PCI outbound address map configuration doesn't match the
PCI memory address range covered by appropriate TLB entry
configuration for canyonlands causing machine check
exceptions while accessing PCI memory regions. This patch
provides a fix for this issue.
Kazuaki Ichinohe observed and reported this issue while
testing display output with PCI ATI video card on canyonlands.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
Total5200 and digsy MTC use I2C port 2 pins as a ATA chip select.
To avoid adding board-specific ifdefs to cpu/mpc5xxx/ide.c new
define CONFIG_SYS_ATA_CS_ON_I2C2 was introduced. It is used by
Total5200 and will be used by digsy MTC and other boards with
ATA CS on I2C pins.
Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
The serial boot dram extended/standard mode register was not
setup and was using default DRAM setup causing the U-boot was
unstable to boot up in serial mode.
Signed-off-by: TsiChung Liew <Tsi-Chung.Liew@freescale.com>
According to coldfire manual data timeout > address time out
also use correct macro to program XARB_CFG
Signed-off-by: Arun C <arunedarath@mistralsolutions.com>
The routine assumed 4 bank SDRAMs, enhance to configure for 4 or 8
bank SDRAMs.
Signed-off-by: Gerald Van Baren <vanbaren@cideas.com>
Acked-by: Dave Liu <daveliu@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
This patch makes sure the correct mask is applied when setting
the encryption and I2C bus 0 clock in SCCR.
Failing to do so may lead to ENCCM being 0 in which case I2C bus 0
won't function.
Signed-off-by: Norbert van Bolhuis <nvbolhuis@aimvalley.nl>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Commit e1be0d25, "32bit BUg fix for DDR2 on 8572" prevented other
sdram_cfg bits (such as ecc and self_refresh_in_sleep) from being set.
Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
add CONFIG_BOOTCOUNT_LIMIT feature for 8360 CPU.
The bootcounter uses 8 bytes from the muram,
because no other memory was found on this
CPU for the bootcount feature. So we must
correct the muram size in DTS before booting
Linux.
This feature is actual only implemented for
MPC8360, because not all 83xx CPU have qe,
and therefore no muram, which this feature
uses.
Signed-off-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
1. RD_TO_PRE missed to add the AL, and need min 2 clocks for
tRTP according to DDR2 JEDEC spec.
2. WRTORD - tWTR need min 2 clocks according to DDR2 JEDEC spec.
3. add the support of DDR2-533,667,800 DIMMs
4. cpo
5. make the AL to min to gain better performance.
The Micron MT9HTF6472CHY-667D1 DIMMs test passed on
MPC837xEMDS platform at 266MHz/333MHz/400MHz data rate.
items 1, 2 and 5:
Acked-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Reported-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Signed-off-by: Dave Liu <daveliu@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
We should use pci_last_busno() in pci_init_bus(), otherwise we'll
erroneously re-use PCI0's first_busno for PCI1 hoses.
NOTE: The patch is untested. All MPC83xx FSL boards I have have
PCI1 in miniPCI form, for which I don't have any cards handy.
But looking in cpu/mpc85xx/pci.c:
...
#ifdef CONFIG_MPC85XX_PCI2
hose = &pci_hose[1];
hose->first_busno = pci_hose[0].last_busno + 1;
And considering that we do the same for MPC83xx PCI-E support,
I think this patch is correct.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
This patch fixes copy-paste issue: pci_hose[0]'s first and last
busnos were used to fixup pci1's nodes.
We don't see this bug triggering only because Linux reenumerate
buses anyway.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
This patch fixes an issue in config space read accessors: we should
fill-in the value even if we fail (e.g. skipping devices), otherwise
CONFIG_PCI_SCAN_SHOW reports bogus values during boot up.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Currently we assign first_busno = 0 for the first PCIe hose, but this
scheme won't work if we have ordinary PCI hose already registered (its
first_busno value is 0 too).
The old code worked fine only because we have PCI disabled on
MPC837XEMDS boards in stand-alone mode (see commit 00f7bbae92
"mpc83xx: fix PCI scan hang on the standalone MPC837xE-MDS boards").
But on MPC837XERDB boards we have PCI and PCIe, so the bug actually
triggers.
So, to fix the issue, we should use pci_last_busno() + 1 for the
first_busno (i.e. last available busno).
Reported-by: Huang Changming <Chang-Ming.Huang@freescale.com>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
This patch adds OMAP3 cpu type auto detection based on OMAP3 register
and removes hardcoded values.
Signed-off-by: Steve Sakoman <sakoman@gmail.com>
Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
Looks like the initcode updates fell out of order during my merges. The
patch that really fixes up this code is part of power-on overhaul and so
is too large for merging at this point. Instead, we can disable the code
as no currently in-tree board depends on it. The next merge window will
fix things up properly.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
In the commit 79b51ff820 ([MIPS] cpu/mips/
Makefile: Split [CS]OBJS onto separate lines), I wrongly deleted a START
line. This patch puts it back.
Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
Fixing the get_timer function to return time in miliseconds instead of
ticks. Also fixed PXA boards to use the conventional value of 1000 for
CONFIG_SYS_HZ.
Signed-off-by: Micha Kalfon <smichak.uv@gmail.com>
The current MMC infrastructure relies on the existence of an
arch-specific header file. This isn't necessary, and a couple
drivers were forced to implement dummy files to meet this requirement.
Instead, we move the stuff in those header files into a more appropriate
place, and eliminate the stubs and the #include of asm/arch/mmc.h
Signed-off-by: Andy Fleming <afleming@freescale.com>
This errata fix is required for 32 bit DDR2 controller on 8572.
May also be required for P10XX20XX platforms
Signed-off-by: Poonam_Agarwal-b10812 <b10812@lc1106.zin33.ap.freescale.net>
Update the 86xx reset sequence to try executing a board-specific reset
function. If the board-specific reset is not implemented or does not
succeed, then assert #HRESET_REQ. Using #HRESET_REQ is a more standard
reset procedure than the previous method and allows all board
peripherals to be reset if needed.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Previously if we >=4G of memory and !CONFIG_PHYS_64BIT we'd report
an error and hang. Instead of doing that since DDR is mapped in the
lowest priority LAWs we setup the DDR controller and the max amount
of memory we report back is what we can map (CONFIG_MAX_MEM_MAPPED)
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Becky Bruce <beckyb@kernel.crashing.org>
Added various p2020 processor specific details:
* SVR for p2020, p2020E
* immap updates for LAWs and DDR on p2020
* LAW defines related to p2020
Signed-off-by: Srikanth Srinivasan <srikanth.srinivasan@freescale.com>
Signed-off-by: Travis Wheatley <Travis.Wheatley@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Previously we only allowed power-of-two memory sizes and didnt
handle >2G of memory. Now we will map up to CONFIG_MAX_MEM_MAPPED
and should properly handle any size that we can make in the TLBs
we have available to us
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
If we only have one controller we can completely ignore how
memctl_intlv_ctl is set. Otherwise other levels of code get confused
and think we have twice as much memory.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
This fixes the initialization of the SDRAM_CODT register in the ppc4xx DDR2
initialization code. It also removes use of the SDRAM_CODT_FEEDBACK_RCV_SINGLE_END
and SDRAM_CODT_FEEDBACK_DRV_SINGLE_END #define's since they are reserved bits.
Signed-off-by: Carolyn Smith <carolyn.smith@tektronix.com>
Signed-off-by: Stefan Roese <sr@denx.de>
The criteria of the AMCC SDRAM Controller DDR autocalibration
U-Boot code is to pick the largest passing write/read/compare
window that also has the smallest SDRAM_RDCC.[RDSS] Read Sample
Cycle Select value.
On some Kilauea boards the DDR autocalibration algorithm can
find a large passing write/read/compare window with a small
SDRAM_RDCC.[RDSS] aggressive value of Read Sample Cycle Select
value "T1 Sample".
This SDRAM_RDCC.[RDSS] Read Sample Cycle Select value of
"T1 Sample" proves to be to aggressive when later on U-Boot
relocates into DDR memory and executes.
The memory traces on the Kilauea board are short so on some
Kilauea boards the SDRAM_RDCC.[RDSS] Read Sample Cycle Select
value of "T1 Sample" shows up as a potentially valid value for
the DDR autocalibratiion algorithm.
The fix is to define a weak default function which provides
the minimum SDRAM_RDCC.[RDSS] Read Sample Cycle Select value
to accept for DDR autocalibration. The default will be the
"T2 Sample" value. A board developer who has a well defined
board and chooses to be more aggressive can always provide
their own board specific string function with the more
aggressive "T1 Sample" value or stick with the default
minimum SDRAM_RDCC.[RDSS] value of "T2".
Also put in a autocalibration loop fix for case where current
write/read/compare passing window size is the same as a prior
window size, then in this case choose the write/read/compare
result that has the associated smallest RDCC T-Sample value.
Signed-off-by: Adam Graham <agraham@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Anatolij Gustschin