702/
    >> For the record, I'm currently reviewing the last two parts here.  The
    >> first one will come soon (I do Kconfig migrations separate so I
    can see
    >> a lack of size change).
    >
    >
    > OK, after Tom merged the last to parts, I can confirm I can
    successfully
    > boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking
    > CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still
    > missing).

    So is this patch still needed ?

Quoted 1 line

Yes, sorry for being unclear there.

Can we do what this patch does with less ifdeffery ? Maybe with simple if (address > FOO && < bar) remap() ?

Signature · 2 lines

Best regards,
Marek Vasut

Simon Goldschmidt

author

2018-10-08 13:17 UTC #6 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 43 lines · Marek Vasut

Quoted 41 lines · Simon Goldschmidt

Marek Vasut <marex at denx.de <mailto:[email protected]>> schrieb am Mo., 8. Okt. 2018, 02:20:

    On 10/07/2018 01:08 PM, Simon Goldschmidt wrote:
    > On 06.10.2018 22:38, Tom Rini wrote:
    >> On Sat, Oct 06, 2018 at 10:33:42PM +0200, Simon Goldschmidt wrote:
    >>
    >>> This patch prevents disabling the FPGA bridges when
    >>> SPL or U-Boot is executed from FPGA onchip RAM.
    >>>
    >>> Signed-off-by: Simon Goldschmidt
    <[email protected]
    <mailto:[email protected]>>
    >>> ---
    >>>
    >>>
    >>> There are other patches required to make boot from
    >>> FPGA work correctly:
    >>> - CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000
    >>>    ideally this is done by moving SPL_TEXT_BASE to Kconfig:
    >>>    https://patchwork.ozlabs.org/patch/976918/
    >>> - spl_ram.c needs this patch by Michal Simek:
    >>>    https://patchwork.ozlabs.org/patch/978694/
    >>> - For U-Boot to be run from FPGA onchip RAM, either
    >>>    CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch
    >>>    is required to relocate gd->env_addr:
    >>>    https://patchwork.ozlabs.org/patch/975702/
    >> For the record, I'm currently reviewing the last two parts here.  The
    >> first one will come soon (I do Kconfig migrations separate so I
    can see
    >> a lack of size change).
    >
    >
    > OK, after Tom merged the last to parts, I can confirm I can
    successfully
    > boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking
    > CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still
    > missing).

Quoted 7 lines

Quoted 4 lines
So is this patch still needed ?
Yes, sorry for being unclear there.
Can we do what this patch does with less ifdeffery ? Maybe with simple if (address > FOO && < bar) remap() ?

Hmm, I can try, but the change in misc_gen5.c depends on whether we build SPL or U-Boot, I'm not sure how to do that without ifdef...

Does 'if (CONFIG_IS_ENABLED(SPL_BUILD))' work? Does it look nice? It's not used anywhere else like that.

Simon

Marek Vasut

2018-10-08 13:22 UTC #7 lore raw

PATCH 1/1 arm: socfpga: fix SPL booting from fpga OnChip RAM

Quoted 49 lines · Simon Goldschmidt

Quoted 48 lines · Marek Vasut

Quoted 45 lines · Simon Goldschmidt

Marek Vasut <marex at denx.de <mailto:[email protected]>> schrieb am Mo., 8. Okt. 2018, 02:20:

    On 10/07/2018 01:08 PM, Simon Goldschmidt wrote:
    > On 06.10.2018 22:38, Tom Rini wrote:
    >> On Sat, Oct 06, 2018 at 10:33:42PM +0200, Simon Goldschmidt wrote:
    >>
    >>> This patch prevents disabling the FPGA bridges when
    >>> SPL or U-Boot is executed from FPGA onchip RAM.
    >>>
    >>> Signed-off-by: Simon Goldschmidt
    <[email protected]
    <mailto:[email protected]>>
    >>> ---
    >>>
    >>>
    >>> There are other patches required to make boot from
    >>> FPGA work correctly:
    >>> - CONFIG_SPL_TEXT_BASE needs to be set to 0xC0000000
    >>>    ideally this is done by moving SPL_TEXT_BASE to Kconfig:
    >>>    https://patchwork.ozlabs.org/patch/976918/
    >>> - spl_ram.c needs this patch by Michal Simek:
    >>>    https://patchwork.ozlabs.org/patch/978694/
    >>> - For U-Boot to be run from FPGA onchip RAM, either
    >>>    CONFIG_SYS_EXTRA_ENV_RELOC needs to be used or this patch
    >>>    is required to relocate gd->env_addr:
    >>>    https://patchwork.ozlabs.org/patch/975702/
    >> For the record, I'm currently reviewing the last two parts here.  The
    >> first one will come soon (I do Kconfig migrations separate so I
    can see
    >> a lack of size change).
    >
    >
    > OK, after Tom merged the last to parts, I can confirm I can
    successfully
    > boot from FPGA by adapting CONFIG_SYS_TEXT_BASE via config and hacking
    > CONFIG_SPL_TEXT_BASE to 0xC0000000 (as moving this to Kconfig is still
    > missing).

    So is this patch still needed ?

Yes, sorry for being unclear there.

Quoted 8 lines

Quoted 2 lines
Can we do what this patch does with less ifdeffery ? Maybe with simple if (address > FOO && < bar) remap() ?
Hmm, I can try, but the change in misc_gen5.c depends on whether we build SPL or U-Boot, I'm not sure how to do that without ifdef...
Does 'if (CONFIG_IS_ENABLED(SPL_BUILD))' work? Does it look nice? It's not used anywhere else like that.

That's fine, I had those address range checks in mind.

Signature · 2 lines

Best regards,
Marek Vasut