Flushing kernel image after decompression was taking 113 milliseconds
with U-boot 2022.10. With U-boot 2023.01 and 2023.04, flushing
the same amount of memory takes approx 1.5 seconds. With
U-boot 2023.07-rc6, it takes 6.5 seconds.
powerpc flush_cache() function used to call WATCHDOG_RESET() after
flushing every cacheline. At that time WATCHDOG_RESET() was light
so the operation was almost seamless.
But commit 29caf9305b ("cyclic: Use schedule() instead of
WATCHDOG_RESET()") replaced WATCHDOG_RESET() by schedule() and that
started to hurt with U-boot 2022.10.
And in U-boot 2023.07-rc6 that's even worse after
commit 26e8ebcd7c ("watchdog: mpc8xxx: Make it generic").
In the meantime commit 729c1fe656 ("powerpc: introduce
CONFIG_CACHE_FLUSH_WATCHDOG_THRESHOLD") gives us the opportinity to
only call schedule() every given chunk of data instead of every
cacheline. As explained in that commit there is no point in pinging
the watchdog after every cacheline flush, so lets define a sensible
default chunk size of 4k which matches to size of a page on most
powerpc platforms.
With that new default threshold, the culprit flushing performed after
kernel image decompression now takes 85 milliseconds on a powerpc 8xx.
Fixes: 29caf9305b ("cyclic: Use schedule() instead of WATCHDOG_RESET()")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Make sure the serial driver is initialized before relocation by tagging
the serial nodes with "bootph-all". Add these u-boot specific properties
to an *-u-boot.dtsi file.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add the serial node descriptions similar to Linux v6.4 for the
t1024rdb board and its dependencies.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Make sure the serial driver is initialized before relocation by tagging
the serial nodes with "bootph-all". Add these u-boot specific properties
to an *-u-boot.dtsi file.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add the serial node descriptions similar to Linux v6.4 for the
t1042d4rdb board and its dependencies.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Make sure the serial driver is initialized before relocation by tagging
the serial nodes with "bootph-all". Add these u-boot specific properties
to an *-u-boot.dtsi file.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add the serial node descriptions similar to Linux v6.4 for the t4240rdb
board and its dependencies.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Make sure the serial driver is initialized before relocation by tagging
the serial nodes with "bootph-all". Add these u-boot specific properties
to an *-u-boot.dtsi file.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add the serial node descriptions similar to Linux v6.4 for the t2080rdb
board and its dependencies.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Powerpc configurations are apparently able to do unaligned accesses. But
in an attempt to clean up and handle unaligned accesses in the same way
we ignore that and use the common asm-generic/unaligned.h directly
instead.
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
As part of various code clean-ups we have on occasion missed removing
unused header files. None of these files are referenced anywhere else
at this point.
Signed-off-by: Tom Rini <trini@konsulko.com>
In order to use QMC mode in the CPM, a SCC requires more space
in parameter RAM.
After SCC1 there is I2C parameter RAM and after SCC2 there is
SPI parameter RAM. MPC866 and MPC885 can already relocate I2C and.
SPI parameter RAM.
But in order to free space after SCC3 and SCC4, SMC1 and SMC2
need to be relocated. In order to do so, a CPM microcode patch
is required.
Binary data for that patch is copied from Linux kernel.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
MPC885 CPU has the following ERRATA:
When the USB controller is configured in Host mode, and the
SOF generation (SFTE=1 in USMOD register) is being used,
there may be false CRC error indication in other SCCs.
Although the data is received correctly, the CRC result
will be corrupted.
Add capability to load the related microcode to fix it.
The microcode binary data is copied from Linux kernel.
Other microcode will be added in following patch so make it
a Kconfig choice.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
sparse reports the following warning:
CHECK arch/powerpc/cpu/mpc8xx/micropatch_usb_sof.c
arch/powerpc/cpu/mpc8xx/micropatch_usb_sof.c:29:9: warning: cast removes address space '<asn:2>' of expression
arch/powerpc/cpu/mpc8xx/micropatch_usb_sof.c:30:9: warning: cast removes address space '<asn:2>' of expression
This is because of (void *) casts for using memcpy() as a substitute.
Do like other architectures, __force the cast to silence the warning
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Using SMC relocation microcode patch or USB-SOF microcode patch
will disable DPRAM memory from 0x2000 to 0x2400 and from 0x2f00
to 0x3000.
At the time being, init RAM is setup to use 0x2800-0x2e00, but
the stack pointer goes beyond 0x2800 and even beyond 0x2400.
For the time being we are not going to use any microcode patch
that uses memory about 0x3000, so reorganise setup to use:
- 0x2800 - 0x2e00 for init malloc and global data and CPM buffers
- 0x3000 - 0x3c00 for init stack
For more details about CPM dual port ram, see
commit b1d62424cb ("powerpc: mpc8xx: redistribute data in CPM dpram")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
With relocation, CPM parameter RAM can be anywhere in the
dual port RAM, so don't split dual port RAM.
Remove dparam and dparam16 members of struct comm_proc
PROFF_XXX become offsets from the start of dual port RAM,
then they are now consistant with the offsets in RPBASE
registers.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
The "Programming Environments Manual for 32-Bit Implementations of the
PowerPC™ Architecture" says "W and G bits are not defined for IBAT
registers. Attempting to write to these bits causes boundedly-undefined
results"
The "e300 Power Architecture™ Core Family Reference Manual" says the
same: "Neither the W or G bits of the IBAT registers should be set.
Attempting to write to these bits causes boundedly-undefined results."
Remove the possibility to set those bytes.
Fixes: 30915ab95d ("mpc83xx: Migrate BATS config to Kconfig")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Mario Six <mario.six@gdsys.cc>
This pull request adds support for the last CPU board from
CS GROUP France (previously CSSI).
That CPU board called CMPCPRO has a mpc8321E CPU (Family PQII PRO hence
its name) and can be plugged in place of the CMPC885 board.
In order to support that new board, the following changes are included
in this series:
- Make the mpc8xx watchdog driver more generic for reusing it
with mpc83xx
- Fix various small problems on mpc83xx platform
- Add a GPIO Driver for QE GPIOs
- Add support for mpc832x into mpc83xx SPI driver
- Refactor existing board code that will be shared with new board
- Add the new board
CSSI has another CPU board, similar to the CMPC885 board
that get plugged on the two base boards MCR3000_2G and MIAE.
That CPU board is called CMPCPRO because it has a MPC8321E CPU,
also known as Power QUICC II PRO.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
On mpc832x, SPI can be either handled by CPU or QE.
In order to work in CPU mode, bit 17 of SPMODE has to
be set to 1, that bit is called OP.
Also, data is located at a different place than the one expected
by the driver today. In 8 bits mode with REV set, data to be
transmitted is located in the most significant byte while
received data is located in second byte. So perform the
necessary shifts.
In order to differentiate with other CPUs, a new compatible is
added for mpc832x: fsl,mpc832x-spi
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
The mpc832x has GPIOs handled by the QUICC Engine.
The registers are different from the one for the
non QE mpc83xx GPIOs.
Implement a GPIO driver for those.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
At startup, some RAM is needed (for instance for stack) before
DRAM is initialised.
One way to offer such RAM, used by mpc83xx, is to lock some entries
in the cache. To do that, MMU needs to be activated.
On mpc83xx having a QUICC Engine an alternative is to user some
part of from the Multi User RAM, like done on mpc8xx for instance.
For that, the MMU is not needed.
Activating the MMU is problematic because exception vectors are not
setup yet so in case of ISI or DSI that CPU will crash and reboot.
At the time being, MMU is activated regardless.
Only activate it when locking cache entries to provide initial RAM.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
There are helpers included in soc.h
Declare them static inline so that soc.h can be
included in several places.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
According to the reference manual, the Reset Configuration
Word Low Register bits 2-3 must be set to 0b10.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Introduce a new compatible "fsl,pq2pro-wdt"
On mpc83xx, the prescaling factor is 0x10000.
Don't write the watchdog configuration register in
start.S as it can be written only once.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
mpc8xx, mpc83xx and mpc86xx have similar watchdog with almost same
memory registers.
Refactor the driver to get the register addresses from the
device tree and use the compatible to know the prescale factor.
Calculate the watchdog setup value from the provided timeout.
Don't declare it anymore as an HW_WATCHDOG, u-boot will start
servicing the watchdog early enough.
On mpc8xx the watchdog configuration register is also used for
configuring the bus monitor. So add it as an option to the watchdog
when it is mpc8xx. When watchdog is not selected, leave the
configuration of the initial SYPCR from Kconfig.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
mpc8xx, mpc83xx and mpc86xx have similar watchdog with almost same
memory registers.
Rename it mpc8xxx which is the generic name used for drivers supporting
several mpc families.
The driver will be made more generic in following patch.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
8xx has CONFIG_8xx_GCLK_FREQ which is similar to
CONFIG_SYS_CLK_FREQ, and doesn't set CONFIG_SYS_CLK_FREQ.
Due to that, get_board_sys_clk() returns 0.
Remove CONFIG_8xx_GCLK_FREQ and use CONFIG_SYS_CLK_FREQ instead.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Last use of CONFIG_SYS_GPIO1_PRELIM was removed by
commit fae2ea5951 ("ppc: Remove MPC8349EMDS board and ARCH_MPC8349
support").
Last use of CONFIG_SYS_GPIO2_PRELIM was removed even before by
commit 6843862342 ("ppc: Remove caddy2 / vme8349 boards")
Those two items were removed from whitelist by
commit 8cca60a2cb ("Kconfig: Remove some symbols from the whitelist")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Fixes: fae2ea5951 ("ppc: Remove MPC8349EMDS board and ARCH_MPC8349 support")
Last (incorrect) use of those CONFIG items was removed by
commit 9fd9abedcc ("TQM834x: remove defines causing gcc4.4 warnings")
Those items are invalid and should have been removed at the
same time because lblaw[] has only 4 elements.
And they were removed from the whitelist by
commit 9c5df7a2a9 ("mpc83xx: Migrate LBLAW_* to Kconfig")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Fixes: 9fd9abedcc ("TQM834x: remove defines causing gcc4.4 warnings")
Commit fe7d654d04 ("mpc83xx: Migrate CONFIG_SYS_{BR, OR}*_PRELIM to
Kconfig") converted CONFIG_SYS_{BRx/ORx}_PRELIM to Kconfig by
implementing a fine-grained selection of every bit in Kconfig.
But commit c7fad78ec0 ("Convert CONFIG_SYS_BR0_PRELIM et al to
Kconfig") reworked it so that you now just have to provide the raw
value of each register in Kconfig. However, all fine-grained
Kconfig items remained allthough they are not used anymore.
Remove them all.
Fixes: c7fad78ec0 ("Convert CONFIG_SYS_BR0_PRELIM et al to Kconfig")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
- A fix for a long standing bug that has been exposed by commit
50128aeb0f ("cyclic: get rid of cyclic_init()") preventing 8xx boards
from booting since u-boot 2023.01
- A GPIO driver for powerpc 8xx chip
- Fixup for powerpc 8xx SPI driver
- A new powerpc 8xx board
- The two devices having that board.
This adds a new board from CS GROUP. The board is called
MCR3000_2G, and has a CPU board called CMPC885.
That CPU board is shared with another equipment that will
be added in a later patch.
That board stores Ethernet MAC addresses in an EEPROM which
is accessed using SPI bus.
This patch was originally written by Charles Frey who's
email address is not valid anymore as he left the company.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: FRANJOU Stephane <stephane.franjou@csgroup.eu>
Ports A, C and D are 16 bits ports.
Ports B and E are 32 bits ports.
The "compatible" is used to determine each port type.
This patch was originally written by Charles Frey who's
email address is not valid anymore as he left the company.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: FRANJOU Stephane <stephane.franjou@csgroup.eu>
Copied from e83a7e9453 ("powerpc/mpc83xx: Zero boot_flags arg for
calling board_init_f()")
The argument boot_flags of board_init_f() is not used at all in the
powerpc specific board.c init sequence. Now with the generic init
sequence, this boot_flags arg is used by board_init_f().
This patch sets the r3 register that is used to pass the boot_flags
argument from the start.S board_init_f() call to 0 prior to the function
call to avoid unknown content to end up in gd->flags.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Fixes: 09f3ca3dd5 ("arm, powerpc: select SYS_GENERIC_BOARD")
convert socrates board to use MPC85XX_HAVE_RESET_VECTOR and
disable CONFIG_OF_BOARD and use common u-boot.dtsi for
creating u-boot-dtb.bin.
Signed-off-by: Heiko Schocher <hs@denx.de>
U-Boot build process for socrates board produces final U-Boot binary in
file u-boot-socrates.bin (by binman) And as a bonus it produces two
unusable broken binaries u-boot-dtb.bin and u-boot.bin (by Makefile).
Clean this up, so final U-Boot binary is in u-boot-dtb.bin
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Heiko Schocher <hs@denx.de>
At this point, the remaining places where we have a symbol that is
defined as CONFIG_... are in fairly odd locations. While as much dead
code has been removed as possible, some of these locations are simply
less obvious at first. In other cases, this code is used, but was
defined in such a way as to have been missed by earlier checks. Perform
a rename of all such remaining symbols to be CFG_... rather than
CONFIG_...
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
As this is used on both ARM and PowerPC platforms, this needs to be
listed in arch/Kconfig.nxp and match how they're currently used by
select'ing them under the required PowerPC ARCH_xxx options.
Signed-off-by: Tom Rini <trini@konsulko.com>
As this is used on both ARM and PowerPC platforms, this needs to be
asked in arch/Kconfig.nxp. Set the PowerPC defaults based on
arch/powerpc/include/asm/config_mpc85xx.h and remove the default set in
drivers/mtd/nand/raw/fsl_ifc_nand.c
Signed-off-by: Tom Rini <trini@konsulko.com>
Remove some CONFIG symbols and related comments, etc, that are unused
within the code itself at this point.
Signed-off-by: Tom Rini <trini@konsulko.com>
