This patch add the capability to configure a PPC440 based IBM SDRAM
Controller with static, compiled-in, values. PPC440 memory subsystem
includes a Memory Queue core.
Signed-off-by: Adam Graham <agraham@amcc.com>
Signed-off-by: Victor Gallardo <vgallardo@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Set the MQ Read Passing & MCIF Cycle limits to the recommended by AMCC
values. This fixes the occasional 440SPe hard locking issues when the 440SPe's
dedicated DMA engines are used (e.g. by the h/w accelerated RAID driver).
Previously the appropriate initialization had been made in Linux, by the
ppc440spe ADMA driver, which is wrong because modifying the MQ configuration
registers after normal operation has begun is not supported and could
have unpredictable results.
Comment from Stefan: This patch doesn't change the resulting value of the
MQ registers. It explicitly sets/clears all bits to the desired state which
better documents the resulting register value instead of relying on pre-set
default values.
Signed-off-by: Yuri Tikhonov <yur@emcraft.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Removed Magic numbers from Initialization preload registers
Tested with Kilauea, Glacier, Canyonlands and Katmai boards
About 5-7% improvement seen for LMBench memtests
Signed-off-by: Prodyut Hazarika <phazarika@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch fixes some minor issues introduced with the patch:
ppc4xx: Optimize PLB4 Arbiter... from Prodyut Hazarika:
- Rework memory-queue and PLB arbiter optimization code, that the
local variable is not needed anymore. This removes one #ifdef.
- Use consistant spacing in ppc4xx.h header (XXX + 0x01 instead
of XXX+ 0x01). This was not introduced by Prodyut, just a
copy-paste problem.
Signed-off-by: Stefan Roese <sr@denx.de>
PPC405EX and PPC460EX/GT/SX
- Read pipeline depth set to 4 for PPC440SP/SPE, PPC405EX, PPC460EX/GT/SX
processors
- Moved PLB4 Arbiter register definitions to ppc4xx.h since it is shared
across processors (405 and 440/460)
- Optimize Memory Queue settings for PPC440SP/SPE and PPC460EX/GT/SX
processors
- Add register bit definitions for Memory Queue Configuration registers
Signed-off-by: Prodyut Hazarika <phazarika@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch completes the preprocessor mneomics for the IBM DDR2 SDRAM
controller registers (MODT and INITPLR) used by the
PowerPC405EX(r). The MMODE and MEMODE registers are unified with their
peer values used for the INITPLR MR and EMR registers,
respectively. Finally, a spelling typo is correct (MANUEL to MANUAL).
With these mnemonics in place, the CFG_SDRAM0_* magic numbers for
Kilauea are replaced by equivalent mnemonics to make it easier to
compare and contrast other 405EX(r)-based boards (e.g. during board
bring-up).
Finally, unified the SDRAM controller register dump routine such that
it can be used across all processor variants that utilize the IBM DDR2
SDRAM controller core. It produces output of the form:
PPC4xx IBM DDR2 Register Dump:
...
SDRAM_MB0CF[40] = 0x00006701
...
which is '<mnemonic>[<DCR #>] = <value>'. The DCR number is included
since it is not uncommon that the DCR values in header files get mixed
up and it helps to validate, at a glance, they match what is printed
in the user manual.
Tested on:
AMCC Kilauea/Haleakala:
- NFS Linux Boot: PASSED
- NAND Linux Boot: PASSED
Signed-off-by: Grant Erickson <gerickson@nuovations.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch fixes ppc4xx related printf format warning. Those warnings are
now visible since patch dc4b0b38d4
[Fix printf errors.] by Andrew Klossner has been applied. Thanks, this is
really helpful.
Signed-off-by: Stefan Roese <sr@denx.de>
Newer PPC's like 440SPe, 460EX/GT can be equipped with more than 2GB of SDRAM.
To support such configurations, we "only" map the first 2GB via the TLB's. We
need some free virtual address space for the remaining peripherals like, SoC
devices, FLASH etc.
Note that ECC is currently not supported on configurations with more than 2GB
SDRAM. This is because we only map the first 2GB on such systems, and therefore
the ECC parity byte of the remaining area can't be written.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch changes the return type of initdram() from long int to phys_size_t.
This is required for a couple of reasons: long int limits the amount of dram
to 2GB, and u-boot in general is moving over to phys_size_t to represent the
size of physical memory. phys_size_t is defined as an unsigned long on almost
all current platforms.
This patch *only* changes the return type of the initdram function (in
include/common.h, as well as in each board's implementation of initdram). It
does not actually modify the code inside the function on any of the platforms;
platforms which wish to support more than 2GB of DRAM will need to modify
their initdram() function code.
Build tested with MAKEALL for ppc, arm, mips, mips-el. Booted on powerpc
MPC8641HPCN.
Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
This patch changes the kilauea and kilauea_nand (for NAND booting)
board port to not use a board specific DDR2 init routine anymore. Now
the common code from cpu/ppc4xx is used.
Thanks to Grant Erickson for all his basic work on this 405EX early
bootup.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch (Part 1 of 2):
* Rolls up a suite of changes to enable correct primordial stack and
global data handling when the data cache is used for such a purpose
for PPC40x-variants (i.e. CFG_INIT_DCACHE_CS).
* Related to the first, unifies DDR2 SDRAM and ECC initialization by
eliminating redundant ECC initialization implementations and moving
redundant SDRAM initialization out of board code into shared 4xx
code.
* Enables MCSR visibility on the 405EX(r).
* Enables the use of the data cache for initial RAM on
both AMCC's Kilauea and Makalu and removes a redundant
CFG_POST_MEMORY flag from each board's CONFIG_POST value.
- Removed, per Stefan Roese's request, defunct memory.c file for
Makalu and rolled sdram_init from it into makalu.c.
With respect to the 4xx DDR initialization and ECC unification, there
is certainly more work that can and should be done (file renaming,
etc.). However, that can be handled at a later date on a second or
third pass. As it stands, this patch moves things forward in an
incremental yet positive way for those platforms that utilize this
code and the features associated with it.
Signed-off-by: Grant Erickson <gerickson@nuovations.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch fixes a problem with DIMMs that have 8 banks. Now the
MCIF0_MBxCF register will be setup correctly for this setup too.
This was noticed with the 512MB DIMM on Canyonlands/Glacier.
Signed-off-by: Stefan Roese <sr@denx.de>
Since the current dflush() implementation is know to have some problems
(as seem on lwmon5 ECC init) this patch removes it completely and replaces
it by using clean_dcache_range().
Tested on Katmai with ECC DIMM.
Signed-off-by: Stefan Roese <sr@denx.de>
On Katmai the complete auto-calibration somehow doesn't seem to
produce the best results, meaning optimal values for RQFD/RFFD.
This was discovered by GDA using a high bandwidth scope,
analyzing the DDR2 signals. GDA provided a fixed value for RQFD,
so now on Katmai "only" RFFD is auto-calibrated.
This patch also adds RDCC calibration as mentioned on page 7 of
the AMCC PowerPC440SP/SPe DDR2 application note:
"DDR1/DDR2 Initialization Sequence and Dynamic Tuning"
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds the CONFIG_4xx_DCACHE options to some SDRAM init files
and to the Sequoia TLB init code. Now the cache can be enabled on 44x
boards by defining CONFIG_4xx_DCACHE in the board config file. This
option will disappear, when more boards use is successfully and no
more known problems exist.
This is tested successfully on Sequoia and Katmai. The only problem that
needs to be fixed is, that USB is not working on Sequoia right now, since
it will need some cache handling code too, similar to the 4xx EMAC driver.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch moves some common 4xx macros and the PPC405_SYS_INFO/
PPC440_SYS_INFO structure into the common ppc4xx.h header.
Lot's of other macros are good candidates to be consolidated this way
in the future.
Signed-off-by: Stefan Roese <sr@denx.de>
platforms wishing to display RAM diagnostics in addition to size,
can do so, on one line, in their own board_add_ram_info()
implementation.
this consequently eliminates CONFIG_ADD_RAM_INFO.
Thanks to Stefan for the hint.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
The new "weak" functions ddr_wrdtr() and ddr_clktr() are added to better
support non default, boardspecific DDR(2) controller configuration.
Signed-off-by: Stefan Roese <sr@denx.de>
- Show on the Status LEDs, some States of the board.
- Get the MAC addresses from the EEProm
- use PREBOOT
- use the CF on the board.
- check the U-Boot image in the Flash with a SHA1
checksum.
- use dynamic TLB entries generation for the SDRAM
Signed-off-by: Heiko Schocher <hs@denx.de>
Now program_tlb() allows to program a TLB (or multiple) with
different virtual and physical addresses. With this change, now one
physical region (e.g. SDRAM) can be mapped 2 times, once with caches
diabled and once with caches enabled.
Signed-off-by: Stefan Roese <sr@denx.de>
Fix a bug in the auto calibration routine. This driver now runs
more reliable with the tested modules. It's also tested with
167MHz PLB frequency (667MHz DDR2 frequency) on the Katmai.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch updates the recently added Katmai board support. The biggest
change is the support of ECC DIMM modules in the 440SP(e) SPD DDR2
driver.
Please note, that still some problems are left with some memory
configurations. See the driver for more details.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds support for the DDR2 controller used on the
440SP and 440SPe. It is tested on the Katmai (440SPe) eval
board and works fine with the following DIMM modules:
- Corsair CM2X512-5400C4 (512MByte per DIMM)
- Kingston ValueRAM KVR667D2N5/512 (512MByte per DIMM)
- Kingston ValueRAM KVR667D2N5K2/2G (1GByte per DIMM)
This patch also adds the nice functionality to dynamically
create the TLB entries for the SDRAM (tlb.c). So we should
never run into such problems with wrong (too short) TLB
initialization again on these platforms.
Signed-off-by: Stefan Roese <sr@denx.de>
