Fix LTO build for some thumb-interwork usecases (such as for
da850evm_defconfig), where inline assmebly such as
mrc p15,0,r2,c1,c0,0
causes the compiler to fail during LTO linking with
Error: selected processor does not support `mrc p15,0,r2,c1,c0,0'
in Thumb mode
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
On ARM, the gd pointer is stored in registers r9 / x18. For this the
-ffixed-r9 / -ffixed-x18 flag is used when compiling, but using global
register variables causes errors when building with LTO, and these
errors are very difficult to overcome.
Richard Biener says [1]:
Note that global register vars shouldn't be used with LTO and if they
are restricted to just a few compilation units the recommended fix is
to build those CUs without -flto.
We cannot do this for U-Boot since all CUs use -ffixed-reg flag.
It seems that with LTO we could in fact store the gd pointer differently
and gain performance or size benefit by allowing the compiler to use
r9 / x18. But this would need more work.
So for now, when building with LTO, go the clang way, and instead of
declaring gd a global register variable, we make it a function call via
macro.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68384
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
The Thumb instruction `ldr` is able to move high registers only from
armv7. For armv5 and armv6 we have to use `mov`.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
In style of linked lists, instead of declaring symbols for boundaries
of getopt options array in the linker script, declare corresponding
sections and retrieve the boundaries via static inline functions.
Without this clang's LTO produces binary without any getopt options,
because for some reason it thinks that array is empty (start and end
symbols are at the same address).
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
Currently we use incremental linking (ld -r) to link several object
files from one directory into one built-in.o object file containing the
linked code from that directory (and its subdirectories).
Linux has, some time ago, moved to thin archives instead.
Thin archives are archives (.a) that do not really contain the object
files, only references to them.
Using thin archives instead of incremental linking
- saves disk space
- apparently works better with dead code elimination
- makes things easier for LTO
The third point is the important one for us. With incremental linking
there are several options how to do LTO, and that would unnecessarily
complicate things.
We have to use the --whole-archive/--no-whole-archive linking option
instead of --start-group/--end-group, otherwise linking may fail because
of unresolved symbols, or the resulting binary will be unusable.
We also need to use the P flag for ar, otherwise final linking may fail.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Simon Glass <sjg@chromium.org>
This commit does the same thing as Linux commit 33def8498fdd.
Use a more generic form for __section that requires quotes to avoid
complications with clang and gcc differences.
Remove the quote operator # from compiler_attributes.h __section macro.
Convert all unquoted __section(foo) uses to quoted __section("foo").
Also convert __attribute__((section("foo"))) uses to __section("foo")
even if the __attribute__ has multiple list entry forms.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
With commit 8678776df6 (arm: mvebu: armada-3720-uDPU: fix PHY mode
definition to sgmii-2500) the PHY mode was switch to "sgmii-2500", even
when this is functionally incorrect since "2500base-x" was not supported
in U-Boot at that time. As this mode is now supported (at least present
in the headers), this patch moves back to the orinal version.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Jakov Petrina <jakov.petrina@sartura.hr>
Cc: Vladimir Vid <vladimir.vid@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Now that we have switched to binman to generate u-boot.itb for all
RISC-V boards, USE_SPL_FIT_GENERATOR is no longer needed and can
be dropped.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Use the new BINMAN_STANDALONE_FDT option for AE350 based SPL defconfigs,
so that binman is now used to generate u-boot.itb.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Rick Chen <rick@andestech.com>
By utilizing the newly introduced BINMAN_STANDALONE_FDT option, along
with a new dedicated device tree source file for the QEMU virt target
used for binman only, we can now use binman to generate u-boot.itb.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Sort the RISC-V DTS build targets by their Kconfig target names in
alphabetical order.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present SiFive Unleashed board uses the Makefile to create the FIT,
using USE_SPL_FIT_GENERATOR, which is deprecated as per the Makefile
warning. Update to use binman instead.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The patch contains several DT files for SOM platform.
Carrier card is sck-kv (KV260) revA/B. SMK-K26 is description for starter
kit which doesn't have EMMC populated. And SM-K26 is full som with EMMC.
Files are divided in this way to make sure that SOM can be plugged to
different carrier card and all peripherals on SOM (or defined by a spec) can
be used by U-Boot. Full DT for SOM+CC can be merged together as overlays.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Mainline kernel has psgtr driver that's why it is good to add description
to DT files. Some boards are just missing description for USB3 and sata.
zc1751-dc1 and p-a2197 are also missing clock descriptions for input
clocks.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
ZynqMP pinctrl Linux driver has been merged to 5.13-rc1 kernel. Based on it
DT files can be extended by pinctrl configurations.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Probe zynqmp firmware driver by adding zynqmp firmware, power &
ipi mailbox device tree nodes for mini emmc.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Add missing xlnx,mio-bank property to sdhci node. Also add properties with
0 value to have it listed in case that files are copied to different
projects where default case doesn't need to be handled in the same way.
That's why explicitly list them too.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
I2C muxes that have the slave devices with same address are
falling into the below problem.
VCK190 system controller (SC) - zynqmp-e-a2197-00-revA.dts
I2C1 (0xff030000) -> Mux1 (@0x74) -> Channel 3 -> 0x50
I2C1 (0xff030000) -> Mux2 (@0x75) -> Channel 0 -> 0x50
1. SC accesses I2C1 - Mux1 (0x74) - Channel 3 and then
2. SC accesses I2C1 - Mux2 (0x75) - Channel 0.
Now it results in 2 slave devices with same address (0x50)
on the I2C bus, making the communication un-reliable.
When ' i2c-mux-idle-disconnect' is in DT, after '1', the Mux
channel output is disconnected, making none of the channels
available to the I2C1. So, there is no question of having the
same addressed slave (0x50) present on the bus when we are doing '2'.
Same pattern is seen in below two boards also.
ZCU208 - zynqmp-zcu208-revA.dts
ZCU216 - zynqmp-zcu216-revA.dts
Signed-off-by: Raviteja Narayanam <raviteja.narayanam@xilinx.com>
Add label to GPIO lines so the user-level applications can find any line
without knowing its physical path on System Controller on VCK190/VMK180.
These labels are describing EMIO gpio connection which depends on PL which
we normally don't describe but that's only way to go for now. Lately this
should be done out of this source code.
Signed-off-by: Saeed Nowshadi <saeed.nowshadi@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The 'silabs,skip-recall' property prevents interruption in operation of
the clock while the driver is being probed. Without this property, the
DDR DIMM clk can cause a failure during Versal's boot.
Signed-off-by: Saeed Nowshadi <saeed.nowshadi@xilinx.com>
All si570 which are used for ps reference clock generation should contain
silabs,skip-recall property not to cause break on ps clock.
On Versal boards this will cause hang on Versal cpu when it is booted at
the same time with SC.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Currently all assembly optimized implementation of memory routines
show up at the top level of the RISC-V architecture Kconfig menu.
Let's group them together into a submenu.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
At present U-Boot SPL fails to boot on SiFive Unleashed board, due
to a load address misaligned exception happens when loading the FIT
image in spl_load_simple_fit(). The exception happens in memmove()
which is called by fdt_splice_().
Commit 8f0dc4cfd1 introduces an assembly version of memmove but
it does take misalignment into account (it checks if length is a
multiple of machine word size but pointers need also be aligned).
As a result it will generate misaligned load/store for the majority
of cases and causes significant performance regression on hardware
that traps misaligned load/store and emulate them using firmware.
The current behaviour of memcpy is that it checks if both src and
dest pointers are co-aligned (aka congruent modular SZ_REG). If
aligned, it will copy data word-by-word after first aligning
pointers to word boundary. If src and dst are not co-aligned,
however, byte-wise copy will be performed.
This patch was taken from the Linux kernel patch [1], which has not
been applied at the time being. It fixes the memmove and optimises
memcpy for misaligned cases. It will first align destination pointer
to word-boundary regardless whether src and dest are co-aligned or
not. If they indeed are, then wordwise copy is performed. If they
are not co-aligned, then it will load two adjacent words from src
and use shifts to assemble a full machine word. Some additional
assembly level micro-optimisation is also performed to ensure more
instructions can be compressed (e.g. prefer a0 to t6).
With this patch, U-Boot boots again on SiFive Unleashed board.
[1] https://patchwork.kernel.org/project/linux-riscv/patch/20210216225555.4976-1-gary@garyguo.net/
Fixes: 8f0dc4cfd1 ("riscv: assembler versions of memcpy, memmove, memset")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
If /chosen was missing, chosen_offset would never get updated with the new
/chosen node. This would cause fdt_setprop_u32 to fail. This patch fixes
this by setting chosen_offset. In addition, log any errors from setting
boot-hartid as well.
Fixes: 5370478d1c ("riscv: Add boot hartid to device tree")
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Rick Chen <rick@andestech.com>
Reviewed-by: Atish Patra <atish.patra@wdc.com>
At present there is only one Kconfig option CONFIG_SIFIVE_CLINT to
control the enabling of SiFive CLINT support in both SPL (M-mode)
and U-Boot proper (S-mode). So for a typical SPL config that the
SiFive CLINT driver is enabled in both SPL and U-Boot proper, that
means the S-mode U-Boot tries to access the memory-mapped CLINT
registers directly, instead of the normal 'rdtime' instruction.
This was not a problem before, as the hardware does not forbid the
access from S-mode. However this becomes an issue now with OpenSBI
commit 8b569803475e ("lib: utils/sys: Add CLINT memregion in the root domain")
that the SiFive CLINT register space is protected by PMP for M-mode
access only. U-Boot proper does not boot any more with the latest
OpenSBI, that access exceptions are fired forever from U-Boot when
trying to read the timer value via the SiFive CLINT driver in U-Boot.
To solve this, we need to split current SiFive CLINT support between
SPL and U-Boot proper, using 2 separate Kconfig options.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
This patch adds the dtsi/dts files needed to support the Marvell
Octeon TX2 CN913x DB. This is only the base port with not all
interfaces supported fully.
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Some of the setups including cn9130 opens mmio window starting from
0xc0000000, reflect it in the u-boot code.
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Signed-off-by: Kostya Porotchkin <kostap@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
There is region left by ATF, which needs to remain in memory to provide RT
services. To prevent overwriting it by u-boot, do not provide any mapping
for this memory region, so any attempt to access it will trigger
synchronous exception.
Update sr 2021-04-12:
Don't update armada3700/cpu.c mmu table, as this has specific changes
included in mainline.
Signed-off-by: Grzegorz Jaszczyk <jaz@semihalf.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This lacks the right u-boot specific DT include to make HDMI work.
Reported-by: B1oHazard <ty3uk@mail.ua>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
This reverts commit bc8bbb77f7.
This commit breaks U-Boot booting on SiFive Unleashed board, as
there is no such CSR on U54 core.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
Testing showed that disabling AI while leaving PLL1 enabled disabled the
aisram. This suggests that AI is a more appropriate clock for that ram
bank.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
This is more consistent with the naming of other ram banks, and matches
what Linux is doing.
Reported-by: Damien Le Moal <Damien.LeMoal@wdc.com>
Signed-off-by: Sean Anderson <seanga2@gmail.com>
These devices are necessary for the clock driver, which is required by the
sram driver, to run pre-relocation.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Update the Kconfig and the board file to make the am335x board compatible
with cape detection.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
The am335x from BeagleBone use i2c EEPROM to detect capes.
The memory is wired to i2c bus 2 therefore it need to be enabled.
Add i2c2 clock, pinmux description and pinmux enable function.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Add the extension_board_scan specific function to scan the information
of the EEPROM on one-wire and fill the extension struct.
Add the Kconfig symbol to enable the needs to detect DIPs.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Reviewed-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Andre Przywara <andre.przywara@arm.com>
This commit extends the sandbox to implement a dummy
extension_board_scan() function and enables the extension command in
the sandbox configuration. It then adds a test that checks the proper
functionality of the extension command by applying two Device Tree
overlays to the sandbox Device Tree.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
[trini: Limit to running on sandbox]
Signed-off-by: Tom Rini <trini@konsulko.com>
Scrub duplicate /memory@* node entries here. Some R-Car DTs might
contain multiple /memory@* nodes, however fdt_fixup_memory_banks()
either generates single /memory node or updates the first /memory
node. Any remaining memory nodes are thus potential duplicates.
However, it is not possible to delete all the memory nodes right
away, since some of those might not be DRAM memory nodes, but some
sort of other memory. Thus, delete only the memory nodes which are
in the R-Car3 DBSC ranges.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
