The fence instruction is used to enforce device I/O and memory ordering
constraints in RISC-V. It can not be relied on to directly affect the
data cache on every CPU.
Andes' AX25 does not have a coherence agent. Its fence instruction
flushes the data cache and is used to keep data in the system coherent.
The implementation of flush_dcache_all in lib/cache.c is therefore
specific to the AX25. Move it into the AX25-specific cache.c in
cpu/ax25/.
This also adds a missing new line between flush_dcache_all and
flush_dcache_range in lib/cache.c.
Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Undefined exceptions are treated as reserved. This is not clearly
communicated to the user. Adjust the error message to clarify that a
reserved exception has occurred and add additional details.
Fixes: e8b522b ("riscv: treat undefined exception codes as reserved")
Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
The following hang is observed on a Hummingboard 2 MicroSOM
i2eX iMX6D - rev 1.3 with no eMMC populated on board:
U-Boot SPL 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)
Trying to boot from MMC1
U-Boot 2018.11+gf6206f8587 (Nov 16 2018 - 00:56:34 +0000)
CPU: Freescale i.MX6D rev1.5 996 MHz (running at 792 MHz)
CPU: Extended Commercial temperature grade (-20C to 105C) at 33C
Reset cause: POR
Board: MX6 HummingBoard2
DRAM: 1 GiB
MMC: FSL_SDHC: 0, FSL_SDHC: 1
Loading Environment from MMC... *** Warning - bad CRC, using default environment
No panel detected: default to HDMI
Display: HDMI (1024x768)
In: serial
Out: serial
Err: serial
---> hangs
which is caused by the following infinite loop inside esdhc_send_cmd_common()
while (!(esdhc_read32(®s->irqstat) & flags))
;
Instead of looping forever, provide an exit path so that a timeout
error can be propagated in the case irqstat does not report
any interrupts, which may happen when no eMMC is populated on
board.
Reported-by: Ricardo Salveti <rsalveti@rsalveti.net>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
Tested-by: Ricardo Salveti <rsalveti@rsalveti.net>
Chris Spencer reports that when enabling ext4 read support without
also enabling write support the following error is seen:
fs/fs.c:198:12: error: 'ext4_write_file' undeclared here (not in a
function); did you mean 'ext4_read_file'?
.write = ext4_write_file,
^~~~~~~~~~~~~~~
Fix this problem by moving these options to Kconfig.
Reported-by: Chris Spencer <spencercw@gmail.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Change CONFIG_ENV_OFFSET for Poplar to be 1MB further into the eMMC
than before. This puts it immediately prior to the space we are
reserving offset 0x200000-0x400000 for UEFI to save its persistent
data. Define CONFIG_ENV_SIZE as a product of env_mmc_nblks and the
sector size, like CONFIG_ENV_OFFSET is.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
This particular commit is causing a regression on stih410-b2260 and
other platforms when reading from FAT16. Noting that I had rebased the
original fix from Thomas onto then-current master, there is also
question from Akashi-san if the change is still needed after other FAT
fixes that have gone in.
This reverts commit a68b0e11ea.
Reported-by: Patrice Chotard <patrice.chotard@st.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Thomas RIENOESSL <thomas.rienoessl@bachmann.info>
Signed-off-by: Tom Rini <trini@konsulko.com>
This reverts commit a7f4b4b344.
As reported by Alex Kiernan the above optimization introduces a
regression in the below use case where:
1. Device has defined 'u-boot,dm-spl' property (@ eMMC DTS node)
2. The device downloads its MLO/SPL via UART (not eMMC - the eMMC pinmux
pins are NOT probed/configured in MLO/SPL).
3. The loaded via UART MLO/SPL wants to load Linux from eMMC. In this case
the DM core and pinctrl uclass checks 'u-boot,dm-spl' and don't
configure pins (as it thinks that those were initialized in MLO/SPL).
As we are very close to release - please revert this commit.
Reported-by: Alex Kiernan <alex.kiernan@gmail.com>
Signed-off-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
The SDMMC reset is missing from DT, so the reset manager cannot unreset
the SDMMC. Add the missing DT reset entry.
Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
Fixes commit 0131162439 ("dm: usb: create a new UCLASS ID for USB gadget
devices")
The UCLASS_DRIVER for id UCLASS_USB_GADGET_GENERIC needs to be declared
even for platforms that do not enable DM_USB_GADGET. Otherwise the driver
for their usb peripheral controller fails to bind.
Reported-by: Priit Laes <plaes@plaes.org>
Reported-by: Jagan Teki <jagan@amarulasolutions.com>
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Tested-by: Priit Laes <plaes@plaes.org>
Acked-by: Jagan Teki <jagan@openedev.com>
There is no need to compile and include this code if it is not used.
CONFIG_USB_GADGET can be used for the purpose.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
dfu, fastbot and other usb gadget commands take the USB port index as a
parameter. Currently this index is assigned in the order of the driver
bindings.
Changing this behavior using the SEQ_ALIAS feature. This option assign to
the device a SEQ number based on its alias (if it exists)
To use it we must set the DM_UC_FLAG_SEQ_ALIAS flag and follow the existing
naming convention: use "usb" for the name of the gadget UCLASS_DRIVER
(same as for the UCLASS_USB).
If no alias is provided, then the index falls back to the order in which
the bindings took place.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reported-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
Tested-by: Sam Protsenko <semen.protsenko@linaro.org>
Existing clock configure code has been followed based on the
legacy MMC dt node definitions and it cannot work with recent
dts(i) sync from Linux.
So, add clock configure code for Allwinner platforms which support
DM_MMC and eventually this will drop once CLK support is in Mainline.
Fixes: 3c92cca3cd ("ARM: dts: sun4i: Update A10 dts(i) files from Linux-v4.18-rc3")
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Priit Laes <plaes@plaes.org> # Gemei G9 A10 Tablet
Tested-by: Marek Kraus <gamelasterv2@gmail.com> # A10-OLinuXino-Lime
OSC_27M_CLK should return 27MHz and OSC_32K_CLK should return
32768Hz to reflect the reality.
This also keeps the values in sync with the Linux clock tree.
Signed-off-by: Fabio Estevam <festevam@gmail.com>
To B0[+] chips, dcd_skip needs to be true. For A0 chip, it needs
to be false, however A0 chip is no longer being supported anymore.
Considering we are moving code from imx-mkimage to uboot mkimage,
to make sure we not introduce some surprise, we still keep dcd_skip
code there.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The CPU banner printed is as following:
CPU: CPU: Freescale i.MX8QXP RevB A35 at 147228 MHz
1. Drop the CPU:
2. Change vendor from Freescale to NXP
Signed-off-by: Peng Fan <peng.fan@nxp.com>
priv data has not been allocated when doing bind, so it is
wrong to use dev_get_priv in bind call back.
Let's switch to use platdata in the driver to fix the issue.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When boot_dev is true, fill boot device. However the original logic is when
boot_dev is false, fill boot device, this will trigger data abort.
Also fix sc_misc_get_control when using pointer val.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
When build clk driver in spl, met the warning:
"
drivers/clk/imx/clk-imx8.c:21:25: warning: ‘imx8_clk_names’ defined but not used [-Wunused-variable]
static struct imx8_clks imx8_clk_names[] = {
^~~~~~~~~~~~~~
"
Fix with wrapping the array with CONFIG_CMD_CLK.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The current defconfig build generates a u-boot.imx file that is too large
for the available space on a eMMC/SD card. Installing this file overwrites
the u-boot environment. So disable the unused FIT support to reduce the
size of the u-boot binary.
Signed-off-by: Soeren Moch <smoch@web.de>
Check the size of the generated u-boot.imx file. Report an error
if it would be too big and overwrite the u-boot environment.
Signed-off-by: Soeren Moch <smoch@web.de>
The mkimage_fit_atf.sh SPL FIT generator script requires
u-boot-nodtb.bin, but this was not enforced by the Makefile. This could
cause the generator script to be executed before u-boot-nodtb.bin has
been created.
Signed-off-by: Chris Spencer <christopher.spencer@sea.co.uk>
Cc: NXP i.MX U-Boot Team <uboot-imx@nxp.com>
The current value CTLUPD_AREF(0) is the reset value of the register,
so there is no need to write a value. If needed, the register can be
written using board specific CR settings.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
The CR80 register has multiple interrupt bits, the code is supposed
to check bit 8 but instead uses a logical and. In most cases this
probably did not affect real operations since at that stage typically
none of the other bits are set.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
According to the data sheet bits 10-8 are PHYDRAM_CK_EN. Fix mask
to allow setting PHYDRAM_CK_EN correctly.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Commit 3f353ceccb ("vf610: refactor DDRMC code") changed on-die
termination (ODT) values from 120 Ohm to 60 Ohm and enabled a static
read/write leveling which has not been tested with this board. This
commit reverts both changes and makes sure that memory gets
initialized as it has been done before the mentioned commit.
Fixes: 3f353ceccb ("vf610: refactor DDRMC code")
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
The Vybrid reference manual VFXXXRM Rev. 0 10/2016 states in chapter
5.2.6.1 DUMMY PADS (DDR/QuadSPI) that those pads need to be programed
for correct operation of DDR. Assume the default DDR pin configuration
which seems to work well on a Colibri VF50.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Safe some space by not selecting CMD_NFS and CMD_LOADS/B.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
After updating the value of dev_context_ptrs[0], we should flush this
from cache to memory. Otherwise the xhci controller won't use it.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Marek Vasut <marex@denx.de>
In order to keep SPL code size below the 32Kb limit,
put under CONFIG_SPL_BUILD flag all unused code in SPL.
This is needed for stm32f7xx board which are using SPL.
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
In case "gpio-ranges" property is not present in device tree,
use default value for gpio_count and gpio_range.
This fixes an issue on stm32 F7 and H7 boards where "pinmux status -a"
command didn't return any pin status due to the fact that both stm32 F7
and H7 board DT doesn't use the gpio-ranges property.
Fixes: dbf928dd2634a6("gpio: stm32f7: Add gpio bank holes management")
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Remove unused Android partitions:
- efs, crypto, cache: we don't use it anymore (images are not built
in AOSP
- ipu1, ipu2: IPU firmware is now a part of vendor image and doesn't
reside as a separate partition
While at it, rename "reserved" partition to "uboot-env", as it's
actually stores U-Boot environment.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Praneeth Bajjuri <praneeth@ti.com>
The env location label ENVL_ESATA is missed in location tables, so
when we configure the ENV in SATA, u-boot fails to get correct env
location and cause boot hang in board_f.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
When running usb dwc3 gadget driver, we meet random USB enumeration failure in fastboot.
The root cause is a cache coherence issue. When it happens, the ctrl_req in
gadget driver is allocated at 0xfe932f40, and the usb_composite_dev (cdev)
is allocated at 0xfe932f60. So after we submit the setup request (cache flushed) to USB
controller, any accessing to usb_composite_dev variable will cause the cache line refill, then
when setup transfer is completed, reading the setup data in ctrl_req will gets old value from
cache not from memory.
The ctrl_req is allocated by API dma_alloc_coherent, but u-boot don't have cohernet memory.
so it still needs cache maintain operations before/after HW accessing. Since the cache flush or
invalidate bases on cache line, so when the allocated memory size is not cache line aligned,
potentially it may meet such issue.
This patch modifies the dma_alloc_coherent API to round the size to cache line aligned.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
I can't continue maintaining the board because I don't have access to the
hardware anymore, so remove myself from the entry and add Javier who has
volunteered to help and maintain the board.
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Javier Martínez Canillas <javier@dowhile0.org>
On our mirror account on github we have gotten a handful of pull
requests. At this time github does not allow you to disable pull
requests on a project. The generally suggested work-around is to add a
pull request template that tells people to not use that workflow. Add
one here that points to the wiki page on submitted patches.
Signed-off-by: Tom Rini <trini@konsulko.com>
Commit d0851c8937 ("blk: Call part_init() in the post_probe() method")
removed the call to part_init() in mmc.c, as this is done by the DM_MMC
framework.
However Allwinner is (still) relying on a non-DM MMC driver, so we are
now missing the implicit partition init, leading to failing MMC accesses
due to the missing partition information.
Bring the call back just for non-DM MMC driver to fix this regression.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Soeren Moch <smoch@web.de>
