The fixup routine must not fixup NULL pointers.
Problem can be seen by
char *testfun(void) __attribute__((weak));
char *(*myfun)(void) = testfun;
Then add
printf("myfun:%p, &myfun:%p\n", myfun, &myfun);
before relocation and after relocation.
myfun should be NULL in both cases but it is not.
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
The change is currently needed to be able to remove the board
configuration scripting from the top level Makefile and replace it by
a simple, table driven script.
Moving this configuration setting into the "CONFIG_*" name space is
also desirable because it is needed if we ever should move forward to
a Kconfig driven configuration system.
Signed-off-by: Wolfgang Denk <wd@denx.de>
If 36-bit physical address is used, move the INIT_RAM_ADDR to higher
address. This frees the low 4GB address space for better use.
Signed-off-by: York Sun <yorksun@freescale.com>
As discussed on the list, move "arch/ppc" to "arch/powerpc" to
better match the Linux directory structure.
Please note that this patch also changes the "ppc" target in
MAKEALL to "powerpc" to match this new infrastructure. But "ppc"
is kept as an alias for now, to not break compatibility with
scripts using this name.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Wolfgang Denk <wd@denx.de>
Acked-by: Detlev Zundel <dzu@denx.de>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
Cc: Peter Tyser <ptyser@xes-inc.com>
Cc: Anatolij Gustschin <agust@denx.de>
The HID1[MBDD] bit is new on rev5.0 or greater cores and will optimize
the performance of mbar/eieio instructions.
Signed-off-by: Sandeep Gopalpet <sandeep.kumar@freescale.com>
Use the same code between primary and secondary cores to init the
L1 cache. We were not enabling cache parity on the secondary cores.
Also, reworked the L1 cache init code to match the e500mc L2 init code
that first invalidates the cache and locks. Than enables the cache and
makes sure its enabled before continuing.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
r14 is not supposed to be clobbered by functions. Switch
to r12 and call GET_GOT when needed. This will allow u-boot
to loose the -ffixed-r14 gcc option.
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Using the GOT in IRQ handlers requires r14 to be -ffixed-r14.
Avoid this by relocatate transfer_to_handler too.
This will allow to free up r14 later on.
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
The e500mc core supports the ability to stash into the L1 or L2 cache,
however we need to uniquely identify the caches with an id.
We use the following equation to set the various stash-ids:
32 + coreID*2 + 0(L1) or 1(L2)
The 0 (for L1) or 1 (for L2) matches the CT field used be various cache
control instructions.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
NULL is an absolute value and should not be relocated.
After this correction code like:
void weak_fun(void) __attribute__((weak));
printf("weak_fun:%p\n", weak_fun);
will still print null after relocation.
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
The MPC8536E is capable of booting form NAND/eSDHC/eSPI, this patch
implements these three bootup methods in a unified way - all of these
use the general cpu/mpc85xx/start.S, and load the main image to L2SRAM
which lets us use the SPD to initialize the SDRAM.
For all three bootup methods, the bootup process can be divided into two
stages: the first stage will initialize the corresponding controller,
configure the L2SRAM, then copy the second stage image to L2SRAM and
jump to it. The second stage image is just like the general U-Boot image
to configure all the hardware and boot up to U-Boot command line.
When boot from NAND, the eLBC controller will first load the first stage
image to internal 4K RAM buffer because it's also stored on the NAND
flash. The first stage image, also call 4K NAND loader, will initialize
the L2SRAM, load the second stage image to L2SRAM and jump to it. The 4K
NAND loader's code comes from the corresponding nand_spl directory, along
with the code twisted by CONFIG_NAND_SPL.
When boot from eSDHC/eSPI, there's no such a first stage image because
the CPU ROM code does the same work. It will initialize the L2SRAM
according to the config addr/word pairs on the fixed address and
initialize the eSDHC/eSPI controller, then load the second stage image
to L2SRAM and jump to it.
The macro CONFIG_SYS_RAMBOOT is used to control the code to produce the
second stage image for all different bootup methods. It's set in the
board config file when one of the bootup methods above is selected.
Signed-off-by: Mingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Factor out the code we use to actually write a tlb entry.
set_tlb is a logical view of the TLB while write_tlb is a low level
matching the MAS registers.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Its reset value is random, and we sometimes read uninitialized TLB
arrays. Make sure that we don't retain MAS8 from reading such an entry
if the VF bit in MAS8 is set, attempts to use the mapping will trap.
Signed-off-by: Scott Wood <scottwood@freescale.com>
We cannot handle any exceptions while running in AS1, as the exceptions
will transition back to AS0 without a valid mapping.
Signed-off-by: Scott Wood <scottwood@freescale.com>
In future Book-E implementations IVORs will most likely go away and be
replaced with fixed offsets. The IVPR will continue to exist to allow
for relocation of the interrupt vectors.
This code adds support to setup the IVORs as their fixed offset values
per the ISA 2.06 spec when we transition from u-boot to another OS
either via 'bootm' or a cpu release.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The MPC8536E is capable of booting form NAND/eSDHC/eSPI, this patch
implements these three bootup methods in a unified way - all of these
use the general cpu/mpc85xx/start.S, and load the main image to L2SRAM
which lets us use the SPD to initialize the SDRAM.
For all three bootup methods, the bootup process can be divided into two
stages: the first stage will initialize the corresponding controller,
configure the L2SRAM, then copy the second stage image to L2SRAM and
jump to it. The second stage image is just like the general U-Boot image
to configure all the hardware and boot up to U-Boot command line.
When boot from NAND, the eLBC controller will first load the first stage
image to internal 4K RAM buffer because it's also stored on the NAND
flash. The first stage image, also call 4K NAND loader, will initialize
the L2SRAM, load the second stage image to L2SRAM and jump to it. The 4K
NAND loader's code comes from the corresponding nand_spl directory, along
with the code twisted by CONFIG_NAND_SPL.
When boot from eSDHC/eSPI, there's no such a first stage image because
the CPU ROM code does the same work. It will initialize the L2SRAM
according to the config addr/word pairs on the fixed address and
initialize the eSDHC/eSPI controller, then load the second stage image
to L2SRAM and jump to it.
The macro CONFIG_SYS_RAMBOOT is used to control the code to produce the
second stage image for all different bootup methods. It's set in the
board config file when one of the bootup methods above is selected.
Signed-off-by: Mingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Factor out the code we use to actually write a tlb entry.
set_tlb is a logical view of the TLB while write_tlb is a low level
matching the MAS registers.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Its reset value is random, and we sometimes read uninitialized TLB
arrays. Make sure that we don't retain MAS8 from reading such an entry
if the VF bit in MAS8 is set, attempts to use the mapping will trap.
Signed-off-by: Scott Wood <scottwood@freescale.com>
We cannot handle any exceptions while running in AS1, as the exceptions
will transition back to AS0 without a valid mapping.
Signed-off-by: Scott Wood <scottwood@freescale.com>
In future Book-E implementations IVORs will most likely go away and be
replaced with fixed offsets. The IVPR will continue to exist to allow
for relocation of the interrupt vectors.
This code adds support to setup the IVORs as their fixed offset values
per the ISA 2.06 spec when we transition from u-boot to another OS
either via 'bootm' or a cpu release.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
There is a workaround for MPC8569 CPU Errata, which needs to set Bit 13 of
LBCR in 4K bootpage. We setup a temp TLB for eLBC controller in bootpage,
then invalidate it after LBCR bit 13 is set.
Signed-off-by: Haiying Wang <Haiying.Wang@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
If one custom board is using the 8MB flash, it is set
as FLASH_BASE = 0xef000000, TEXT_BASE = 0xef780000.
The current start.S code will be broken at switch_as.
It is because the TLB1[15] is set as 16MB page size,
EPN = TEXT_BASE & 0xff000000, RPN = 0xff000000.
For the 8MB flash case, the EPN = 0xefxxxxxx,
RPN = 0xffxxxxxx. Assume the virt address of switch_as
is 0xef7ff18c, the real address of the instruction at
switch_as should be 0xff7ff18c. the 0xff7ff18c is out
of the range of the default 8MB boot LAW window
0xff800000 - 0xffffffff.
So when we switch to AS1 address space at switch_as,
the core can't fetch the instruction at switch_as any
more. It will cause broken issue.
Signed-off-by: Dave Liu <daveliu@freescale.com>
Introduce CONFIG_E500MC to deal with the minor differences between
e500v2 and e500mc.
* Certain fields of HID0/1 don't exist anymore on e500mc
* Cache line size is 64-bytes on e500mc
* reset value of PIR is different
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Using CONFIG_SYS_CACHELINE_SIZE instead of 31 means we can handle
e500mc's 64-byte cacheline properly when it gets added.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Added the ability for C code to invalidate the i/d-cache's and
to flush the d-cache. This allows us to more efficient change mappings
from cache-able to cache-inhibited.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
With a page size of BOOKE_PAGESZ_16M, both the real and effective
addresses must be multiples of 16MB. The hardware silently truncates
them so the code happens to work. This patch clarifies the situation
by establishing addresses that the hardware doesn't need to truncate.
Signed-off-by: Andrew Klossner <andrew@cesa.opbu.xerox.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
This commit gets rid of a huge amount of silly white-space issues.
Especially, all sequences of SPACEs followed by TAB characters get
removed (unless they appear in print statements).
Also remove all embedded "vim:" and "vi:" statements which hide
indentation problems.
Signed-off-by: Wolfgang Denk <wd@denx.de>
There is no reason to icbi when invalidating the temporary stack in
the d-cache. Its impossible on e500 to have the i-cache contain
any addresses in the temp stack and it can be problematic in generating
transactions on the bus to non-valid addresses.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Commit 0db37dc... (and some others) changed the INIT_RAM TLB
mappings to be unguarded. This collided with an existing "bug"
where the mappings for the INIT_RAM were being kept around.
This meant that speculative loads to those addresses were
succeeding in the TLB, and going out to the bus, where they
were causing an exception (there's nothing at that address). The
Flash code was coincidentally causing such a speculative load.
Rather than go back to mapping the INIT RAM as guarded, we fix
it so that the entries for the INIT_RAM are invalidated. Thus
the speculative loads will fail in the TLB, and have no effect.
Signed-off-by: Andy Fleming <afleming@freescale.com>
Reworked the initial processor initialzation sequence:
* introduced cpu_early_init_f that is run in address space 1 (AS=1)
* Moved TLB/LAW and CCSR init into cpu_early_init_f()
* Reworked initial asm code to do most of the core init before TLBs
The main reasons for these changes are to allow handling of 36-bit phys
addresses in the future and some of the issues that will exist when we
do that.
There are a few caveats on what can be initialized via the LAW and TLB
static tables:
* TLB entry 14/15 can't be initialized via the TLB table
* any LAW that covers the implicit boot window (4G-8M to 4G) must map to
the code that is currently executing.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Move the initialization of the LAWs into C code and provide an API
to allow modification of LAWs after init.
Board code is responsible to provide a law_table and num_law_entries.
We should be able to use the same code on 86xx as well.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Either use the standard defines in asm/cache.h or grab the information
at runtime from the L1CFG SPR.
Also, minor cleanup in cache.h to make the code a bit more readable.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The mpc85xx start code uses some magic numbers that we actually
have #defines for in <config.h> so use those instead.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
On e500 when we leave the 4k boot page we should use an absolute address since
we don't know where the board code may want us to be really running at.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
From: Ed Swarthout <Ed.Swarthout@freescale.com>
Support external interrupts from platform to eliminate system hangs.
Define CONFIG_INTERRUPTS board configure option to enable.
Enable ecm, ddr, lbc, and pci/pcie error interrupts in PIC.
Remove extra cpu initialization redundant with hardware initialization.
Whitespace cleanup.
Define and use _START_OFFSET consistent with other processors using
ppc_asm.tmpl
Move additional code from .text to boot page to make room for
exception vectors at start of image.
Handle Machine Check, External and Critical exceptions.
Fix e500 machine check error determination in traps.c
TEXT_BASE can now be 0xfffc_0000 - which cuts binary image in half.
Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
Acked-by: Andy Fleming <afleming@freescale.com>
Remove the fixed TLB and LAW entry nubmer. Use actually TLB and LAW
entry number to control the loop. This can reduce the potential risk
for the 85xx processor increasing its TLB adn LAW entry number.
Signed-off-by: Swarthout Edward <swarthout@freescale.com>
Signed-off-by: Roy Zang <tie-fei.zang@freescale.com>
Joakim Tjernlund
Wolfgang Denk
Peter Tyser
york
Stefan Roese
Sandeep Gopalpet
Kumar Gala
Mingkai Hu
Scott Wood
Haiying Wang
Dave Liu
Jean-Christophe PLAGNIOL-VILLARD
Andrew Klossner
Andy Fleming
Rafal Jaworowski
urwithsughosh@gmail.com
Ed Swarthout
Zang Roy-r61911
Wolfgang Denk