Commit graph

182 commits

Author SHA1 Message Date
Patrick Delaunay
aad8414794 stm32mp: update the mmu configuration for SPL and prereloc
Overidde the weak function dram_bank_mmu_setup() to set the DDR
(preloc case) or the SYSRAM (in SPL case) executable before to enable
the MMU and configure DACR.

This weak function is called in dcache_enable/mmu_setup.

This patchs avoids a permission access issue when the DDR is marked
executable (by calling mmu_set_region_dcache_behaviour with
DCACHE_DEFAULT_OPTION) after MMU setup and domain access permission
activation with DACR in dcache_enable.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
2021-03-02 15:53:37 -05:00
Patrick Delaunay
1419e5b516 stm32mp: update MMU config before the relocation
Mark the top of ram, used for relocated U-Boot as a normal memory
(cacheable and executable) to avoid permission access issue when
U-Boot jumps to this relocated code.

When MMU is activated in pre-reloc stage; only the beginning of
DDR is marked executable.

This patch avoids access issue when DACR is correctly managed.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
2021-03-02 15:53:37 -05:00
Simon Glass
401d1c4f5d common: Drop asm/global_data.h from common header
Move this out of the common header and include it only where needed.  In
a number of cases this requires adding "struct udevice;" to avoid adding
another large header or in other cases replacing / adding missing header
files that had been pulled in, very indirectly.   Finally, we have a few
cases where we did not need to include <asm/global_data.h> at all, so
remove that include.

Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
2021-02-02 15:33:42 -05:00
Patrick Delaunay
751f918bdb arm: stm32mp: stm32prog: always flush DFU on start command for uart
Remove the test on data->dfu_seq, because dfu_seq=0 not only when
the DFU is not started (mask with 0xffff). This flush is mandatory
as the final treatment, common with USB, is done in DFU callback.

This patch avoids issue if the received length is a multiple of
the DFU packet.

For example if size of bootfs partition is egual to 0x4000000,
data->dfu_seq=0 at the end of the partition, the flush it not
requested and the phase is not increased in the callback.
U-Boot continue to request the bootfs in the next GetPhase command.

Fixes: 468f0508b5 ("stm32mp: stm32prog: add serial link support")
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
2021-01-13 09:52:58 +01:00
Patrick Delaunay
b66bfdf238 arm: stm32mp: bsec: migrate trace to log macro
Define LOG_CATEGORY, change pr_debug to dev_dbg and remove "bsec:"
header as it is managed by log macro (dev->name is displayed)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2021-01-13 09:52:58 +01:00
Patrick Delaunay
711b5bc0a2 arm: stm32mp: migrate cmd_stm32prog to log macro
Change debug and pr_ macro to log macro.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2021-01-13 09:52:58 +01:00
Patrick Delaunay
eb653acd63 arm: stm32mp: migrate trace to log macro
Change debug and pr_ macro to log macro and define LOG_CATEGORY.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2021-01-13 09:52:58 +01:00
Simon Glass
65e25bea59 dm: Rename DM_GET_DRIVER() to DM_DRIVER_GET()
In the spirit of using the same base name for all of these related macros,
rename this to have the operation at the end. This is not widely used so
the impact is fairly small.

Signed-off-by: Simon Glass <sjg@chromium.org>
2021-01-05 12:26:35 -07:00
Simon Glass
f10643cf8a dm: core: Access device ofnode through functions
At present ofnode is present in the device even if it is never used. With
of-platdata this field is not used, so can be removed. In preparation for
this, change the access to go through inline functions.

Signed-off-by: Simon Glass <sjg@chromium.org>
2021-01-05 12:24:41 -07:00
Simon Glass
0fd3d91152 dm: Use access methods for dev/uclass private data
Most drivers use these access methods but a few do not. Update them.

In some cases the access is not permitted, so mark those with a FIXME tag
for the maintainer to check.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Pratyush Yadav <p.yadav@ti.com>
2021-01-05 12:24:40 -07:00
Simon Glass
8a8d24bdf1 dm: treewide: Rename ..._platdata variables to just ..._plat
Try to maintain some consistency between these variables by using _plat as
a suffix for them.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 16:51:09 -07:00
Simon Glass
d1998a9fde dm: treewide: Rename ofdata_to_platdata() to of_to_plat()
This name is far too long. Rename it to remove the 'data' bits. This makes
it consistent with the platdata->plat rename.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 16:51:09 -07:00
Simon Glass
c69cda25c9 dm: treewide: Rename dev_get_platdata() to dev_get_plat()
Rename this to be consistent with the change from 'platdata'.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 16:51:09 -07:00
Simon Glass
caa4daa2ae dm: treewide: Rename 'platdata' variables to just 'plat'
We use 'priv' for private data but often use 'platdata' for platform data.
We can't really use 'pdata' since that is ambiguous (it could mean private
or platform data).

Rename some of the latter variables to end with 'plat' for consistency.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 16:51:08 -07:00
Simon Glass
41575d8e4c dm: treewide: Rename auto_alloc_size members to be shorter
This construct is quite long-winded. In earlier days it made some sense
since auto-allocation was a strange concept. But with driver model now
used pretty universally, we can shorten this to 'auto'. This reduces
verbosity and makes it easier to read.

Coincidentally it also ensures that every declaration is on one line,
thus making dtoc's job easier.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 08:00:25 -07:00
Richard Genoud
40426d6f9a SPL: stm32mp1: fix spl_mmc_boot_partition not defined
spl_mmc_boot_partition is only defined when
CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_USE_PARTITION is defined.

Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-11-25 10:56:09 +01:00
Patrick Delaunay
77c077e171 arm: stm32mp: correct the ALIGN macro usage
Correct the ALIGN macro usage in mmu_set_region_dcache_behaviour
call: the address must use ALIGN_DOWN and size can use ALIGN macro.

With STM32_SYSRAM_BASE=0x2FFC0000 and MMU_SECTION_SIZE=0x100000 for
STM32MP15x the computed address was 30000000 instead of 2ff00000.

Fixes: 43fe9d2fda ("stm32mp1: mmu_set_region_dcache_behaviour")
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-11-25 10:33:16 +01:00
Patrick Delaunay
53de79fecc stm32mp: stm32prog: accept device without partition
When partitions are not available on a device the command stm32prog raises
an error but a device can have no partition to check in init_device()
and the command need to continue to the next part_id.

This patch correct an issue for ram0 target, when block_dev and mtd
are NULL.

For example with the simple flashlayout file:

Opt	Part	Name	Type	Device	Offset	Binary
-	0x01	fsbl	Binary	none	0x0	tf-a-serialboot.stm32
-	0x03	ssbl	Binary	none	0x0	u-boot.stm32
P	0x10	kernel	System	ram0	0xC2000000	uImage.bin
P	0x11	dtb	FileSytem	ram0	0xC4000000	stm32mp157f-ev1.dtb

Fixes: ffc405e63b ("stm32mp: stm32prog: add upport of partial update")
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-10-21 18:12:20 +02:00
Patrick Delaunay
6af78d03d3 arm: stm32: cleanup arch gpio.h
Cosmetic update of gpio.h:
- remove enumerate: stm32_gpio_port, stm32_gpio_pin
  because STM32_GPIO_XXX values are unused
- move STM32_GPIOS_PER_BANK in stm32_gpio.c
  as its value is IP dependent and not arch dependent

No functional change as number of banks and number of gpio by banks
is managed by device tree since since DM migration and
commit 8f651ca60b ("pinctrl: stm32: Add get_pins_count() ops").

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-10-21 18:12:20 +02:00
Patrick Delaunay
67f9f11f19 stm32mp: limit size of cacheable DDR in pre-reloc stage
In pre-reloc stage, U-Boot marks cacheable the DDR limited by
the new config CONFIG_DDR_CACHEABLE_SIZE.

This patch allows to avoid any speculative access to DDR protected by
firewall and used by OP-TEE; the "no-map" reserved memory
node in DT are assumed after this limit:
STM32_DDR_BASE + DDR_CACHEABLE_SIZE.

Without security, in basic boot, the value is equal to STM32_DDR_SIZE.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-10-21 18:12:20 +02:00
Patrick Delaunay
cf8df34015 arm: stm32mp: cleanup test on eth_env_set_enetaddr result
Remove the unnecessary inversion on the eth_env_set_enetaddr() result which
only make complex the code of setup_mac_address() and display an invalid
value in the associated pr_err.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Marek Vasut <marex@denx.de>
2020-09-09 14:19:12 +02:00
Patrick Delaunay
9a2b0540dc arm: stm32mp: stm32prog: use IS_ENABLED to prevent ifdef
Use IS_ENABLED to prevent ifdef in stm32prog command.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-08-13 09:52:49 +02:00
Patrick Delaunay
4e9e358fe8 arm: stm32mp: bsec: use IS_ENABLED to prevent ifdef
Use IS_ENABLED to prevent ifdef in bsec driver.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-08-13 09:52:49 +02:00
Simon Glass
a00867b47a sf: Drop dm.h header file from spi_flash.h
This header file should not be included in other header files. Remove it
and use a forward declaration instead.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-08-03 22:19:54 -04:00
Tom Rini
719f42190d Use binman instead of one of the Rockchip build scripts
Refactor to allow any arch to create SPI-flash images
 New button uclass
 -----BEGIN PGP SIGNATURE-----
 
 iQFFBAABCgAvFiEEslwAIq+Gp8wWVbYnfxc6PpAIreYFAl8g33sRHHNqZ0BjaHJv
 bWl1bS5vcmcACgkQfxc6PpAIreaeEgf/QhmJjd6lEE3ABncRa3texyqq7j9eu+9P
 2JMCahrIzps83LLOkzNSGs1+m1hRuTcdnYNV8FxdmkAjDTYwV5RHOfwtpAr2nkH2
 HBbJRA2qMPleqT8wcymouhaCIRTTCGfUkFgZ+OowwSlHnQ4d2oGU/59Z+AkbpfX0
 3svA1Mlr5rUTZ3qTvdsepuZckuYOTGYZ5jgjbgr+G8F4iW4Ov+OkBxqDz9zFCveF
 md5CHKP1SV+bq67Q9uqUEwatLzYpy1p24j1Qe4G08U1+XCi6r+95VEc7TWxggboj
 eV/5d/sZbGESAtJiJ4TmqpGVcOLzBLI0FOunky0JNFwdcyRS/6KcCg==
 =Xjil
 -----END PGP SIGNATURE-----

Merge tag 'dm-pull-28jul20' of git://git.denx.de/u-boot-dm

Use binman instead of one of the Rockchip build scripts
Refactor to allow any arch to create SPI-flash images
New button uclass
2020-07-29 21:16:08 -04:00
Simon Glass
e6385c7e9c Makefile: Rename ALL-y to INPUTS-y
When binman is in use, most of the targets built by the Makefile are
inputs to binman. We then need a final rule to run binman to produce the
final outputs.

Rename the variable to indicate this, and add a new 'inputs' target.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
2020-07-28 19:30:39 -06:00
Patrick Delaunay
346034a7a8 arm: stm32mp: use correct weak function name spl_board_prepare_for_linux
Replace the function spl_board_prepare_for_boot_linux by the correct
name of the weak function spl_board_prepare_for_linux defined in spl.h.

This patch avoids warning with W=1 option:

u-boot/arch/arm/mach-stm32mp/spl.c:150:6:
warning: no previous prototype for ‘spl_board_prepare_for_boot_linux’
[-Wmissing-prototypes]

Fixes: dc7e5f190d ("arm: stm32mp: activate data cache on DDR in SPL")
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-28 18:14:15 +02:00
Patrick Delaunay
ded48bec05 arm: stm32mp: stm32prog: check result of find_mmc_device
Check result of find_mmc_device() before calling mmc_init().

This patch avoid a data abort when the command try to acces
to sd/mmc device deactivated in device tree.

Fixes: aff4c5dd82 ("stm32mp: stm32prog: add MMC device")
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-28 17:43:46 +02:00
Patrick Delaunay
43fe9d2fda stm32mp1: mmu_set_region_dcache_behaviour
Since commit d877f8fd0f ("arm: provide a function for boards init
code to modify MMU virtual-physical map") the parameter of
mmu_set_region_dcache_behaviour need to be MMU_SECTION_SIZE
aligned.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-28 17:21:37 +02:00
Patrick Delaunay
97f7e39def arm: stm32mp: move dbgmcu_init call when DT is ready
As the dbgmcu_init use the function bsec_dbgswenable which is based
on the DM and DT, its call can't be done before the spl is initialized
(driver model, DT and malloc) in board_init_f::spl_early_init().

This function call is moved later in spl_board_init().

Fixes: bd3f60d29c ("arm: stm32mp: protect DBGMCU_IDC access with BSEC")

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-28 17:21:37 +02:00
Masahiro Yamada
b75d8dc564 treewide: convert bd_t to struct bd_info by coccinelle
The Linux coding style guide (Documentation/process/coding-style.rst)
clearly says:

  It's a **mistake** to use typedef for structures and pointers.

Besides, using typedef for structures is annoying when you try to make
headers self-contained.

Let's say you have the following function declaration in a header:

  void foo(bd_t *bd);

This is not self-contained since bd_t is not defined.

To tell the compiler what 'bd_t' is, you need to include <asm/u-boot.h>

  #include <asm/u-boot.h>
  void foo(bd_t *bd);

Then, the include direcective pulls in more bloat needlessly.

If you use 'struct bd_info' instead, it is enough to put a forward
declaration as follows:

  struct bd_info;
  void foo(struct bd_info *bd);

Right, typedef'ing bd_t is a mistake.

I used coccinelle to generate this commit.

The semantic patch that makes this change is as follows:

  <smpl>
  @@
  typedef bd_t;
  @@
  -bd_t
  +struct bd_info
  </smpl>

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2020-07-17 09:30:13 -04:00
Bin Meng
7d5de35b6f arm: cmd_stm32prog: Fix the CONFIG_IS_ENABLED() usage
Add parentheses around CONFIG_IS_ENABLED() in the if statement, to
fix potential build failures.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-07-09 12:33:24 +08:00
Patrick Delaunay
c16cba88bd stm32mp1: use the command env info in env_check
Activate CMD_NVEDIT_INFO and use the new command "env info -d -p -q"
to automatically save the environment on first boot.

This patch allows to remove the env_default variable.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
bd3f60d29c arm: stm32mp: protect DBGMCU_IDC access with BSEC
As debugger must be totally closed on Sec closed chip,
the DBGMCU_IDC register is no more accessible (self
hosted debug is disabled with OTP).

This patch adds a function bsec_dbgswenable() to check
if the DBGMCU registers are available before to access them:
BSEC_DENABLE.DBGSWENABLE = self hosted debug status.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
03c4e6224a arm: stm32mp: stm32prog: add "Device Name" in iproduct during DFU USB enumeration
Add "Device Name" in iproduct during DFU USB enumeration
to have this information in STM32CubeProgrammer trace
(this tools is compatible with @Name since v2.3)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
d1a4b09de6 board: st: stpmic1: add function stpmic1_init
Add a function stmpic_init to early initialize the PMIC STPMIC1
- keep vdd on during the reset cycle (to avoid issue when backup battery
  is absent)
- Check if debug is enabled to program PMIC according to the bit

This patch allows to remove the compilation of spl.c file from stm32mp1
board in dh_stm32mp1.

CONFIG_SPL_BOARD_INIT is removed as the new function is called earlier
in SPL, in the function board_early_init_f.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
4e62642aef arm: stm32mp: add weak function to save vddcore
Add a weak functions to save the vddcore voltage value provided
in the OPP node when the clock tree is initialized.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
95bd49a5aa arm: stm32mp: spl: add bsec driver in SPL
Add the bsec driver in SPL, as it is needed by SOC part number detection
to found the supported OPP.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-07-07 16:01:23 +02:00
Etienne Carriere
164873a0c0 board: stm32mp1: fix handling of DT OP-TEE reserved memory nodes
Fix the sequence in stm32mp1 fdt.c that disables OP-TEE resources
defined in FDT when U-boot detects OP-TEE firmware is not present.

Before this change, helper function stm32_fdt_disable_optee()
set property status to "disabled" for the OP-TEE reserved memory
nodes but this has no impact since Linux kernel does not consider
the status property for reserved-memory subnodes. This change
make U-Boot to attempt to delete the node instead.

Fixes: 4a1b975dac ("board: stm32mp1: reserve memory for OP-TEE in device tree")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-06-16 10:38:09 +02:00
Simon Glass
cd93d625fd common: Drop linux/bitops.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:23 -04:00
Simon Glass
f09f1ecbe7 Use __ASSEMBLY__ as the assembly macros
Some places use __ASSEMBLER__ instead which does not work since the
Makefile does not define it. Fix them.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:23 -04:00
Simon Glass
c05ed00afb common: Drop linux/delay.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:23 -04:00
Simon Glass
f7ae49fc4f common: Drop log.h from common header
Move this header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:18 -04:00
Simon Glass
0914011310 command: Remove the cmd_tbl_t typedef
We should not use typedefs in U-Boot. They cannot be used as forward
declarations which means that header files must include the full header to
access them.

Drop the typedef and rename the struct to remove the _s suffix which is
now not useful.

This requires quite a few header-file additions.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 18:36:55 -04:00
Simon Glass
691d719db7 common: Drop init.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 17:33:33 -04:00
Simon Glass
4d72caa5b9 common: Drop image.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 17:33:33 -04:00
Simon Glass
0528979fa7 part: Drop disk_partition_t typedef
We should not be using typedefs and these make it harder to use
forward declarations (to reduce header file inclusions). Drop the typedef.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 17:33:33 -04:00
Simon Glass
90526e9fba common: Drop net.h from common header
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.

Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 17:33:31 -04:00
Patrick Delaunay
dc7e5f190d arm: stm32mp: activate data cache on DDR in SPL
Activate cache on DDR to improve the accesses to DDR used by SPL:
- CONFIG_SPL_BSS_START_ADDR
- CONFIG_SYS_SPL_MALLOC_START

Cache is configured only when DDR is fully initialized,
to avoid speculative access and issue in get_ram_size().
Data cache is deactivated at the end of SPL, to flush the data cache
and the TLB.

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-05-14 09:02:12 +02:00
Patrick Delaunay
7e8471cae5 arm: stm32mp: activate data cache in SPL and before relocation
Activate the data cache in SPL and in U-Boot before relocation.

In arch_cpu_init(), the function early_enable_caches() sets the early
TLB, early_tlb[] located .init section, and set cacheable:
- for SPL, all the SYSRAM
- for U-Boot, all the DDR

After relocation, the function enable_caches() (called by board_r)
reconfigures the MMU with new TLB location (reserved in
board_f.c::reserve_mmu) and re-enable the data cache.

This patch allows to reduce the execution time, particularly
- for the device tree parsing in U-Boot pre-reloc stage
  (dm_extended_scan_fd =>dm_scan_fdt)
- in I2C timing computation in SPL (stm32_i2c_choose_solution())

For example, the result on STM32MP157C-DK2 board is:
   1,6s gain for trusted boot chain with TF-A
   2,2s gain for basic boot chain with SPL

For information, as TLB is added in .data section, the binary size
increased and the SPL load time by ROM code increased (30ms on DK2).

But early malloc can't be used for TLB because arch_cpu_init()
is executed before the early poll initialization done in spl_common_init()
called by spl_early_init() So it too late for this use case.
And if I initialize the MMU and the cache after this function it is
too late, as dm_init_and_scan and fdt parsing is also called in
spl_common_init().

And .BSS can be used in board_init_f(): only stack and global can use
before BSS init done in board_init_r().

So .data is the better solution without hardcoded location but if you
have size issue for SPL you can deactivate cache for SPL only
(with CONFIG_SPL_SYS_DCACHE_OFF).

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-05-14 09:02:12 +02:00