ctrl_regs.c:31:5: warning: symbol 'fsl_ddr_get_version' was not declared. Should it be static?
cpu.c:135:14: warning: non-ANSI function declaration of function 'cpu_mask'
cpu.c:154:18: warning: non-ANSI function declaration of function 'cpu_numcores'
cpu.c:37:17: warning: symbol 'cpu_type_list' was not declared. Should it be static?
cpu.c:117:17: warning: symbol 'cpu_type_unknown' was not declared. Should it be static?
fsl_lbc.c:14:6: warning: symbol '__lbc_sdram_init' was not declared. Should it be static?
and:
lc_common_dimm_params.c:15:1: warning: symbol 'compute_cas_latency_ddr3' was not declared. Should it be static?
making it static produces the following compiler warning:
lc_common_dimm_params.c:15:1: warning: 'compute_cas_latency_ddr3' defined but not used [-Wunused-function]
so we protect it with the preprocessor.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Since empty DIMM slot is allowed on other than the first slot, remove the
error message if SPD is not found in this case.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Based on populated DIMMs, automatically select from cs0_cs1_cs2_cs3 or
cs0_cs1 interleaving, or non-interleaving if not available.
Fix the message of interleaving disabled if controller interleaving
is enabled but DIMMs don't support it.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Boot space translation utilizes the pre-translation address to select
the DDR controller target. However, the post-translation address will be
presented to the selected DDR controller. It is possible that the pre-
translation address selects one DDR controller but the post-translation
address exists in a different DDR controller when using certain DDR
controller interleaving modes. The device may fail to boot under these
circumstances. Note that a DDR MSE error will not be detected since DDR
controller bounds registers are programmed to be the same when configured
for DDR controller interleaving.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
When ECC is enabled, DDR controller needs to initialize the data and ecc.
The wait time can be calcuated with total memory size, bus width, bus speed
and interleaving mode. If it went wrong, it is bettert to timeout than
waiting for D_INIT to clear, where it probably hangs.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Fix handling quad-rank DIMMs in a system with two DIMM slots and first
slot supports both dual-rank DIMM and quad-rank DIMM.
For systems with quad-rank DIMM and double dual-rank DIMMs, cs_config
registers need to be enabled to maintain proper ODT operation. The
inactive CS should have bnds registers cleared.
Fix the turnaround timing for systems with all chip-selects enabled. This
wasn't an issue before because DDR was running lower than 1600MT/s with
this interleaving mode.
Fix DDR address calculation. It wasn't an issue until we have multiple
controllers with each more than 4GB and interleaving is disabled.
It also fixes the message of DDR: 2 GiB (DDR3, 64-bit, CL=0.5, ECC off)
when debugging DDR and first DDR controller is disabled. With the fix,
the first enabled controller information will be displayed.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
DDRC ver 4.7 adds DDR_SLOW bit in sdram_cfg_2 register. This bit needs to be
set for speed lower than 1250MT/s.
CDR1 and CDR2 are control driver registers. ODT termination valueis for
IOs are defined. Starting from DDRC 4.7, the decoding of ODT for IOs is
000 -> Termsel off
001 -> 120 Ohm
010 -> 180 Ohm
011 -> 75 Ohm
100 -> 110 Ohm
101 -> 60 Ohm
110 -> 70 Ohm
111 -> 47 Ohm
Add two write leveling registers. Each QDS now has its own write leveling
start value. In case of zero value, the value of QDS0 will be used. These
values are board-specific and are set in board files.
Extend DDR register timing_cfg_1 to have 4 bits for each field.
DDR control driver registers and write leveling registers are added to
interactive debugging for easy access.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Before proper environment is setup, we extract hwconfig and put it into a
buffer with size HWCONFIG_BUFFER_SIZE. We need to enlarge the buffer to
accommodate longer string. Since this macro is used in multiple files, we
move it into arch/powerpc/include/asm/config.h.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Add one more argument to call function readline_into_buffer().
Fix print SPD format for negative values.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Only the first DIMM of first controller should fall back to raw timing
parameters if SPD is missing or corrupted.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Empty slot should be skipped when calculating CAS latency.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Faster DDR3 timing requires parameters exceeding previously defined
range. Extended parameters are fixed. Added some debug messages.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Restructure DDR interleaving option to support 3 and 4 DDR controllers
for 2-, 3- and 4-way interleaving.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Required by JEDEC 79-3E for high speed DDR3.
Also change "CSn disabled" message to debug.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
When the DDR3 speed goes higher, we need to utilize fine offset
from SPD.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
fsl_ddr_get_dimm_params() should be wrapped by
CONFIG_SYS_DDR_RAW_TIMING, otherwise, when using fixed_sdram() instead of
using SPD, it will cause compile error.
Signed-off-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Acked-by: York Sun <yorksun@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Fix:
interactive.c: In function 'fsl_ddr_interactive':
interactive.c:1357:15: warning: variable 'len' set but not used [-Wunused-but-set-variable]
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Newer JEDEC DDR3 SPD Specifications define several additional values for
the DDR3 module_type field which were undefined when this code was
written. Update the code to handle the newer module types.
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Cc: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Fix:
ctrl_regs.c: In function 'set_ddr_sdram_cfg_2':
ctrl_regs.c:641:15: warning: variable 'rcw_en' set but not used [-Wunused-but-set-variable]
ctrl_regs.c: In function 'compute_fsl_memctl_config_regs':
ctrl_regs.c:951:31: warning: array subscript is above array bounds [-Warray-bounds]
ctrl_regs.c:752:34: warning: array subscript is above array bounds [-Warray-bounds]
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Fix:
options.c: In function 'populate_memctl_options':
options.c:486:28: warning: variable 'pdodt' set but not used [-Wunused-but-set-variable]
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
ddr1_dimm_params.c: In function 'compute_ranksize':
ddr1_dimm_params.c:44: warning: format '%08x' expects type 'unsigned int', but
argument 2 has type 'long long unsigned int'
ddr2_dimm_params.c: In function 'compute_ranksize':
ddr2_dimm_params.c:43: warning: format '%08x' expects type 'unsigned int', but
argument 2 has type 'long long unsigned int'
ddr3_dimm_params.c: In function 'compute_ranksize':
ddr3_dimm_params.c:74: warning: format '%16lx' expects type 'long unsigned int',
but argument 2 has type 'long long unsigned int'
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
Interactive DDR debugging provides a user interface to view and modify SPD,
DIMM parameters, board options and DDR controller registers before DDR is
initialized. With this feature, developers can fine-tune DDR for board
bringup and other debugging without frequently having to reprogram the flash.
To enable this feature, define CONFIG_FSL_DDR_INTERACTIVE in board header
file and set an environment variable to activate it. Syntax:
setenv ddr_interactive on
After reset, U-boot prompts before initializing DDR controllers
FSL DDR>
The available commands are
print print SPD and intermediate computed data
reset reboot machine
recompute reload SPD and options to default and recompute regs
edit modify spd, parameter, or option
compute recompute registers from current next_step to end
next_step shows current next_step
help this message
go program the memory controller and continue with u-boot
The first command should be "compute", which reads data from DIMM SPDs and
board options, performs the calculation then stops before setting DDR
controller. A user can use "print" and "edit" commands to view and modify
anything. "Go" picks up from current step with any modification and
compltes the calculation then enables the DDR controller to continue u-boot.
"Recompute" does it over from fresh reading.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Unified DDR driver is maintained for better performance, robustness and bug
fixes. Upgrading to use unified DDR driver for MPC83xx takes advantage of
overall improvement. It requires changes for board files to customize
platform-dependent parameters.
To utilize the unified DDR driver, a board needs to define CONFIG_FSL_DDRx
in the header file. No more boards will be accepted without such definition.
Note: the workaround for erratum DDR6 for the very old MPC834x Rev 1.0/1.1
and MPC8360 Rev 1.1/1.2 parts is not migrated to unified driver.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
DDR2 has different ODT table and values. Adding table according to Samsung
application note.
Fix additive latency calculation to avoid interger underflow.
Also converted typedef dynamic_odt_t to struct dynamic_odt.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The two slots on the same controller have different addresses.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Check second DIMM slot in case the first one is empty.
Honor DQS enable option for SDRAM mode register.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
DDR RCW varies at different speeds. It is common for all platform. Move it
out from corenet_ds.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Extend CAS write Latency (CWL) table to comply with DDR3 spec
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Add this option to allow boards to override the default read-to-write
turnaround time for better performance.
Signed-off-by: York Sun <yorksun@freescale.com>
Checking width before setting DDR controller. SPD for DDR1 and DDR2 has
data width and primary sdram width. The latter one has different meaning
for DDR3.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
In case of empty SPD or checksum error, fallback to raw timing on
supported boards.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
We used to have fixed parameters for soldered DDR chips. This patch
introduces CONFIG_SYS_DDR_RAW_TIMING to enable calculation based on timing
data from DDR chip datasheet, implemneted in board-specific files or header
files.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Add support for 16-bit DDR bus. Also deal with system using 64- and 32-bit
DDR devices.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Only use DDR DIMM part number if SPD has valid length, to prevent from
display garbage in case SPD doesn't cover these fields.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
If the bus width is 32-bit, burst chop should be disabled and burst length
should be 8. Read from SPD or other source to determine the width.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Reword "The DIMM max tCKmin is ..." to "The DDR clock is faster than the slowest
DIMM(s) can support". Fixed interger type in printf as well.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The numeric constants in the switch statements are replaced by #defines
added to the common ddr_spd.h header. This dramatically improves the
readability of the switch statments.
In addition, a few of the longer lines were cleaned up, and the DDR2
type for an SO-RDIMM module was added to the DDR2 switch statement.
Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Cc: Andy Fleming <afleming@gmail.com>
Cc: Kim Phillips <kim.phillips@freescale.com>
Acked-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The current FreeScale MPC-8xxx DDR SPD interpreter is using full 64-bit
integer divide operations to convert between nanoseconds and DDR clock
cycles given arbitrary DDR clock frequencies.
Since all of the inputs to this are 32-bit (nanoseconds, clock cycles,
and DDR frequencies), we can easily restructure the computation to use
the "do_div()" function to perform 64-bit/32-bit divide operations.
On 64-bit this change is basically a no-op, because do_div is
implemented as a literal 64-bit divide operation and the instruction
scheduling works out almost the same.
On 32-bit PowerPC a fully accurate 64/64 divide (__udivdi3 in libgcc) is
over 1.1kB of code and thousands of heavily dependent cycles to compute,
all of which is linked from libgcc. Another 1.2kB of code comes in for
the function __umoddi3.
It should be noted that nothing else in U-Boot or the Linux kernel seems
to require a full 64-bit divide on my 32-bit PowerPC.
Build-and-boot-tested on the HWW-1U-1A board using DDR2 SPD detection.
Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Acked-by: York Sun <yorksun@freescale.com>
Cc: Andy Fleming <afleming@gmail.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Beside displaying RDIMM or UDIMM, this patch adds display of the model numbers
embedded in SPD.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Move fsl_ddr_get_spd into common mpc8xxx/ddr/main.c as most boards
pretty much do the same thing. The only variations are in how many
controllers or DIMMs per controller exist. To make this work we
standardize on the names of the SPD_EEPROM_ADDRESS defines based on the
use case of the board.
We allow boards to override get_spd to either do board specific fixups
to the SPD data or deal with any unique behavior of how the SPD eeproms
are wired up.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Every 85xx board implements fsl_ddr_get_mem_data_rate via get_ddr_freq()
and every 86xx board uses get_bus_freq(). If implement get_ddr_freq()
as a static inline to call get_bus_freq() we can remove
fsl_ddr_get_mem_data_rate altogether and just call get_ddr_freq()
directly.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
rcw_en bit is only available for DDR3 controllers. It is a reserved bit on
DDR1 and DDR2 controllers.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
To recognize DIMMs with ECC capability by testing ECC bit only. Not to be
confused by Address Parity bit.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
The write recovery time of both registers should match. Since mode register
doesn't support cycles of 9,11,13,15, we should use next higher number for
both registers.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
When DDR data rate is higher than 1200MT/s or controller interleaving is
enabled, additional cycle for write-to-read turnaround is needed to satisfy
dynamic ODT timing.
Signed-off-by: York Sun <yorksun@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
ctrl_regs.c: In function 'set_ddr_sdram_mode_2':
ctrl_regs.c:690:6: warning: unused variable 'i'
'i' is only used by DDR3 code.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Kim Phillips
York Sun
Shaohui Xie
Kumar Gala
Ira W. Snyder
Marek Vasut
Kyle Moffett
Poonam Aggrwal