oweals/u-boot.git
5 years agopci: pci.h: add missing maskbit
Ramon Fried [Sat, 6 Apr 2019 02:12:01 +0000 (05:12 +0300)]
pci: pci.h: add missing maskbit

PCI_MSI_FLAGS_MASKBIT was missing from include file,
add it.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agofdtdec: Use fdt_setprop_u32() for fdtdec_set_phandle()
Thierry Reding [Mon, 15 Apr 2019 08:08:20 +0000 (10:08 +0200)]
fdtdec: Use fdt_setprop_u32() for fdtdec_set_phandle()

The fdt_setprop_u32() function does everything that we need, so we
really only use the function as a convenience wrapper, in which case it
can simply be a static inline function.

Signed-off-by: Thierry Reding <treding@nvidia.com>
5 years agodm: core: Change platform specific translation-offset handling
Stefan Roese [Fri, 12 Apr 2019 14:42:28 +0000 (16:42 +0200)]
dm: core: Change platform specific translation-offset handling

Testing has shown that the current DM implementation of a platform /
board specific translation offset, as its needed for the SPL on MVEBU
platforms is buggy. The translation offset is confingured too late,
after the driver bind functions are run. This may result in incorrect
address translations. With the current implementation its not possible
to configure the offset earlier, as the DM code has not run at all.

This patch now removed the set_/get_translation_offset() calls and
moves the translation offset into the GD variable translation_offset.
This variable will get used when CONFIG_TRANSLATION_OFFSET is enabled.
This option is enabled only for MVEBU on ARM32 platforms, where its
currenty needed and configured in the SPL.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Pierre Bourdon <delroth@gmail.com>
Cc: Baruch Siach <baruch@tkos.co.il>
Cc: Simon Glass <sjg@chromium.org>
Cc: Heiko Schocher <hs@denx.de>
Cc: Tom Rini <trini@konsulko.com>
Tested-by: Pierre Bourdon <delroth@gmail.com>
Tested-by: Baruch Siach <baruch@tkos.co.il>
5 years agosimple-bus: add DM_FLAG_PRE_RELOC flag to simple-bus driver
Lukas Auer [Wed, 10 Apr 2019 12:46:07 +0000 (14:46 +0200)]
simple-bus: add DM_FLAG_PRE_RELOC flag to simple-bus driver

Boards such as qemu-riscv, which receive their device tree at runtime,
for example from QEMU or firmware, are unable to add the appropriate
device tree properties to make devices available pre relocation.
Instead, they must rely on the DM_FLAG_PRE_RELOC flag to be set for the
required drivers.

Add the DM_FLAG_PRE_RELOC flag to the simple-bus driver to make devices
under it with drivers that have set the flag as well available pre
relocation for these boards.

Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
5 years agosandbox: Enable the 'trace' command when tracing is used
Simon Glass [Mon, 8 Apr 2019 19:20:53 +0000 (13:20 -0600)]
sandbox: Enable the 'trace' command when tracing is used

Enable this by default so that tracing can be inspected if enabled. This
cannot rely on the 'imply' in lib/Kconfig since this method of enabling
tracing relates on an environment variable (FTRACE) and does not use
Kconfig.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agoConvert CONFIG_TRACE_BUFFER_SIZE et al to Kconfig
Simon Glass [Mon, 8 Apr 2019 19:20:52 +0000 (13:20 -0600)]
Convert CONFIG_TRACE_BUFFER_SIZE et al to Kconfig

This converts the following to Kconfig:
   CONFIG_TRACE_BUFFER_SIZE
   CONFIG_TRACE_EARLY_SIZE
   CONFIG_TRACE_EARLY
   CONFIG_TRACE_EARLY_ADDR

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agoConvert CONFIG_TRACE to Kconfig
Simon Glass [Mon, 8 Apr 2019 19:20:51 +0000 (13:20 -0600)]
Convert CONFIG_TRACE to Kconfig

This converts the following to Kconfig:
   CONFIG_TRACE

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agotrace: Tidy up error returns
Simon Glass [Mon, 8 Apr 2019 19:20:50 +0000 (13:20 -0600)]
trace: Tidy up error returns

At present many functions in this file return -1. Update them to return a
valid error code. Also tidy up the 'return' statements at the same time,
since these should have a blank line before them.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agodiv64: Don't instrument the division function
Simon Glass [Mon, 8 Apr 2019 19:20:49 +0000 (13:20 -0600)]
div64: Don't instrument the division function

This function may be called from tracing code, since that code needs to
read the timer and this often requires calling do_div(), which calls
__div64_32(). If this function is instrumented it causes an infinite loop,
since emitting a trace record requests the time, which in turn emits a
trace record, etc.

Update the prototype to prevent instrumentation code being added.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agodiv64: Use kernel types
Simon Glass [Mon, 8 Apr 2019 19:20:48 +0000 (13:20 -0600)]
div64: Use kernel types

These functions still use uint32_t and uint64_t but checkpatch now
requests that the kernel types be used instead. Update them as well as a
few resulting checkpatch errors.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agoinitcall: Drop use of header files
Simon Glass [Mon, 8 Apr 2019 19:20:47 +0000 (13:20 -0600)]
initcall: Drop use of header files

This file should not include header files. They have already been included
by the time initcall.h is included. Also, document how to enable debugging
in this file.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agosandbox: Increase the early-trace-buffer size
Simon Glass [Mon, 8 Apr 2019 19:20:46 +0000 (13:20 -0600)]
sandbox: Increase the early-trace-buffer size

This buffer is too small now that sandbox has grown in size. Increase it.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agobootstage: Allow calling bootstage_mark() before bootstage_init()
Simon Glass [Mon, 8 Apr 2019 19:20:45 +0000 (13:20 -0600)]
bootstage: Allow calling bootstage_mark() before bootstage_init()

It is possible for this to happen if something goes wrong very early in
the init sequence. Add a check for this.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agosandbox: Move pre-console buffer out of the way of tracing
Simon Glass [Mon, 8 Apr 2019 19:20:44 +0000 (13:20 -0600)]
sandbox: Move pre-console buffer out of the way of tracing

These two buffers currently conflict if tracing is enabled. Move the
pre-console buffer and update the documentation.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agosandbox: Drop the printf() in setup_ram_buf()
Simon Glass [Mon, 8 Apr 2019 19:20:43 +0000 (13:20 -0600)]
sandbox: Drop the printf() in setup_ram_buf()

This was really intended for debugging. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agosandbox: Correct maths in allocation routines
Simon Glass [Mon, 8 Apr 2019 19:20:42 +0000 (13:20 -0600)]
sandbox: Correct maths in allocation routines

Allocation routines were adjusted to ensure that the returned addresses
are a multiple of the page size, but the header code was not updated to
take account of this. These routines assume that the header size is the
same as the page size which is unlikely.

At present os_realloc() does not work correctly due to this bug. The only
user is the hostfs 'ls' command, and only if the directory contains a
unusually long filename, which likely explains why this bug was not
caught earlier.

Fix this by doing the calculations using the obtained page size.

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agosandbox: Improve debugging in initcall_run_list()
Simon Glass [Mon, 8 Apr 2019 19:20:41 +0000 (13:20 -0600)]
sandbox: Improve debugging in initcall_run_list()

At present if one of the initcalls fails on sandbox the address printing
is not help, e.g.:

  initcall sequence 0000557678967c80 failed at call 00005576709dfe1f (err=-96)

This is because U-Boot gets relocated high into memory and the relocation
offset (gd->reloc_off) does not work correctly for sandbox.

Add support for finding the base address of the text region (at least on
Linux) and use that to set the relocation offset. This makes the output
better:

  initcall sequence 0000560775957c80 failed at call 0000000000048134 (err=-96)

Then you use can use grep to see which init call failed, e.g.:

   $ grep 0000000000048134 u-boot.map
   stdio_add_devices

Of course another option is to run it with a debugger such as gdb:

   $ gdb u-boot
   ...
   (gdb) br initcall.h:41
   Breakpoint 1 at 0x4db9d: initcall.h:41. (2 locations)

Note that two locations are reported, since this function is used in both
board_init_f() and board_init_r().

   (gdb) r
   Starting program: /tmp/b/sandbox/u-boot
   [Thread debugging using libthread_db enabled]
   Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

   U-Boot 2018.09-00264-ge0c2ba9814-dirty (Sep 22 2018 - 12:21:46 -0600)

   DRAM:  128 MiB
   MMC:

Breakpoint 1, initcall_run_list (init_sequence=0x5555559619e0 <init_sequence_f>)
    at /scratch/sglass/cosarm/src/third_party/u-boot/files/include/initcall.h:41
41 printf("initcall sequence %p failed at call %p (err=%d)\n",
   (gdb) print *init_fnc_ptr
   $1 = (const init_fnc_t) 0x55555559c114 <stdio_add_devices>
   (gdb)

Signed-off-by: Simon Glass <sjg@chromium.org>
5 years agocommon: command: Add command execution tracer.
Christoph Muellner [Fri, 5 Apr 2019 11:03:46 +0000 (13:03 +0200)]
common: command: Add command execution tracer.

When using boot scripts it can become quite hard to understand
which commands are actually executed during bootup (e.g. where
is a kernel image loaded from or which DTB is in use).

Shell scripts suffer from a similar problem and many shells address
this problem with a command execution tracer (e.g. BASH has xtrace,
which can be enabled by "set -x").

This patch introduces a command tracer for U-Boot, which prints
every command with its arguments before it is executed.

Signed-off-by: Christoph Muellner <christoph.muellner@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
5 years agofdt: boot_get_fdt: android: use ENV 'fdtaddr' as fallback
Eugeniu Rosca [Mon, 1 Apr 2019 10:52:52 +0000 (12:52 +0200)]
fdt: boot_get_fdt: android: use ENV 'fdtaddr' as fallback

Our platform doesn't store the DTB into the Android image second area,
but rather copies the DTB to RAM from a dedicated dtb.img partition [0],
prior to booting the Android image by calling bootm.

Similar to [1], we find it useful to just call 'bootm' and have the
right DTB being passed to OS (assuming its address has been previously
stored in 'fdtaddr' by calling `fdt addr <dtb-addr>`).

Booting Android with DTB from 'fdtaddr' will only occur if:
 - No DTB is embedded in the second area of Android image
 - 'fdtaddr' points to a valid DTB in RAM

[0] https://source.android.com/devices/architecture/dto/partitions
[1] https://patchwork.ozlabs.org/patch/1046652/
    ("Support boot Android image without address on bootm command")

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agofdt: boot_get_fdt: android: compress handling (non-functional)
Eugeniu Rosca [Mon, 1 Apr 2019 10:45:36 +0000 (12:45 +0200)]
fdt: boot_get_fdt: android: compress handling (non-functional)

Prepare for booting Android images which lack any DTB in the second
area by using 'fdtaddr' environment variable as source/address of FDT.
No functional/behavioral change expected in this patch.

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agofdt: boot_get_fdt: simplify no_fdt handling (non-functional)
Eugeniu Rosca [Mon, 1 Apr 2019 10:45:35 +0000 (12:45 +0200)]
fdt: boot_get_fdt: simplify no_fdt handling (non-functional)

Increase the readability of boot_get_fdt().
No change in behavior is expected.

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agofdt: boot_get_fdt: really boot w/o FDT when "goto no_fdt"
Eugeniu Rosca [Mon, 1 Apr 2019 10:45:34 +0000 (12:45 +0200)]
fdt: boot_get_fdt: really boot w/o FDT when "goto no_fdt"

The 'no_fdt' goto label was introduced by v2015.01 commit [0] and it
had two review stages [1-2]. The *documented* purpose behind commit [0]
is (excerpt from commit description):

> allows both FDT and non-FDT kernels to boot by making the
> third parameter to the bootm/bootz optional

While [1] and [2] share the same goal, they have very different
implementations:
 - [1] was based on a very simple 'argc' check at function error out
   with returning success to the caller if the third parameter was NOT
   passed to bootm/bootz command. This approach had the downside of
   returning success to the caller even in case of legitimate internal
   errors, which should halt booting.
 - [2] added the "no_fdt" label and several "goto no_fdt" statements.
   This allowed to report the legitimate internal errors to the caller.

IOW the major difference between [1] and [2] is:
 - [1] boot w/o FDT if FDT address is not passed to boot{m,z,*}
 - [2] give *freedom* to the developer to boot w/o FDT from any
   (more or less) arbitrary point in the function flow (and here
   comes the peculiar aspect, which looks to be a leftover from [1])
   with the precondition that the 3rd argument (FDT address) is NOT
   provided to boot{m,z,*}. In practice, this means that only a subset
   of "goto no_fdt" end up booting w/o FDT while the other subset is
   returning an error to the caller.

This patch removes the peculiar behavior described above, such that
"goto no_fdt" performs really what it tells to the developer.

The motivation of this patch is to decrease the unneeded complexity
and increase the readability of boot_get_fdt().

[0] 48aead71c1ad ("fdt: Allow non-FDT kernels to boot when CONFIG_OF_LIBFDT is defined")
[1] https://patchwork.ozlabs.org/patch/412923/
    ("[U-Boot,v1] fdt: Allow non-FDT kernels to boot when CONFIG_OF_LIBFDT is defined")
[2] https://patchwork.ozlabs.org/patch/415635/
    ("[U-Boot,v2] fdt: Allow non-FDT kernels to boot when CONFIG_OF_LIBFDT is defined")

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agofdt: boot_get_fdt: remove redundant zeroing out
Eugeniu Rosca [Mon, 1 Apr 2019 10:45:33 +0000 (12:45 +0200)]
fdt: boot_get_fdt: remove redundant zeroing out

Paranoid programming [1] lies at the foundation of proper software
development, but the repetitive zeroing-out of output arguments in the
context of the same function rather clutters the code and inhibits
further refactoring/optimization than is doing any good.

In boot_get_fdt(), we already perform zero/NULL-initialization of
*of_flat_tree and *of_size at the beginning of the function, so doing
the same at function error-out is redundant/superfluous.

Moreover, keeping the code unchanged might encourage the developers to
update *of_flat_tree and *of_size during some interim computations,
which is against the current design of boot_get_fdt(). Currently,
writing useful data into these arguments happens just before
successfully returning from boot_get_fdt() and it should better stay so.

[1] https://blog.regehr.org/archives/1106

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoMerge branch 'master' of git://git.denx.de/u-boot-usb
Tom Rini [Sun, 21 Apr 2019 23:00:04 +0000 (19:00 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-usb

- Various fastboot, dwc2/stm32 updates

5 years agoMerge branch 'master' of git://git.denx.de/u-boot-socfpga
Tom Rini [Sun, 21 Apr 2019 22:59:30 +0000 (18:59 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-socfpga

- Various stratix10, gen5 updates

5 years agousb: dwc2: fix gadget disconnect
Fabrice Gasnier [Wed, 17 Apr 2019 14:46:13 +0000 (16:46 +0200)]
usb: dwc2: fix gadget disconnect

This fixes a disconnect issue detected with fastboot command, when using
dwc2 driver.
- On u-boot side:
uboot>$ fastboot 0
- On USB host PC side, few seconds after
PC>$ fastboot reboot # Get stuck, uboot target never reboots

By enabling DEBUG_ISR logs, the bus suspend interrupt is seen before the
PC command has been issued. When the USB bus suspend occurs, there's a HACK
that disables the fastboot (composite driver). Here is the call stack
upon USB bus suspend:
- dwc2_handle_usb_suspend_intr()
  - dev->driver->disconnect()
    - composite_disconnect()
      - reset_config()
        - f->disable()
          - fastboot_disable()
            - usb_ep_disable(f_fb->out_ep);
            - usb_ep_disable(f_fb->in_ep);
            .. other disable calls.

When the resume interrupt happens, everything has been disabled, then
nothing happens. fastboot command gets stuck on HOST side.

Remove original HACK, that disconnects the composite driver upon
USB bus suspend. Implement disconnect detection instead:
- check GINTSTS OTG interrupt
- read GOTGINT register
- check GOTGINT, SesEndDet bit (e.g. session end)
This is inspired by what is implemented currently in Linux dwc2 driver.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@st.com>
Reviewed-by: Marek Vasut <marex@denx.de>
5 years agousb: reload watchdog during ums command
Patrick Delaunay [Thu, 18 Apr 2019 15:18:39 +0000 (17:18 +0200)]
usb: reload watchdog during ums command

Reload the watchdog in the mass storage command ums
to avoid reboot during the usb waiting loop
when the host doesn't send any request.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agofastboot: add support for 'getvar platform'
Eugeniu Rosca [Tue, 9 Apr 2019 19:11:40 +0000 (21:11 +0200)]
fastboot: add support for 'getvar platform'

Our R-Car3 Android userspace relies on the output of 'fastboot
getvar platform' and U-Boot currently is not able to handle it:

host $> fastboot getvar platform
getvar:platform FAILED (remote: Variable not implemented)
finished. total time: 0.001s

We either have the option of adding 'fastboot.platform' variable
to the default/saved environment as a workaround or add proper
'fastboot getvar platform' support in U-Boot via this patch.
In the latter case, other platforms can benefit from it too.

Note that R-Car3 already exports 'platform' environment variable via
v2019.01 commit 00e4b57e9e71c3 ("ARM: rmobile: Set environment variable
containing CPU type").

Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
5 years agofastboot: Replace literal 32 with PART_NAME_LEN
Alex Kiernan [Tue, 9 Apr 2019 05:30:05 +0000 (05:30 +0000)]
fastboot: Replace literal 32 with PART_NAME_LEN

Where we have to compute partition names, rather than using a hardcoded
32 for the partition name length, replace with PART_NAME_LEN.

Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Reviewed-by: Eugeniu Rosca <rosca.eugeniu@gmail.com>
5 years agostm32mp1: add stusb1600 support for DK1 and DK2 board
Patrick Delaunay [Fri, 29 Mar 2019 14:42:24 +0000 (15:42 +0100)]
stm32mp1: add stusb1600 support for DK1 and DK2 board

The DK1 and DK2 boards use the USB Type-C controller STUSB1600.
This patch updates:
- the device tree to add the I2C node in the DT
- the board stm32mp1 to probe this I2C device and use this controller
  to check cable detection.
- the DWC2 driver to support a new dt property
  "u-boot,force-b-session-valid" which forces B session and
  device mode; it is a workaround because the VBUS sensing and
  ID detection isn't available with stusb1600.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agostm32mp1: migrate USBOTG device to driver model
Patrick Delaunay [Fri, 29 Mar 2019 14:42:23 +0000 (15:42 +0100)]
stm32mp1: migrate USBOTG device to driver model

Use the DWC2 device driver with DM_USB_GADGET support and
cleanup the USB support in STM32MP1 board.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agostm32mp1: remove CONFIG_USB_DWC2, HOST support for USBO
Patrick Delaunay [Fri, 29 Mar 2019 14:42:22 +0000 (15:42 +0100)]
stm32mp1: remove CONFIG_USB_DWC2, HOST support for USBO

Remove the HOST support for STM32MP1 USBO device = OTG DWC2.
The current DWC2 driver have no dynamic detection of device,
So it is dangerous to have start 3V3 when PC is
connected to the micro USB connector.

=> it is preferable to have only DEVICE support
   CONFIG_USB_GADGET_DWC2_OTG for OTG port

See DWC3 driver for clean dual role support...

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agousb: dwc2: add support for STM32MP1
Patrick Delaunay [Fri, 29 Mar 2019 14:42:21 +0000 (15:42 +0100)]
usb: dwc2: add support for STM32MP1

Add compatible "st,stm32mp1-hsotg" and associated driver data to manage
the usb33d-supply and the ST specific register for VBus sensing.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
# Conflicts:
# drivers/usb/gadget/dwc2_udc_otg.c
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2_udc_otg: Add tx_fifo_sz array support
Patrice Chotard [Fri, 29 Mar 2019 14:42:20 +0000 (15:42 +0100)]
usb: dwc2_udc_otg: Add tx_fifo_sz array support

All TX fifo size can be different, add tx_fifo_sz_array[]
into dwc2_plat_otg_data to be able to set them.

tx_fifo_sz_array[] is 17 Bytes long and can contains max 16
tx fifo size (synopsys IP supports max 16 IN endpoints).
First entry of tx_fifo_sz_array[] is the number of valid
fifo size the array contains.

In case of tx_fifo_sz_array[] doesn't contains the same
number of element than max hardware endpoint, display
a warning message.

Compatibility with board which doesn't use tx_fifo_sz_array[]
(Rockchip rk322x/rk3128/rv1108/rk3288/rk3036) is kept.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2_udc_otg: Read MAX_HW_ENDPOINT from HWCFG4 register
Patrick Delaunay [Fri, 29 Mar 2019 14:42:19 +0000 (15:42 +0100)]
usb: dwc2_udc_otg: Read MAX_HW_ENDPOINT from HWCFG4 register

Some DWC2 ip variant doesn't use 16 hardware endpoint as hardcoded
in the driver. Bits INEps [29:26] of HWCFG4 register allows to get
this information.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2: Add function for session B check
Patrick Delaunay [Fri, 29 Mar 2019 14:42:18 +0000 (15:42 +0100)]
usb: dwc2: Add function for session B check

Add a new function to check the session B validity, to be use to check
cable connection.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2: Add force-b-session-valid support
Patrick Delaunay [Fri, 29 Mar 2019 14:42:17 +0000 (15:42 +0100)]
usb: dwc2: Add force-b-session-valid support

Handle "force-b-session-valid" property from DT.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2: force reset assert before to probe the driver
Patrick Delaunay [Fri, 29 Mar 2019 14:42:16 +0000 (15:42 +0100)]
usb: dwc2: force reset assert before to probe the driver

Reset the hardware to be sure of the device state.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2: convert driver to DM_USB_GADGET
Patrick Delaunay [Fri, 29 Mar 2019 14:42:15 +0000 (15:42 +0100)]
usb: dwc2: convert driver to DM_USB_GADGET

Minimal conversion to driver model by using the uclass
UCLASS_USB_GADGET_GENERIC based on:
- reset uclass
- clock uclass
- generic uclass.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: dwc2: remove unused variable regs_otg
Patrick Delaunay [Fri, 29 Mar 2019 14:42:14 +0000 (15:42 +0100)]
usb: dwc2: remove unused variable regs_otg

Remove the global regs_otg variable.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
5 years agophy: usbphyc: increase PLL wait timeout
Patrick Delaunay [Fri, 29 Mar 2019 14:42:13 +0000 (15:42 +0100)]
phy: usbphyc: increase PLL wait timeout

wait 200us to solve USB init issue on device mode
(ums and stm32prog commands)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agophy: usbphyc: move vdda1v1 and vdda1v8 in phy_init
Patrick Delaunay [Fri, 29 Mar 2019 14:42:12 +0000 (15:42 +0100)]
phy: usbphyc: move vdda1v1 and vdda1v8 in phy_init

vdda1v1 and vdda1v8 are used by the PLL.
Both need to be enabled before starting the PLL.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agophy: usbphyc: Binding update of vdda supply
Patrick Delaunay [Fri, 29 Mar 2019 14:42:11 +0000 (15:42 +0100)]
phy: usbphyc: Binding update of vdda supply

Move supply vdda1v1 and vdda1v8 in usbphyc node and
no more in port

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agophy: usbphyc: update xlate with DT binding
Patrick Delaunay [Fri, 29 Mar 2019 14:42:10 +0000 (15:42 +0100)]
phy: usbphyc: update xlate with DT binding

Parameter added for port 1, for example:

&usbh_ehci {
phys = <&usbphyc_port0>;
phy-names = "usb";
vbus-supply = <&vbus_sw>;
status = "okay";
};

&usbotg_hs {
pinctrl-names = "default";
pinctrl-0 = <&usbotg_hs_pins_a>;
phys = <&usbphyc_port1 0>;
phy-names = "usb2-phy";
status = "okay";
};

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agophy: usbphyc: remove unused variable index
Patrick Delaunay [Fri, 29 Mar 2019 14:42:09 +0000 (15:42 +0100)]
phy: usbphyc: remove unused variable index

Remove unused field index in struct stm32_usbphyc_phy.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
5 years agofastboot: Improve error reporting on 'getvar partition-{size, type}'
Eugeniu Rosca [Thu, 28 Mar 2019 13:31:33 +0000 (14:31 +0100)]
fastboot: Improve error reporting on 'getvar partition-{size, type}'

Currently U-Boot reports the same error message in all below cases:
[A] host> fastboot getvar partition-type
[B] host> fastboot getvar partition-size
[C] host> fastboot getvar partition-type:
[D] host> fastboot getvar partition-size:
[E] host> fastboot getvar partition-type:<invalid-part>
[F] host> fastboot getvar partition-size:<invalid-part>

The message looks like:
host> fastboot getvar partition-size:
getvar:partition-size: FAILED (remote: partition not found)
Finished. Total time: 0.003s

Be more user friendly and output:
 - "partition not given" for [A-D]
 - "partition not found" for [E-F]

Fixes: f73a7df984a9 ("net: fastboot: Merge AOSP UDP fastboot")
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Acked-by: Alex Kiernan <alex.kiernan@gmail.com>
5 years agofastboot: getvar: correct/rename "has_slot" to "has-slot"
Eugeniu Rosca [Tue, 26 Mar 2019 16:46:14 +0000 (17:46 +0100)]
fastboot: getvar: correct/rename "has_slot" to "has-slot"

Since its inception in upstream fastboot android-n-preview-1 [1],
"has-slot" option has never taken the form of "has_slot". Amongst the
users of "getvar has-slot:" is the upstream bootloadertest.py [2].

Current U-Boot "has_slot" version must be a typo. Fix it.

[1] https://android.googlesource.com/platform/system/core/+/a797479bd51c
    ("Fix fastboot variable name")
[2] https://android.googlesource.com/platform/system/extras/+/72de393e118e3
    ("Bootloader verification for AndroidThings.")

Fixes: f73a7df984a9 ("net: fastboot: Merge AOSP UDP fastboot")
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Acked-by: Alex Kiernan <alex.kiernan@gmail.com>
5 years agodfu: Avoid declaring unused variables and absent parameters
Andy Shevchenko [Mon, 4 Mar 2019 14:04:44 +0000 (16:04 +0200)]
dfu: Avoid declaring unused variables and absent parameters

The compiler is not happy when neither USB nor TFTP transport for DFU defined:

cmd/dfu.c: In function ‘do_dfu’:
cmd/dfu.c:31:8: warning: unused variable ‘devstring’ [-Wunused-variable]
  char *devstring = argv[3];
        ^~~~~~~~~
cmd/dfu.c:30:8: warning: unused variable ‘interface’ [-Wunused-variable]
    char *interface = argv[2];
          ^~~~~~~~~

Surround those variables by #ifdef expression.

More serious, that comes under same circumstances, is a compilation error due
to absence of macro parameter:

In file included from include/image.h:45,
                 from include/common.h:35,
                 from cmd/dfu.c:13:
include/command.h:207:24: error: expected expression before ‘,’ token
 # define _CMD_HELP(x) x,
                        ^
include/command.h:286:18: note: in expansion of macro ‘_CMD_HELP’
    _cmd, _usage, _CMD_HELP(_help) _CMD_COMPLETE(_comp) }
                  ^~~~~~~~~
include/command.h:290:3: note: in expansion of macro ‘U_BOOT_CMD_MKENT_COMPLETE’
   U_BOOT_CMD_MKENT_COMPLETE(_name, _maxargs, _rep, _cmd, \
   ^~~~~~~~~~~~~~~~~~~~~~~~~
include/command.h:332:2: note: in expansion of macro ‘U_BOOT_CMD_COMPLETE’
  U_BOOT_CMD_COMPLETE(_name, _maxargs, _rep, _cmd, _usage, _help, NULL)
  ^~~~~~~~~~~~~~~~~~~
cmd/dfu.c:70:1: note: in expansion of macro ‘U_BOOT_CMD’
 U_BOOT_CMD(dfu, CONFIG_SYS_MAXARGS, 1, do_dfu,
 ^~~~~~~~~~
make[1]: *** [scripts/Makefile.build:279: cmd/dfu.o] Error 1
make: *** [Makefile:1518: cmd] Error 2

Put empty string unconditionally to have macro parameter present.

Fixes: 0f44d33536a5 ("dfu: Fix up the Kconfig mess")
Cc: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Lukasz Majewski <lukma@denx.de>
5 years agousb: Select USB_MUSB_DSPS with USB_MUSB_TI
Alex Kiernan [Thu, 18 Apr 2019 11:10:50 +0000 (11:10 +0000)]
usb: Select USB_MUSB_DSPS with USB_MUSB_TI

USB_MUSB_TI requires USB_MUSB_DSPS, failing at link time if it's not
selected:

  drivers/usb/musb-new/built-in.o: In function `ti_musb_host_ofdata_to_platdata':
  drivers/usb/musb-new/ti-musb.c:193: undefined reference to `musb_dsps_ops'

or if OF_CONTROL is not selected:

  arch/arm/mach-omap2/built-in.o:(.data.usb0+0x24): undefined reference to `musb_dsps_ops'

Reviewed-by: Hannes Schmelzer <oe5hpm@oevsv.at>
Tested-by: Hannes Schmelzer <oe5hpm@oevsv.at>
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
5 years agoconfigs: Migrate USB_MUSB_DISABLE_BULK_COMBINE_SPLIT to Kconfig
Alex Kiernan [Fri, 12 Apr 2019 10:51:05 +0000 (10:51 +0000)]
configs: Migrate USB_MUSB_DISABLE_BULK_COMBINE_SPLIT to Kconfig

Migrate support for disable MUSB bulk split/combine to Kconfig

Green Travis build:

https://travis-ci.org/akiernan/u-boot/builds/519101867

Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
5 years agoMerge tag 'arc-for-2019.07' of git://git.denx.de/u-boot-arc
Tom Rini [Thu, 18 Apr 2019 16:12:16 +0000 (12:12 -0400)]
Merge tag 'arc-for-2019.07' of git://git.denx.de/u-boot-arc

In this small series we migrate ARC boards to DM_MMC
so we're hopefully are good now and our boards will be kept
in U-Boot for some more time :)

5 years agoARC: [plat-axs10x]: migrate to DM_MMC
Eugeniy Paltsev [Thu, 21 Mar 2019 13:37:23 +0000 (16:37 +0300)]
ARC: [plat-axs10x]: migrate to DM_MMC

Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
5 years agoARC: [plat-hsdk]: migrate to DM_MMC
Eugeniy Paltsev [Mon, 25 Feb 2019 15:35:29 +0000 (18:35 +0300)]
ARC: [plat-hsdk]: migrate to DM_MMC

Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
5 years agoARC: dwmmc: Adding DesignWare MMC driver support for ARC devboards
Eugeniy Paltsev [Mon, 25 Feb 2019 15:35:28 +0000 (18:35 +0300)]
ARC: dwmmc: Adding DesignWare MMC driver support for ARC devboards

Add the DM_MMC-compatible DesignWare MMC driver support for Synopsys
ARC devboards. It is created to switch ARC devboards to use DM_MMC.

It required information such as clocks (Bus Interface Unit clock,
Card Interface Unit clock) and SDIO bus width.

Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
5 years agotimer: dw-apb: Add missing 64bit up-conversion
Marek Vasut [Wed, 10 Apr 2019 11:44:05 +0000 (13:44 +0200)]
timer: dw-apb: Add missing 64bit up-conversion

The generic timer count is an incrementing 64bit value and a timer driver
must return an incrementing 64bit value. The DW APB timer only provides a
32bit timer counting down, thus the result must be inverted and converted
to a 64bit value. The current implementation is however missing the 64bit
up-conversion and this results in random timer roll-overs, which in turn
triggers random timeouts throughout the codebase.

This patch adds the missing 64bit up-conversion to fix the issue.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <chin.liang.see@intel.com>
Cc: Dinh Nguyen <dinguyen@kernel.org>
Cc: Ley Foon Tan <ley.foon.tan@intel.com>
Cc: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Cc: Tien Fong Chee <tien.fong.chee@intel.com>
5 years agoarm: dts: Stratix10: Add QSPI node
Ley Foon Tan [Wed, 3 Apr 2019 05:45:02 +0000 (13:45 +0800)]
arm: dts: Stratix10: Add QSPI node

Merge qspi dts node from Linux.
Commit 0cb140d07fc75fb (arm64: dts: stratix10: Add QSPI support for Stratix10)

Add -u-boot.dtsi files for non Linux dts properties and
update properties for Uboot.
- add u-boot,dm-pre-reloc
- add alias for spi0
- change compatible for flash
- support quad read and quad write
- change maximum frequency to 100MHz

Tested on Stratix 10 SoC devkit.
SOCFPGA_STRATIX10 # sf probe 0:0
SF: Detected mt25qu02g with page size 256 Bytes, erase size 64 KiB, total 256 MiB

Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoddr: altera: Stratix10: Add ECC memory scrubbing
Ley Foon Tan [Thu, 21 Mar 2019 17:24:05 +0000 (01:24 +0800)]
ddr: altera: Stratix10: Add ECC memory scrubbing

Scrub memory content if ECC is enabled and it is not
from warm reset boot.

Enable icache and dcache before scrub memory
and use "DC ZVA" instruction to clear memory
to zeros. This instruction writes a cache line
at a time and it can prevent false ECC error
trigger if write cache line partially.

Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoarm: socfpga: stratix10: Add cpu_has_been_warmreset()
Ley Foon Tan [Thu, 21 Mar 2019 17:24:04 +0000 (01:24 +0800)]
arm: socfpga: stratix10: Add cpu_has_been_warmreset()

Add helper function cpu_has_been_warmreset() to check
if CPU is from warm reset boot.

Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoarm: dts: Stratix10: Modify stratix10 socdk memory node
Ley Foon Tan [Thu, 21 Mar 2019 17:24:03 +0000 (01:24 +0800)]
arm: dts: Stratix10: Modify stratix10 socdk memory node

The stratix10 socdk ships with 4GB of memory.  Modify the
device tree to represent this.  Note that to access 4GB of
memory in Stratix 10, due to the IO space from 2GB to 4GB,
we use the fact that the DDR controller ignores upper address
bits outside of the configured DRAM's size.  This means that
, the 4GB DRAM is mapped to memory every 4GB.

For an 8GB memory, you can either live with the 2GB IO space,
and loose access to that memory from the processor, or use
the same trick:

Loose 2GB of memory:
        memory {
                device_type = "memory";
                /* 8GB */
/* first 2GB */
                reg = <0 0x00000000 0 0x80000000>,
/* last 4GB */
                      <1 0x00000000 1 0x00000000>;
                u-boot,dm-pre-reloc;
        };

or to map it all:
        memory {
                device_type = "memory";
                /* 8GB */
/* first 2GB */
                reg = <0 0x00000000 0 0x80000000>,
/* next 6GB */
                      <2 0x80000000 1 0x80000000>;
                u-boot,dm-pre-reloc;
        };

Signed-off-by: Dalon Westergreen <dalon.westergreen@intel.com>
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoconfigs: stratix10: Change CONFIG_NR_DRAM_BANKS to 2
Ley Foon Tan [Thu, 21 Mar 2019 17:24:02 +0000 (01:24 +0800)]
configs: stratix10: Change CONFIG_NR_DRAM_BANKS to 2

Stratix10 maps dram in 2 address spans, from 0-2GB and from
2GB up to 128GB.

Signed-off-by: Dalon Westergreen <dalon.westergreen@intel.com>
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoddr: altera: Stratix10: Add multi-banks DRAM size check
Ley Foon Tan [Thu, 21 Mar 2019 17:24:01 +0000 (01:24 +0800)]
ddr: altera: Stratix10: Add multi-banks DRAM size check

Stratix 10 maps dram from 0 to 128GB.  There is a 2GB hole
in the memory for peripherals and other IO from 2GB to 4GB.
However the dram controller ignores upper address bits for
smaller dram configurations.  Example: a 4GB dram
maps to multiple locations, every 4GB on the address.

Signed-off-by: Dalon Westergreen <dalon.westergreen@intel.com>
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoddr: altera: stratix10: Move SDRAM size check to SDRAM driver
Ley Foon Tan [Thu, 21 Mar 2019 17:24:00 +0000 (01:24 +0800)]
ddr: altera: stratix10: Move SDRAM size check to SDRAM driver

Move SDRAM size check to SDRAM driver. sdram_calculate_size()
is called in SDRAM initialization already, avoid calling
twice in size check function.

Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
5 years agoarm: socfpga: implement proper peripheral reset
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:36 +0000 (20:12 +0100)]
arm: socfpga: implement proper peripheral reset

This commit removes ad-hoc reset handling for peripheral resets from SPL
for socfpga gen5.

This is done because as U-Boot drivers support reset handling by now.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agospi: cadence_qspi: add reset handling
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:35 +0000 (20:12 +0100)]
spi: cadence_qspi: add reset handling

This adds reset handling to the cadence qspi driver.

For backwards compatibility, only a warning is printed when failing to
get reset handles.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agomtd: rawnand: denali: add reset handling
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:34 +0000 (20:12 +0100)]
mtd: rawnand: denali: add reset handling

This adds reset handling to the devicetree-enabled Denali NAND driver.

For backwards compatibility, only a warning is printed when failing to
get reset handles.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoarm: socfpga: move gen5 SDR driver to DM
Simon Goldschmidt [Tue, 16 Apr 2019 20:04:39 +0000 (22:04 +0200)]
arm: socfpga: move gen5 SDR driver to DM

To clean up reset handling for socfpga gen5, port the DDR driver to DM
using UCLASS_RAM and implement proper reset handling.

This gets us rid of one ad-hoc call to socfpga_per_reset().

The gen5 driver is implemented in 2 distinct files. One of it (containing
the calibration training) is not touched much and is kept at using
hard coded addresses since the code grows even more otherwise.

SPL is changed from calling hard into the DDR driver code to just
probing UCLASS_RESET and UCLASS_RAM. It is happy after finding a RAM
driver after that.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoreset: socfpga: add reset handling for old kernels
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:32 +0000 (20:12 +0100)]
reset: socfpga: add reset handling for old kernels

This adds code to take peripherals out of reset based on an environment
variable. This is in preparation for removing the code that does this from
SPL.

However, some drivers even in current Linux cannot handle peripheral reset,
so until this works, we need a compatibility workaround.

This workaround is implemented in the 'assert' and 'remove' callbacks of
this reset driver: the 'assert' callback does not disable peripherals that
were already taken out of reset, while the 'remove' callback, which is
called on OS_PREPARE, deasserts all peripheral resets if the environment
variable "socfpga_legacy_reset_compat" is set to 1, which is what the gen5
SPL did up to now.

This is in preparation to clean up the SPL and implementing proper reset
handling for U-Boot.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoarm: socfpga: gen5: deassert peripheral reset by default
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:31 +0000 (20:12 +0100)]
arm: socfpga: gen5: deassert peripheral reset by default

To keep the current behaviour of taking all peripherals out of reset
before booting the OS before removing that code from socfpga gen5 SPL,
this enables the new behaviour by default for all gen5 boards by adding
the environment variable "socfpga_legacy_reset_compat=1" to the default
environment.

This can be overridden in board config files or by saving an environment
without this variable enabled.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoreset: socfpga: rename membase ptr to modrst_base
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:30 +0000 (20:12 +0100)]
reset: socfpga: rename membase ptr to modrst_base

The only member of this driver's priv struct is a pointer, which is
called 'membase'. However, since this driver handles multiple sub-
architectures, this is not the base address from dts but the base
address of some common registers of those sub-arches.

Reflect this better in sourcecode by renaming 'membase' to 'modrst_base'.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoarm: socfpga: gen5: add reset & sdr node to SPL devicetrees
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:29 +0000 (20:12 +0100)]
arm: socfpga: gen5: add reset & sdr node to SPL devicetrees

The SPL for socfpga gen5 currently takes all peripherals out of reset
unconditionally. To implement proper reset handling for peripherals,
the reset node has to be provided with the SPL dts.

In preparation to move the DDR driver to DM, the sdr node is required
in SPL, too.

This patch adds "u-boot,dm-pre-reloc" to U-Boot specific dtsi addon
files so that the reset manager and SDR driver correctly probe in SPL.
It centralizes these settings into a common file since in contrast to
boot-type specific nodes, "soc", "rst" and "sdr" are always needed.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoarm: socfpga: gen5: sync devicetrees to Linux
Simon Goldschmidt [Fri, 1 Mar 2019 19:12:28 +0000 (20:12 +0100)]
arm: socfpga: gen5: sync devicetrees to Linux

This is again a sync to linux-next + pending patches in Dinh's tree at
commit 1c909b2dfe6a ("ARM: dts: socfpga: update more missing reset
properties")'

It adds missing peripheral reset properties to socfpga.dtsi and removes
U-Boot specific leftovers from socfpga_cyclone5_socrates.dts.

Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
5 years agoMerge branch 'master' of git://git.denx.de/u-boot-spi
Tom Rini [Wed, 17 Apr 2019 13:21:32 +0000 (09:21 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-spi

- drop non-DM code from ti_qspi
- support spi-mem for ti_qspi

5 years agoarm: am57xx: cl-som-am57x: remove board support
Uri Mashiach [Sun, 14 Apr 2019 09:17:53 +0000 (12:17 +0300)]
arm: am57xx: cl-som-am57x: remove board support

U-Boot support for the CL-SOM-AM57x module is no longer required.

Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
5 years agoMerge branch 'master' of git://git.denx.de/u-boot-sunxi
Tom Rini [Wed, 17 Apr 2019 13:19:45 +0000 (09:19 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-sunxi

- Convert DM_MMC and DM_SCSI
- A20, R40, H6 Linux dts(i) sync
- CLK, RESET support for sunxi, sun8_emac net drivers

5 years agoMerge tag 'xilinx-for-v2019.07' of git://git.denx.de/u-boot-microblaze
Tom Rini [Wed, 17 Apr 2019 13:19:13 +0000 (09:19 -0400)]
Merge tag 'xilinx-for-v2019.07' of git://git.denx.de/u-boot-microblaze

Xilinx/FPGA changes for v2019.07

fpga:
- Add support for external data in FIT
- Extend testing for external data case
- Inform user about a need to run post config on Zynq

arm:
- Tune zynq command functions
- Fix internal variable setting

arm64:
- Add support for zc39dr decoding
- Disable WDT for zcu100
- Small changes in reset_reason()
- Some DT changes (spi)
- Tune qspi-mini configuration
- Remove useless eeprom setting
- Fix two sdhci boot case

spi:
- Fix tap delay programming

clk:
- Enable i2c in SPL

net:
- Fix gem phydev handling
- Remove phy detection code from gem driver

general:
- Correct EXT_DTB usage for MULTI_DTB_FIT configuration

5 years agoMerge tag 'uniphier-v2019.07' of git://git.denx.de/u-boot-uniphier
Tom Rini [Wed, 17 Apr 2019 13:16:38 +0000 (09:16 -0400)]
Merge tag 'uniphier-v2019.07' of git://git.denx.de/u-boot-uniphier

UniPhier SoC updates for v2019.07

- Sync DT with Linux 5.1-rc4

- Enable CONFIG_SUPPORT_EMMC_RPMB for uniphier_v8_defconfig

5 years agospi: ti_qspi: Convert to spi-mem ops
Vignesh Raghavendra [Tue, 16 Apr 2019 16:02:00 +0000 (21:32 +0530)]
spi: ti_qspi: Convert to spi-mem ops

Convert driver to use  spi-mem ops in order to support accelerated MMIO
flash interface in generic way and for better performance.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agospi: ti_qspi: Drop non DM code
Vignesh Raghavendra [Tue, 16 Apr 2019 16:01:59 +0000 (21:31 +0530)]
spi: ti_qspi: Drop non DM code

Now that all boards using TI QSPI have moved to DM and DT, drop non DM
code completely.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
[jagan: update MIGRATION.txt, rebase config_whitelist.txt]
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agosunxi: update SATA driver to always use DM_SCSI
Andre Przywara [Fri, 12 Apr 2019 10:58:54 +0000 (16:28 +0530)]
sunxi: update SATA driver to always use DM_SCSI

It seems like the Allwinner SATA driver is already quite capable of
using the driver model, so we can force this on all boards and can
remove support for a non-DM_SCSI build.
This removes the warning about boards with SATA ports not being
DM_SCSI compliant.

It also takes the opportunity to move the driver out of the board/sunxi
directory to join its siblings in drivers/ata, and to make it a proper
Kconfig citizen.

The board defconfigs stay untouched.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jagan Teki <jagan@openedev.com>
[jagan: select DM_SCSI separately]
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoarm: sunxi: Enable DM_MMC and DM_SCSI
Jagan Teki [Fri, 12 Apr 2019 11:18:25 +0000 (16:48 +0530)]
arm: sunxi: Enable DM_MMC and DM_SCSI

- Enable DM_MMC if MMC defined
- Enable DM_SCSI if SCSI defined

globally through Allwinner platform, the effected SoC families
and boards will make use of MMC and SCSI subsystems in driver-model.

Tested DM_MMC in one board from A64, H6, H5, H3, R40, A83T, A20, A10
SoCs.

Tested-by: Pablo Sebastián Greco <pgreco@centosproject.org> # BPI-M2-Ultra
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoARM: dts: sun8i-r40-bananapi-m2-berry: Enable AHCI
Jagan Teki [Mon, 15 Apr 2019 06:12:32 +0000 (11:42 +0530)]
ARM: dts: sun8i-r40-bananapi-m2-berry: Enable AHCI

Enable ahci node for BPI-M2-Berry, this would require since
we have DM_SCSI enabled on the respective SoC.

Unable to sync the same node from Linux, since the similar change
is still in Linux ML.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoARM: dts: a20-wits-pro-a20-dkt: Enable AHCI
Jagan Teki [Fri, 12 Apr 2019 12:11:58 +0000 (17:41 +0530)]
ARM: dts: a20-wits-pro-a20-dkt: Enable AHCI

Enable ahci node for a20-wits-pro-a20-dkt, this would require since
we have DM_SCSI enabled on the respective SoC.

Right now, ahci enabled in -u-boot.dtsi and will remove once same
supported by Linux.

Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoARM: dts: a20-m5: Enable AHCI
Jagan Teki [Fri, 12 Apr 2019 12:05:24 +0000 (17:35 +0530)]
ARM: dts: a20-m5: Enable AHCI

Enable ahci node for sun7i-a20-m5.dts, this would require since
we have DM_SCSI enabled on the respective SoC.

No need to send patch to Linux for this change, since this
dts is U-Boot specific.

Cc: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoboard: sunxi: Add R40 sata compatible
Jagan Teki [Fri, 12 Apr 2019 11:17:56 +0000 (16:47 +0530)]
board: sunxi: Add R40 sata compatible

Add sata compatible for R40.

Cc: Pablo Sebastián Greco <pgreco@centosproject.org>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoarm: allwinner: dts: a20: Sync A20 dts(i) files from Linux 5.1-rc2
Jagan Teki [Fri, 12 Apr 2019 10:49:34 +0000 (16:19 +0530)]
arm: allwinner: dts: a20: Sync A20 dts(i) files from Linux 5.1-rc2

Sync sun7i-a20 dts(i) files from Linux 5.1-rc2

Linux commit details about the sun7i-a20* sync:
"ARM: dts: sun7i: bananapi: Add GPIO banks regulators"
(sha1: 09c6572290f018d73ec2e812e28bada34d41815f)

Here are U-Boot specific dts changes.

- s/uart0_pins_a/uart0_pb_pins for
  sun7i-a20-ainol-aw1.dts
  sun7i-a20-m5.dts
  sun7i-a20-primo73.dts
  sun7i-a20-yones-toptech-bd1078.dts
  sunxi-itead-core-common.dtsi
- s/gmac_pins_mii_a/gmac_rgmii_pins for
  sun7i-a20-m5.dts
- drop i2c0, i2c1 pins from
  sunxi-itead-core-common.dtsi
- drop mmc0 pins from
  sun7i-a20-primo73.dts

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoarm: allwinner: r40: Sync R40 dts(i) files from Linux 5.1-rc2
Jagan Teki [Tue, 9 Apr 2019 09:38:51 +0000 (15:08 +0530)]
arm: allwinner: r40: Sync R40 dts(i) files from Linux 5.1-rc2

Sync sun8i-r40 dts(i) files from Linux 5.1-rc2

Linux commit details about the sun8i-r40* sync:
"ARM: dts: sun8i: r40: bananapi-m2-ultra: Add Bluetooth device node"
(sha1: 1e5f1db4ccd8348a21da55bff82f4263000879ef)

Linux commit details about the sun8i-v40* sync:
"ARM: dts: sunxi: Fix I2C bus warnings"
(sha1: 0729b4af5753b65aa031f58c435da53dbbf56d19)

Cc: Pablo Sebastián Greco <pgreco@centosproject.org>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoARM: uniphier_v8: enable CONFIG_SUPPORT_EMMC_RPMB
Masahiro Yamada [Fri, 12 Apr 2019 09:55:51 +0000 (18:55 +0900)]
ARM: uniphier_v8: enable CONFIG_SUPPORT_EMMC_RPMB

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agoARM: dts: uniphier: sync with Linux 5.1-rc4
Masahiro Yamada [Fri, 12 Apr 2019 09:55:50 +0000 (18:55 +0900)]
ARM: dts: uniphier: sync with Linux 5.1-rc4

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agoboard: sunxi: gmac: Remove Ethernet clock and reset
Jagan Teki [Wed, 27 Feb 2019 18:57:01 +0000 (00:27 +0530)]
board: sunxi: gmac: Remove Ethernet clock and reset

Since Ethernet clock and reset is now handling via
CLK and RESET frameworks via driver API's remove
explicit ccm writes.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agonet: sun8i_emac: Add CLK and RESET support
Jagan Teki [Wed, 27 Feb 2019 18:56:58 +0000 (00:26 +0530)]
net: sun8i_emac: Add CLK and RESET support

Add CLK and RESET support for sun8i_emac driver to
enable TX clock and reset pins via CLK and RESET
framework.

Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Lothar Felten <lothar.felten@gmail.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
5 years agonet: sun8i_emac: Retrieve GMAC clock via 'syscon' phandle
Jagan Teki [Wed, 27 Feb 2019 18:56:51 +0000 (00:26 +0530)]
net: sun8i_emac: Retrieve GMAC clock via 'syscon' phandle

Unlike other Allwinner SoC's R40 GMAC clock control register
is locate in CCU, but rest located via syscon itself. Since
the phandle property for current code look for 'syscon' and
it will grab the respective ccu or syscon base address based
on DT property defined in respective SoC dtsi.

So, use the existing 'syscon' code even for R40 for retrieving
GMAC clock via CCU and update the register directly in
sun8i_emac_set_syscon instead of writing it separately using
ccm base.

Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Lothar Felten <lothar.felten@gmail.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agonet: sunxi_emac: Add CLK support
Jagan Teki [Wed, 27 Feb 2019 18:56:50 +0000 (00:26 +0530)]
net: sunxi_emac: Add CLK support

Add CLk support for sunxi_emac to enable AHB_EMAC clock
via CLK framework.

Cc: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
5 years agoclk: sunxi: r40: Fix GMAC reset reg offset
Jagan Teki [Mon, 15 Apr 2019 11:12:16 +0000 (16:42 +0530)]
clk: sunxi: r40: Fix GMAC reset reg offset

GMAC reset reg offset added by below commit seems to assume
it as EMAC but R40 indeed using GMAC.
"clk: sunxi: Implement EMAC, GMAC clocks, resets"
(sha1: 68620c9698f109c1f001f80d282138a5c67cabef)

So, fix by updating the reg offset for RST_BUS_GMAC.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
5 years agoarm64: zynqmp: fix preprocessor check for SPL_ZYNQMP_TWO_SDHCI
Luca Ceresoli [Mon, 15 Apr 2019 14:18:18 +0000 (16:18 +0200)]
arm64: zynqmp: fix preprocessor check for SPL_ZYNQMP_TWO_SDHCI

A missing CONFIG_ prefix while checking for this Kconfig variable makes the
check always fail. Fix it. While there also switch from the '#if defined'
form to the '#ifdef' form as the other checks in this function.

Fixes: 35e2b92344b1 ("arm64: zynqmp: Fix logic around CONFIG_ZYNQ_SDHCI")

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agoarm64: zynqmp: Remove eeprom setting
Michal Simek [Mon, 25 Feb 2019 09:01:22 +0000 (10:01 +0100)]
arm64: zynqmp: Remove eeprom setting

By moving to DM_I2C there is no need to specify any eeprom configuration
because it is read from DT.

Reported-by: Sreeja Vadakattu <sreeja.vadakattu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agonet: gem: Remove phy autodetection code
Michal Simek [Fri, 29 Mar 2019 08:25:09 +0000 (09:25 +0100)]
net: gem: Remove phy autodetection code

There is no reason to detect phy when core is doing it for us.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agonet: zynq_gem: Modify phy supported features after max-speed was set
Siva Durga Prasad Paladugu [Wed, 27 Mar 2019 12:09:59 +0000 (17:39 +0530)]
net: zynq_gem: Modify phy supported features after max-speed was set

The phydev supported features were reset in phy_set_supported() so,
move the setting of driver supported features after this so that it
wont lost in phy_set_supported().

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agoMakefile: Prioritize external dtb if defined
Michal Simek [Sat, 23 Mar 2019 05:43:00 +0000 (11:13 +0530)]
Makefile: Prioritize external dtb if defined

Prioritize external dtb if its passed via EXT_DTB
than the dtb that was built in the tree. With this
patch it appends the specified external dtb to
the u-boot image.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
5 years agoarm: zynq: Add an info message about post config
Siva Durga Prasad Paladugu [Sat, 23 Mar 2019 10:31:36 +0000 (16:01 +0530)]
arm: zynq: Add an info message about post config

Post configuration cant be run at u-boot as u-boot
didn't has any info about the design.So,this patch
adds an info message that post config was not run
and needs to be run manually if needed.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agoarm64: zynqmp: Add idcode for new RFSoC silicon ZU39DR
Siva Durga Prasad Paladugu [Sat, 23 Mar 2019 09:30:06 +0000 (15:00 +0530)]
arm64: zynqmp: Add idcode for new RFSoC silicon ZU39DR

This patch adds "zu39dr" to the list of zynqmp devices
The zu39DR is the new RFSoC silicon with id value of 0x66.

Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>