The workaround for ARM errata 725233 had been lost since
commit 45bf05854b (armv7: adapt omap3 to the new cache
maintenance framework). Bring it back in order to avoid
very difficult to reproduce, but actually encountered in
the wild CPU deadlocks when running software rendered
X11 desktop on OMAP3530 hardware.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
[trini: Migrate to Kconfig]
Signed-off-by: Tom Rini <trini@konsulko.com>
On some of the SoCs one cannot enable hypervisor mode directly from the
u-boot because the ROM code puts the chip to supervisor mode after it
jumps to boot loader. Hence introduce a weak function which can be
overridden based on the SoC type and switch to hypervisor mode in a
custom way.
Cc: beagleboard-x15@googlegroups.com
Signed-off-by: Keerthy <j-keerthy@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
At present CONFIG_SKIP_LOWLEVEL_INIT prevents U-Boot from calling
lowlevel_init(). This means that the instruction cache is not enabled and
the board runs very slowly.
What is really needed in many cases is to skip the call to lowlevel_init()
but still perform CP15 init. Add an option to handle this.
Reviewed-by: Heiko Schocher <hs@denx.de>
Tested-on: smartweb, corvus, taurus, axm
Tested-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andreas Bießmann <andreas@biessmann.org>
Add workaround for Cortex-A15 ARM erratum 801819 which says in summary
that "A livelock can occur in the L2 cache arbitration that might
prevent a snoop from completing. Under certain conditions this can
cause the system to deadlock. "
Recommended workaround is as follows:
Do both of the following:
1) Do not use the write-back no-allocate memory type.
2) Do not issue write-back cacheable stores at any time when the cache
is disabled (SCTLR.C=0) and the MMU is enabled (SCTLR.M=1). Because it
is implementation defined whether cacheable stores update the cache when
the cache is disabled it is not expected that any portable code will
execute cacheable stores when the cache is disabled.
For implementations of Cortex-A15 configured without the “L2 arbitration
register slice” option (typically one or two core systems), you must
also do the following:
3) Disable write-streaming in each CPU by setting ACTLR[28:25] = 0b1111
So, we provide an option to disable write streaming on OMAP5 and DRA7.
It is a rare condition to occur and may be enabled selectively based
on platform acceptance of risk.
Applies to: A15 revisions r2p0, r2p1, r2p2, r2p3 or r2p4 and REVIDR[3]
is set to 0.
Note: certain unicore SoCs *might* not have REVIDR[3] not set, but
might not meet the condition for the erratum to occur when they donot
have ACP (Accelerator Coherency Port) hooked to ACE (AXI Coherency
Extensions). Such SoCs will need the work around handled in the SoC
specific manner, since there is no ARM generic manner to detect such
configurations.
Based on ARM errata Document revision 18.0 (22 Nov 2013)
Suggested-by: Richard Woodruff <r-woodruff2@ti.com>
Suggested-by: Brad Griffis <bgriffis@ti.com>
Reviewed-by: Brad Griffis <bgriffis@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Various files are needlessly rebuilt every time due to the version and
build time changing. As version.h is not actually needed, remove the
include.
Signed-off-by: Rob Herring <robh@kernel.org>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Marek Vasut <marex@denx.de>
Cc: Tom Warren <twarren@nvidia.com>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Macpaul Lin <macpaul@andestech.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: York Sun <yorksun@freescale.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Cc: Simon Glass <sjg@chromium.org>
Cc: Philippe Reynes <tremyfr@yahoo.fr>
Cc: Eric Jarrige <eric.jarrige@armadeus.org>
Cc: "David Müller" <d.mueller@elsoft.ch>
Cc: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Robert Baldyga <r.baldyga@samsung.com>
Cc: Torsten Koschorrek <koschorrek@synertronixx.de>
Cc: Anatolij Gustschin <agust@denx.de>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Łukasz Majewski <l.majewski@samsung.com>
621766: Under a specific set of conditions, executing a sequence of
NEON or vfp load instructions can cause processor deadlock
Impacts: Every Cortex-A8 processors with revision lower than r2p1
Work around: Set L1NEON to 1
Based on ARM errata Document revision 20.0 (13 Nov 2010)
Signed-off-by: Nishanth Menon <nm@ti.com>
Tested-by: Matt Porter <mporter@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
430973: Stale prediction on replaced inter working branch causes
Cortex-A8 to execute in the wrong ARM/Thumb state
Impacts: Every Cortex-A8 processors with revision lower than r2p1
Work around: Set IBE to 1
Based on ARM errata Document revision 20.0 (13 Nov 2010)
Signed-off-by: Nishanth Menon <nm@ti.com>
Tested-by: Matt Porter <mporter@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
454179: Stale prediction may inhibit target address misprediction on
next predicted taken branch
Impacts: Every Cortex-A8 processors with revision lower than r2p1
Work around: Set IBE and disable branch size mispredict to 1
Also provide a hook for SoC specific handling to take place if needed.
Based on ARM errata Document revision 20.0 (13 Nov 2010)
Signed-off-by: Nishanth Menon <nm@ti.com>
Tested-by: Matt Porter <mporter@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Add workaround for Cortex-A15 ARM erratum 798870 which says
"If back-to-back speculative cache line fills (fill A and fill B) are
issued from the L1 data cache of a CPU to the L2 cache, the second
request (fill B) is then cancelled, and the second request would have
detected a hazard against a recent write or eviction (write B) to the
same cache line as fill B then the L2 logic might deadlock."
Implementations for SoC families such as Exynos, OMAP5/DRA7 etc
will be widely different.
Every SoC has slightly different manner of setting up access to L2ACLR
and similar registers since the Secure Monitor handling of Secure
Monitor Call(smc) is diverse. Hence an weak function is introduced
which may be overriden to implement SoC specific accessor implementation.
Based on ARM errata Document revision 18.0 (22 Nov 2013)
Signed-off-by: Nishanth Menon <nm@ti.com>
Tested-by: Matt Porter <mporter@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
So that the CONFIG_SPL_FEL option is not needed anymore. And the regular
SPL binary, generated by the default u-boot build, is now also bootable
over USB in the FEL mode. The SPL still can boot from the SD card too.
A bunch of system registers need to be saved/restored in order to ensure
that the IRQ handler still works in the BROM FEL code after getting
control back from the SPL. This is done in the sunxi code instead of
abusing ifdefs in 'start.S'.
The decision whether to load the main u-boot binary from the SD card or
return to the FEL code in the BROM is done at runtime.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
[hdegoede@redhat.com: Since we now restore various regs before returning to
the FEL BROM code we can drop the sunxi specific #ifdefs in start.S]
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Make sunxi's FEL code fit with the normal U-Boot boot sequence instead of
creating its own. There are some #ifdefs required in start.S. Future work
will hopefully remove these.
This series is available at u-boot-dm, branch sunxi-working.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The link register value can be required on some boards (e.g. FEL mode on
sunxi) so use a branch instruction to jump to save_boot_params() instead
of a branch link.
This requires a branch back to save_boot_params_ret so adjust the users
to deal with this. For exynos just drop the function since it doesn't
do anything.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This commit relocates the exception vectors.
As ARM1176 and ARMv7 have the security extensions, it uses VBAR. For
the other ARM processors, it copies the relocated exception vectors to
the correct address: 0x00000000 or 0xFFFF0000.
Signed-off-by: Georges Savoundararadj <savoundg@gmail.com>
Acked-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Tom Warren <twarren@nvidia.com>
Exception handling is basically identical for all ARM targets.
Factorize it out of the various start.S files and into a
single vectors.S file, and adjust linker scripts accordingly.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Full cache line writes to the same memory region from at least two
processors might deadlock the processor. Exists on r1, r2, r3
revisions.
Signed-off-by: Nitin Garg <nitin.garg@freescale.com>
Acked-by: Fabio Estevam <fabio.estevam@freescale.com>
A short loop including a DMB instruction might cause a denial of
service on another processor which executes a CP15 broadcast operation.
Exists on r1, r2, r3, r4 revisions.
Signed-off-by: Nitin Garg <nitin.garg@freescale.com>
Acked-by: Dirk Behme <dirk.behme@de.bosch.com>
Remove the last uses of symbol offsets in ARM U-Boot.
Remove some needless uses of _TEXT_BASE.
Remove all _TEXT_BASE definitions.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Make indirect vectors addresses global, so they can be replaced by
various code that needs to do so. For example the MX6 PCI express
driver needs to temporarily replace data abort handler when reading
the config space.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Eric Nelson <eric.nelson@boundarydevices.com>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Stefano Babic <sbabic@denx.de>
Replace all relocate_code routines from ARM start.S files
with a single instance in file arch/arm/lib/relocate.S.
For PXA, this requires moving the dcache unlocking code
from within relocate_code into c_runtime_cpu_setup.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
Reviewed-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Tested-by: Simon Glass <sjg@chromium.org>
A start vector fix was added by AneeshV for OMAP4 (commit 0d479b53),
and caused the old monilithic Tegra builds to hang due to an undefined
instruction trap. Previously, the code needed to run on both the
AVP (ARM7TDI) and A9, and the AVP doesn't have a CP15 register.
I corrected this in commit 6d6c0bae w/#ifndef CONFIG_TEGRA, but
now that we use SPL, and boot the AVP w/o any ARMv7 code, I can
revert my change, and make Aneesh's change apply to Tegra.
Signed-off-by: Tom Warren <twarren@nvidia.com>
The KVM and Xen hypervisors for the Cortex-A15 virtualization
implementation need to be entered in HYP mode. Should the primary
board firmware already enter HYP mode (Calxeda firmware does that),
we should not deliberately drop back to SVC mode.
Since U-boot does not use the MMU, running in HYP mode is just fine.
Signed-off-by: Andre Przywara <andre.przywara@linaro.org>
When swi instruction is executed, it is expected to get message
"software interrupt" in console and dump registers and reboot, as
do_software_interrupt() in arch/arm/lib/interrupts.c.
But, actually it causes data abort accessing wrong address in get_bad_stack_swi
macro in arch/arm/cpu/v7/start.S.
This patch fixes this problem.
The same mistake in arch/arm/cpu/{arm1136,arm1176,pxa}/start.S.
Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Commit e05e5de7fa made the 2 1st parameters of
ARM's relocate_code() useless since it moved the code handling them to crt0.S.
So, drop these parameters.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Use __image_copy_end instead of __bss_start for the end of the image to
relocate. This is the same as commit 033ca72, but applied to all ARM start.S.
This is a more appropriate symbol naming for an image copy & relocate feature,
and this also saves a useless copy of data put between __image_copy_end and
__bss_start in linker scripts (e.g. relocation information, or MMU
initialization tables used only before jumping to the relocated image).
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Commit e05e5de7fa made ARM's relocate_code()
return to its caller, but it did not update its declaration accordingly.
Fixing this function declaration fixes dropped C code following calls to
relocate_code().
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
_TEXT_BASE must be set to CONFIG_SPL_TEXT_BASE for generic SPL, and to
CONFIG_SYS_TEXT_BASE for non-SPL builds.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Reviewed-by: Tom Rini <trini@ti.com>
The _start symbol is already relocated, so do not add the relocation the second
time in c_runtime_cpu_setup.
This fixes e.g. the abort exception handling path, which ended in double fault
due to bad address in VBAR.
Signed-off-by: Vincent Stehlé <v-stehle@ti.com>
Reported-by: Lubomir Popov <lpopov@mm-sol.com>
Albert's rework of the linker scripts conflicted with Simon's making
everyone use __bss_end. We also had a minor conflict over
README.scrapyard being added to in mainline and enhanced in
u-boot-arm/master with proper formatting.
Conflicts:
arch/arm/cpu/ixp/u-boot.lds
arch/arm/cpu/u-boot.lds
arch/arm/lib/Makefile
board/actux1/u-boot.lds
board/actux2/u-boot.lds
board/actux3/u-boot.lds
board/dvlhost/u-boot.lds
board/freescale/mx31ads/u-boot.lds
doc/README.scrapyard
include/configs/tegra-common.h
Build tested for all of ARM and run-time tested on am335x_evm.
Signed-off-by: Tom Rini <trini@ti.com>
Note this is a tree-wide change affecting multiple architectures.
At present we use __bss_start, but mostly __bss_end__. This seems
inconsistent and in a number of places __bss_end is used instead.
Change to use __bss_end for the BSS end symbol throughout U-Boot. This
makes it possible to use the asm-generic/sections.h file on all
archs.
Signed-off-by: Simon Glass <sjg@chromium.org>
Various errata exist in the Cortex-A9 CPU, and may be worked around by
setting some bits in a CP15 diagnostic register. Add code to implement
the workarounds, enabled by new CONFIG_ options.
This code was taken from the Linux kernel, v3.8, arch/arm/mm/proc-v7.S,
and modified to remove the logic to conditionally apply the WAR (since we
know exactly which CPU we're running on given the U-Boot configuration),
and use r0 instead of r10 for consistency with the rest of U-Boot's
cpu_init_cp15().
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
This patch adds basic Tegra30 (T30) build support - no specific
board is targeted.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Move all the C runtime setup code from every start.S
in arch/arm into arch/arm/lib/crt0.S. This covers
the code sequence from setting up the initial stack
to calling into board_init_r().
Also, rewrite the C runtime setup and make functions
board_init_*() and relocate_code() behave according to
normal C semantics (no jumping across the C stack any
more, etc).
Some SPL targets had to be touched because they use
start.S explicitly or for some reason; the relevant
maintainers and custodians are cc:ed.
Signed-off-by: Albert ARIBAUD <albert.u.boot@aribaud.net>
We don't need to return to cpu_init_crit after calling lowlevel_init, so
lowlevel_init can directly return to the caller of cpu_init_crit.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
In SPL (CONFIG_SPL_FRAMEWORK) board_init_f must setup the stack pointer,
clear the BSS and call board_init_r. We mark this as weak as some
platforms may need to perform additional initalization at this point.
We provide a gd that we know will be in a usable location, once the BSS
has been cleared to help with this as well. Finally, we no longer call
relocate_code so remove that from the armv7 version.
Next, both board_init_f and jump_to_image_linux are going to be
inherently arch-specific, so move these versions to arch/arm/lib/spl.c
Signed-off-by: Tom Rini <trini@ti.com>
save_boot_params_default() in cpu.c accesses uninitialized stack area
when it compiled with -O0 (not optimized).
This patch removes save_boot_params_default() and put the equivalent in start.S
Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Acked-by: Tom Rini <trini@ti.com>
Add SPL options to tegra20 config files and enable SPL build for
tegra20 boards. Also remove redundant code from u-boot that is not
contained in SPL.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Tom Warren <twarren@nvidia.com>
This is make naming consistent with the kernel and devicetree and in
preparation of pulling out the common tegra20 code.
Signed-off-by: Allen Martin <amartin@nvidia.com>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
Tested-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Tom Warren <twarren@nvidia.com>
In currently, when __bss_start is equal to __bss_end__,
The bss loop will clear all the things in memory space.
But just only when __bss_end__ greater than __bss_start__,
we do the clear bss section operation.
Signed-off-by: Zhong Hongbo <bocui107@gmail.com>
Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>
Interrupts and exceptions doesn't work in relocated code.
It badly use IRQ_STACK_START_IN in rom area as interrupt stack.
It is because the vecotr table is not moved to ram area.
This patch moves vector table before jumping relocated code.
Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Tested-by: Tom Rini <trini@ti.com>
Use ENTRY and ENDPROC with assembly functions to ensure
necessary assembler directives for all functions.
Signed-off-by: Aneesh V <aneesh@ti.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
commit 0d479b53 (Aneesh V) added code for OMAP4 that doesn't
execute on Tegra, due to the AVP (ARM7TDI) not having a CP15.
Result was an undefined instruction hang just after reset.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
Acked-by: Aneesh V <aneesh@ti.com>
Since we have cache support built in we can remove Tegra's existing cache
initialization code amd other related dead code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>
