Chris Packham [Mon, 3 Dec 2018 01:26:49 +0000 (14:26 +1300)]
ARM: mvebu: a38x: sync ddr training code with mv_ddr-armada-18.09.02
This syncs drivers/ddr/marvell/a38x/ with the mv_ddr-armada-18.09 branch
of https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell.git.
Specifically this syncs with commit 99d772547314 ("Bump mv_ddr to
release armada-18.09.2").
The complete log of changes is best obtained from the mv-ddr-marvell.git
repository but some relevant highlights are:
The upstream code is incorporated omitting the portions not relevant to
Armada-38x and DDR3. After that a semi-automated step is used to drop
unused features with unifdef
Felix Brack [Wed, 5 Dec 2018 13:53:42 +0000 (14:53 +0100)]
arm: dts: am33xx: Sync dts with Linux 4.20.0
This patch synchronizes the am33xx SoC specific files with those from
Linux 4.20.0. Hence all board maintainers of am33xx based boards are
on the cc list.
The main purpose of this patch is to prevent further diverging of the
dts files from U-Boot and those from Linux. It aims to set the stage
for the synchronization of board specific dts files. Example: I'm the
maintainer of the PDU001 board: once this patch is applied successfully
I will make changes to the board specific dts file in Linux only and
then post a patch with a copy of this exact dts file to U-Boot. This
will make U-Boot and Linux remain in sync.
The stumbling block of https://patchwork.ozlabs.org/patch/943627 was
removed by the patch https://patchwork.ozlabs.org/patch/962428 from
Lokesh Vutla (many thanks!). This omap-serial driver allows using the
Linux am33xx.dtsi file in U-Boot.
Other changes to dts and dtsi files made by this patch are mainly to
prevent _new_ warnings during the build process. Especially the warning
at pinmux@800 stating 'unnecessary #address-cells/#size-cells without
"ranges" or child "reg"' was not removed. This warning is a good example
showing the benefit of the synchronization: if it needs to be fixed it
will be fixed in Linux and ported back to U-Boot.
Buildman reports all 46 am33xx SoC based boards to build fine, with
warnings of course. Nevertheless this patch should be tested thoroughly
on as many boards as possible to prevent any collateral damage.
Heiko Schocher [Wed, 5 Dec 2018 10:29:54 +0000 (11:29 +0100)]
spl/tpl: change banner into upper case
commit d6330064634a ("spl: Add a define for SPL_TPL_PROMPT")
changes the SPL/TPL banner from upper case into lower
case. As SPL and TPL are three-letter acronyms and they
are written in upper case, change it back to upper case.
Patrice Chotard [Mon, 3 Dec 2018 09:52:51 +0000 (10:52 +0100)]
gpio: stm32f7: Add gpio bank holes management
In some STM32 SoC packages, GPIO bank has not always 16 gpios.
Several cases can occur, gpio hole can be located at the beginning,
middle or end of the gpio bank or a combination of these 3
configurations.
For that, gpio bindings offer the gpio-ranges DT property which
described the gpio bank mapping.
Felix Brack [Fri, 30 Nov 2018 09:23:36 +0000 (10:23 +0100)]
arm: am335x-pdu001: Enable CONFIG_BLK and CONFIG_DM_MMC
This patch enables CONFIG_BLK as well as CONFIG_DM_MMC for the PDU001
board. It depends on Patrice Chotard's patch 'power: regulator: denied
disable on always-on regulator' which prevents power cycling the vmmc
supply. Without this patch the board will not boot as vmmc is
unfortunately used by other board components, not just eMMC and micro SD
card. Furthermore my patch 'dts: am335x-pdu001: Fix polarity of card
detection input' is required to boot from external micro SD card. Without
this patch no SD card will be detected and hence booting will fail.
Felix Brack [Thu, 29 Nov 2018 12:45:06 +0000 (13:45 +0100)]
dts: am335x-pdu001: Fix polarity of card detection input
When a micro SD card is inserted in the PDU001 card cage, the card
detection switch is opened and the corresponding GPIO input is driven
by a pull-up. Hence change the active level of the card detection
input from low to high.
ARM: at91: lds: add test for SPL binary size and bss size
Add test for the SPL binary size and the bss section size.
This will throw an error at build time if the SPL sections
do not fit in the designated RAM area, thus avoiding oversizing the SPL.
Protect configuration registers with a hardware spinlock.
If a hwspinlock is defined in the device-tree node used it
to be sure that none of the others processors on the SoC could
change the configuration at the same time.
Keerthy [Tue, 27 Nov 2018 12:22:41 +0000 (17:52 +0530)]
board: ti: ks2_evm: Over ride spl_get_load_buffer function
Currently k2 spi boot is broken as the image header
is getting copied to an invalid memory location
CONFIG_SYS_TEXT_BASE - sizeof (struct image_size)
which maps to 0xc000000 - 0x40 = 0xbffffc0 being a reserved
location.
We cannot change the CONFIG_SYS_TEXT_BASE address as the single
stage boots like UART boot will need the address to be 0xc000000
hence override the spl_get_load_buffer to have image_header
address as CONFIG_SYS_TEXT_BASE aka 0xc000000
Philipp Tomsich [Mon, 26 Nov 2018 19:20:19 +0000 (20:20 +0100)]
clk: Allow clock defaults to be set during re-reloc state for SPL only
In commit e5e06b65ad65 ("clk: Allow clock defaults to be set also
during re-reloc state") the earlier guard against setting clock
defaults in pre-reloc state was removed. While it is easy to filter
'assigned-clocks' properties for SPL using CONFIG_OF_SPL_REMOVE_PROPS,
no such mechanism exists for the pre-reloc stage of the full U-Boot.
With the default defconfig for the RK3399-Q7 (which filter the
'assigned-clocks' property for the DTS used by SPL anyway), this
caused a pause during startup of the full U-Boot stage that lasted for
almost 10s (due to the CPU not having been clocked up yet).
This reintroduces the guard from commit f4fcba5c5baa ("clk: Allow
clock defaults to be set also during re-reloc state") and extends it
to only apply outside of a TPL/SPL build: i.e. clk_set_defaults will
now run in pre-reloc state for SPL, but only after reloc for the full
U-Boot.
References: commit f4fcba5c5baa ("clk: implement clk_set_defaults()")
References: commit e5e06b65ad65 ("clk: Allow clock defaults to be set
also during re-reloc state") Signed-off-by: Philipp Tomsich <[email protected]>
Patrick Wildt [Mon, 26 Nov 2018 14:58:13 +0000 (15:58 +0100)]
fs: fix FAT name extraction
The long name apparently can be accumulated using multiple
13-byte slots. Unfortunately we never checked how many we
can actually fit in the buffer we are reading to.
Patrick Wildt [Mon, 26 Nov 2018 14:56:57 +0000 (15:56 +0100)]
fs: check FAT cluster size
The cluster size specifies how many sectors make up a cluster. A
cluster size of zero makes no sense, as it would mean that the
cluster is made up of no sectors. This will later lead into a
division by zero in sect_to_clust(), so better take care of that
early.
The MAX_CLUSTSIZE define can reduced using a define to make some
room in low-memory system. Unfortunately if the code reads a
filesystem with a bigger cluster size it will overflow the buffer.
Tom Rini [Thu, 6 Dec 2018 15:15:08 +0000 (10:15 -0500)]
Merge tag 'for-master-20181206' of git://git.denx.de/u-boot-rockchip
- Changes the declaration of regs_phy in dwc2-otg to uintptr_t
to ensure it can be cast to void* for use with writel().
- Add the Rock960 and Ficus boards.
rockchip: rk3399: Add common Rock960 family from Vamrs
Rock960 is a family of boards based on Rockchip RK3399 SoC from Vamrs.
It consists of Rock960 (Consumer Edition) and Ficus (Enterprise Edition)
96Boards.
Below are some of the key differences between both Rock960 and Ficus
boards:
1. Different host enable GPIO for USB
2. Different power and reset GPIO for PCI-E
3. No Ethernet port on Rock960
The common board support will be utilized by both boards. The device
tree has been organized in such a way that only the properties which
differ between both boards are placed in the board specific dts and
the reset of the nodes are placed in common dtsi file.
Philipp Tomsich [Thu, 6 Dec 2018 00:26:39 +0000 (01:26 +0100)]
usb: dwc2-otg: make regs_phy (in platdata) a uintptr_t
The regs_phy field of the platform data structure for dwc2-otg is
today declared an unsigned int, but will eventually be cast into a
void* for a writel operation. This triggers errors on modern GCC
versions.
E.g. we get the following error with GCC 6.3:
drivers/usb/phy/rockchip_usb2_phy.c: In function 'property_enable':
arch/arm/include/asm/io.h:49:29: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
#define __arch_putl(v,a) (*(volatile unsigned int *)(a) = (v))
^
arch/arm/include/asm/io.h:117:48: note: in expansion of macro '__arch_putl'
#define writel(v,c) ({ u32 __v = v; __iowmb(); __arch_putl(__v,c); __v; })
^~~~~~~~~~~
drivers/usb/phy/rockchip_usb2_phy.c:61:2: note: in expansion of macro 'writel'
writel(val, pdata->regs_phy + reg->offset);
^~~~~~
This commit changes regs_phy to be a uintptr_t to ensure that it is
large enough to hold any valid pointer (and fix the associated
warning).
Boris Brezillon [Sun, 2 Dec 2018 09:54:32 +0000 (10:54 +0100)]
mtd: sf: Make sf_mtd.c more robust
SPI flash based MTD devs can be registered/unregistered at any time
through the sf probe command or the spi_flash_free() function.
This commit does not try to fix the root cause as it would probably
require rewriting most of the code and have an mtd_info object
instance per spi_flash object (not to mention that the the spi-flash
layer is likely to be replaced by a spi-nor layer ported from Linux).
Instead, we try to be as safe as can be by checking the code returned
by del_mtd_device() and complain loudly when there's nothing we can
do about the deregistration failure. When that happens we also reset
sf_mtd_info.priv to NULL, and check for NULL pointer in the mtd hooks
so that -ENODEV is returned instead of hitting a NULL pointer
dereference exception when the MTD instance is later accessed by a user.
Boris Brezillon [Sun, 2 Dec 2018 09:54:31 +0000 (10:54 +0100)]
mtd: sf: Unregister the MTD device prior to removing the spi_flash obj
The DM implementation of spi_flash_free() does not unregister the MTD
device before removing the spi dev object. This leads to a use-after-free
bug when the MTD device is later accessed by a MTD user (observed when
attaching the device to UBI after env_sf_load() has called
spi_flash_free()).
Implement ->remove() and call spi_flash_mtd_unregister() from there.
Boris Brezillon [Sun, 2 Dec 2018 09:54:30 +0000 (10:54 +0100)]
mtd: Don't stop MTD partition creation when it fails on one device
MTD partition creation code is a bit tricky. It tries to figure out
when things have changed (either MTD dev list or mtdparts/mtdids vars)
and when that happens it first deletes all the partitions that had been
previously created and then creates the new ones based on the new
mtdparts/mtdids values.
But before deleting the old partitions, it ensures that none of the
currently registered parts are being used and bails out when that's
not the case. So, we end up in a situation where, if at least one MTD
dev has one of its partitions used by someone (UBI for instance), the
partitions update logic no longer works for other devs.
Rework the code to relax the logic and allow updates of MTD parts on
devices that are not being used (we still refuse to updates parts on
devices who have at least one of their partitions used by someone).
Boris Brezillon [Sun, 2 Dec 2018 09:54:29 +0000 (10:54 +0100)]
mtd: Make sure we don't parse MTD partitions belonging to another dev
The mtdparts variable might contain partition definitions for several
MTD devices. Each partition layout is separated by a ';', so let's
make sure we don't pick a wrong name when mtdparts is malformed.
Boris Brezillon [Sun, 2 Dec 2018 09:54:26 +0000 (10:54 +0100)]
mtd: Use get_mtdids() instead of env_get("mtdids") in mtd_search_alternate_name()
The environment is not guaranteed to contain a valid mtdids variable
when called from mtd_search_alternate_name(). Call get_mtdids() instead
of env_get("mtdids").
Fixes: ff4afa8a981e ("mtd: uboot: search for an equivalent MTD name with the mtdids") Signed-off-by: Boris Brezillon <[email protected]> Reviewed-by: Miquel Raynal <[email protected]> Tested-by: Heiko Schocher <[email protected]>
Boris Brezillon [Sun, 2 Dec 2018 09:54:25 +0000 (10:54 +0100)]
mtd: sf: Make sure we don't register the same device twice
spi_flash_mtd_register() can be called several times and each time it
will register the same mtd_info instance like if it was a new one.
The MTD ID allocation gets crazy when that happens, so let's track the
status of the sf_mtd_info object to avoid that.
Boris Brezillon [Sun, 2 Dec 2018 09:54:22 +0000 (10:54 +0100)]
mtd: Add a function to report when the MTD dev list has been updated
We need to parse mtdparts/mtids again everytime a device has been
added/removed from the MTD list, but there's currently no way to know
when such an update has been done.
Add an ->updated field to the idr struct that we set to true every time
a device is added/removed and expose a function returning the value
of this field and resetting it to false.
Andy Shevchenko [Tue, 20 Nov 2018 21:52:33 +0000 (23:52 +0200)]
dm: serial: Introduce ->getinfo() callback
New callback will give a necessary information to fill up ACPI SPCR table,
for example. Maybe used later for other purposes.
Signed-off-by: Andy Shevchenko <[email protected]> Reviewed-by: Simon Glass <[email protected]>
Change ADR_SPACE_SYSTEM_IO to SERIAL_ADDRESS_SPACE_IO to fix build error: Signed-off-by: Simon Glass <[email protected]>
Andy Shevchenko [Tue, 20 Nov 2018 21:52:32 +0000 (23:52 +0200)]
dm: serial: Add ->getconfig() callback
In some cases it would be good to know the settings, such as parity,
of current serial console. One example might be an ACPI SPCR table
to generate using these parameters.
dm: core: add functions to get/remap I/O addresses by name
This functions allow us to get and remap I/O addresses by name, which is useful when there are multiple reg addresses indexed by reg-names property.
This is needed in bmips dma/eth patch series, but can also be used on many
other drivers.
Philipp Tomsich [Sun, 25 Nov 2018 18:38:54 +0000 (19:38 +0100)]
dm: (re)sort uclass ids alphabetically
The comment in uclass-id.h states that
"U-Boot uclasses start here - in alphabetical order"
but the subsequent list is not sorted alphabetically.
This reestablishes order.
Notice that the first pattern has dashes throughout, while the second has
dashes throughout except just before the target architecture which has an
underscore.
The "dash throughout" versions from kernel.org are:
8.1.0, 6.4.0, 5.5.0, 4.9.4, 4.8.5, 4.6.1
while the "dash and underscore" versions from kernel.org are:
This tweak allows the code to handle both versions. Note that this tweak also
causes the architecture parsing to get confused and find the following two
bogus architectures, "2.0" and "64", which are explicitly checked for, and
removed.
Trevor Woerner [Wed, 21 Nov 2018 08:31:12 +0000 (03:31 -0500)]
buildman/toolchain.py: fix toolchain directory
The hexagon toolchain (4.6.1) from kernel.org, for example, was packaged in
a way that is different from most toolchains. The first entry when unpacking
most toolchain tarballs is:
gcc-<version>-nolib/<targetarch>-<system>
e.g.:
gcc-8.1.0-nolibc/aarch64-linux/
The first entry of the hexagon toolchain, however, is:
gcc-4.6.1-nolibc/
This causes the buildman logic in toolchain.py::ScanPath() to not be able to
find the "*gcc" executable since it looks in gcc-4.6.1-nolib/{.|bin|usr/bin}
instead of gcc-4.6.1/hexagon-linux/{.|bin|usr/bin}. Therefore when buildman
tries to download a set of toolchains that includes hexagon, the script fails.
This update takes the second line of the tarball unpacking (which works for
all the toolchains I've tested from kernel.org) and parses it to take the
first two elements, separated by '/'. It makes this logic a bit more robust.
Simon Glass [Sat, 24 Nov 2018 04:29:43 +0000 (21:29 -0700)]
Add inttypes.h
Even if U-Boot does not use this, some libraries do. Add back this header
file so that the build does not fall back to using the host version, which
may include stdint.h and break the build due to conflicts with uint64_t,
etc.
This partially reverts commit dee37fc99d94 ("Remove <inttypes.h> includes
and PRI* usages in printf() entirely")
The only change from the file that was in U-Boot until recently is that it
now comes twice as close to passing checkpatch. The remaining warnings
pertain to the typedefs, which checkpatch does not like.
Simon Glass [Sat, 24 Nov 2018 04:29:42 +0000 (21:29 -0700)]
Add UINT32_MAX and UINT64_MAX
These constants are defined by stdint.h but not by kernel.h, which is
its stand-in in U-Boot. Add the definitions so that libraries which expect
stdint.h constants can work.
Simon Glass [Sat, 24 Nov 2018 04:29:40 +0000 (21:29 -0700)]
time: Update mdelay() to delay in one large chunk
The current function delays in one millisecond at a time. This does not
work well on sandbox since it results in lots of calls to usleep(1000) in
a tight loop. This makes the sleep duration quite variable since each call
results in a sleep of *at least* 1000us, but possibly more. Depending on
how busy the machine is, the sleep time can change quite a bit.
We cannot fix this in general, but we can reduce the effect by doing a
single sleep. The multiplication works fine with an unsigned long argument
up until a sleep time of about 4m milliseconds. This is over an hour and
we can be sure that delays of that length are not useful.
Update the mdelay() function to call udelay() only once with the
calculated delay value.
Simon Glass [Sat, 24 Nov 2018 04:29:38 +0000 (21:29 -0700)]
input: i8042: Use remove() instead of exported functions
We should not have exported functions in a driver. The i8042_disable()
function is used to disable the keyboard. Provide a remove() method
instead, which is the standard way of disabling a device.
We could potentially add a method to flush input but that does not seem
necessary.
Simon Glass [Sat, 24 Nov 2018 04:29:37 +0000 (21:29 -0700)]
cros_ec: Adjust to use v1 vboot context only
At present there are no users of the 64-byte v2 context. The v1 context is
only 16 bytes long and currently an error is raised if too much data is
returned from the EC.
Simon Glass [Sat, 24 Nov 2018 04:29:32 +0000 (21:29 -0700)]
tpm: Remove use of build-time TPM versions
There is only one place in the code which assumes at build-time that we
are using either a v1 or a v2 TPM. Fix this up and add a new function to
return the version of a TPM.
Supported TPM versions (v1 and v2) can be enabled independently and it is
possible to use both versions at once. This is useful for sandbox when
running tests.
Simon Glass [Sat, 24 Nov 2018 04:29:30 +0000 (21:29 -0700)]
sandbox: Use 'extras' to specify 'head' files
At present sandbox has a start.o in the 'start' target but also includes
it in the normal target list. This is not how this is normally handled. It
is needed because sandbox does not include the u-boot-init variable in its
link rule.
Update the rule and move start.o from the normal target list to the
'extras' list.
Simon Glass [Sat, 24 Nov 2018 04:29:29 +0000 (21:29 -0700)]
sandbox: Zero the ram buffer on startup
At present the RAM buffer is not inited unless it is read from a file,
likely produced by an earlier phase of U-Boot. This causes valgrind
warnings whenever the RAM buffer is used. Correct this by initing it if
needed.
Simon Glass [Sat, 24 Nov 2018 04:29:28 +0000 (21:29 -0700)]
sandbox: sysreset: Update to support power-on reset
If U-Boot is started from SPL or TPL, then those earlier phases deal with
the reset cause. On real hardware this cause may be lost once it is read.
Emulate that behaviour in sandbox by reporting a warm reset when a
previous phase has run since start-up.
Simon Glass [Sat, 24 Nov 2018 04:29:27 +0000 (21:29 -0700)]
sandbox: net: Correct name copy in eth_raw_bus_post_bind()
We cannot be sure that the interface name takes up the full length of the
space available to it. Use strcpy() instead of memcpy() in this case. This
corrects a valgrind warning.
Rick Chen [Mon, 3 Dec 2018 09:48:20 +0000 (17:48 +0800)]
riscv: ax25-ae350: Pass dtb address to u-boot with a1 register
ax25-ae350 use CONFIG_OF_BOARD via a2 and CONFIG_SYS_SDRAM_BASE
to boot from ram which allow the board to override the fdt
address originally.
But after this patch
riscv: save hart ID and device tree passed by prior boot stage
It provide prior_stage_fdt_address which offer a temporary
memory address to keep the dtb address passing from loader(gdb)
to u-boot with a1.
So passing via a2 and CONFIG_SYS_SDRAM_BASE is redundant and
can be removed. And it also somehow may corrupted BBL if it
was be arranged in CONFIG_SYS_SDRAM_BASE.
In board_fdt_blob_setup()
When boting from ram:
prior_stage_fdt_address will be use to reserved dtb temporarily.
When booting from ROM:
dtb will be pre-burned in CONFIG_SYS_FDT_BASE, if it is flash base.
Or CONFIG_SYS_FDT_BASE maybe a memory map space (NOT RAM or ROM)
which is provided by HW.
Anup Patel [Mon, 3 Dec 2018 05:27:41 +0000 (10:57 +0530)]
riscv: qemu: Use different SYS_TEXT_BASE for S-mode
When u-boot runs in S-mode, the M-mode runtime firmware
(BBL or equivalent) uses memory range in 0x80000000 to
0x80200000. Due to this, we cannot use 0x80000000 as
SYS_TEXT_BASE when running in S-mode. Instead for S-mode,
we use 0x80200000 as SYS_TEXT_BASE.
Even Linux RISC-V kernel ignores/reserves memory range
0x80000000 to 0x80200000 because it runs in S-mode.
Anup Patel [Mon, 3 Dec 2018 05:27:40 +0000 (10:57 +0530)]
riscv: Add kconfig option to run U-Boot in S-mode
This patch adds kconfig option RISCV_SMODE to run U-Boot in
S-mode. When this opition is enabled we use s<xyz> CSRs instead
of m<xyz> CSRs.
It is important to note that there is no equivalent S-mode CSR
for misa and mhartid CSRs so we expect M-mode runtime firmware
(BBL or equivalent) to emulate misa and mhartid CSR read.
In-future, we will have more patches to avoid accessing misa and
mhartid CSRs from S-mode.
projects / J-u-boot.git / log