Commit graph

51 commits

Author SHA1 Message Date
Patrick Delaunay
b2ac9645e6 ram: stm32mp1: remove __maybe_unused on stm32mp1_ddr_setup
Since the commit f42045b2e7 ("stm32mp15: replace CONFIG_TFABOOT when
it is possible") the function stm32mp1_ddr_setup is always called so the
__maybe_unused can be removed.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
2021-11-30 16:43:28 +01:00
Patrick Delaunay
d72e7bbe7c ram: stm32mp1: compute DDR size from DDRCTL registers
Compute the DDR size from DDR controller register (mstr and addrmap)
in U-Boot proper as the DDR information are useful only for SPL
but not for U-Boot proper, for example with TFABOOT.

This patch simplify U-Boot DT when several DDR size are supported
and support of next SOC in STM32MP family.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
2021-11-30 16:43:28 +01:00
Patrick Delaunay
b3c29dc9e5 stm32mp1: ram: remove tuning support
Remove the DDR interactive command tuning, as the support of a predefined
DDR PHY tuning is removed for STM32MP1 driver in SPL and in TF-A
and the result of this tuning will be never used.

Moreover this SW tuning procedure can failed on some hardware
configuration (to many BIST errors and no convergence); it will be no
more supported in the next delivery of the DDR utilities included in
the CubeMX tool of STMicroelectronics.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
2021-11-30 16:43:28 +01:00
Patrick Delaunay
9819fe345c stm32mp1: ram: remove the support of calibration result
The support of a predefined DDR PHY tuning result is removed for
STM32MP1 driver because it is not needed at the supported frequency
when built-in calibration is executed.

The calibration parameters were provided in the device tree by the
optional node "st,phy-cal", activated in ddr helper file by the
compilation flag DDR_PHY_CAL_SKIP and filled with values generated
by the CubeMX DDR utilities.

This patch
- updates the binding file to remove "st,phy-cal" support
- updates the device trees and remove the associated defines
- simplifies the STM32MP1 DDR driver and remove the support of
  the optional parameter "st,phy-cal"

After this patch, the built-in calibration is always executed
and the calibration registers are moved in the phy dynamic part;
that allows manual tests.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
2021-11-30 16:43:28 +01:00
Patrick Delaunay
4831ba2903 stm32mp1: ram: add read valid training support
Add the read data eye training = training for optimal read valid placement
(RVTRN) when the built-in calibration is executed for LPDDR2 and LPDDR3.

This training is supported on the PUBL integrated in the STM32MP15x
DDR subsystem and it is not required for DDR3.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
2021-11-30 16:43:28 +01:00
Patrick Delaunay
f42045b2e7 stm32mp15: replace CONFIG_TFABOOT when it is possible
In some part of STM32MP15 support the CONFIG_TFABOOT can be replaced
by other config: CONFIG_ARMV7_PSCI and CONFIG_ARM_SMCCC.

This patch also simplifies the code in cpu.c, stm32mp1_ram.c and
clk_stml32mp1.c as execution of U-Boot in sysram (boot without SPL and
without TFA) is not supported: the associated initialization code is
present only in SPL.

This cleanup patch is a preliminary step to support SPL load of OP-TEE
in secure world, with SPL in secure world and U-Boot in no-secure world.

Reported-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
2021-11-10 17:07:31 +01:00
Michal Simek
b4c2c151b1 Kconfig: Remove all default n/no options
default n/no doesn't need to be specified. It is default option anyway.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
[trini: Rework FSP_USE_UPD portion]
Signed-off-by: Tom Rini <trini@konsulko.com>
2021-08-31 17:47:49 -04:00
Simon Glass
401d1c4f5d common: Drop asm/global_data.h from common header
Move this out of the common header and include it only where needed.  In
a number of cases this requires adding "struct udevice;" to avoid adding
another large header or in other cases replacing / adding missing header
files that had been pulled in, very indirectly.   Finally, we have a few
cases where we did not need to include <asm/global_data.h> at all, so
remove that include.

Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
2021-02-02 15:33:42 -05:00
Patrick Delaunay
66b3b9db69 ram: stm32mp1: migrate trace to dev or log macro
Define LOG_CATEGORY, use dev_ macro when it is possible
and migrate other trace to log_ macro.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2021-01-13 09:52:58 +01:00
Simon Glass
41575d8e4c dm: treewide: Rename auto_alloc_size members to be shorter
This construct is quite long-winded. In earlier days it made some sense
since auto-allocation was a strange concept. But with driver model now
used pretty universally, we can shorten this to 'auto'. This reduces
verbosity and makes it easier to read.

Coincidentally it also ensures that every declaration is on one line,
thus making dtoc's job easier.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-13 08:00:25 -07:00
Heinrich Schuchardt
c670aeee3d common: rename getc() to getchar()
The sandbox is built with the SDL2 library with invokes the X11 library
which in turn calls getc(). But getc() in glibc is defined as

    int getc(FILE *)

This does not match our definition.

    int getc(void)

The sandbox crashes when called with parameter -l.

Rename our library symbol getc() to getchar().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2020-10-22 09:54:53 -04:00
Patrick Delaunay
6c393e8c0f ram: stm32mp1: add size and addr parameter to test all
Add size and addr parameter to test "All" to override the default
value (4kB and STM32_DDR_BASE) used in tests with these optional
parameters: [size] or [addr].

When other optional parameters are present before [addr],
they are replaced by default value:
- [loop] = "1"
- [pattern] = "-" (new: force default pattern)

Example to use:

DDR>test 0 1 0x20000

DDR>test 0 1 0x1000000 0xD0000000

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
81b66b9033 ram: stm32mp1: use the DDR size by default in the test addressBus
The test 4 = "AddressBus [size] [addr]" without parameter
detects alias for any address bit only when:
- size = real size of DDR
- addr = start of DDR = 0xC0000000

These value must be the default value when parameters are absent.

This patch sets bufsize to STM32_DDR_SIZE and get_bufsize() selects
the correct value for bufsize when this parameter is absent =
full size of the DDDR

On EV1 board :
DDR> test 4
running at 0xC0000000 length 0x40000000

On DK2 board
DDR> test 4
running at 0xC0000000 length 0x20000000

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
fcd4890829 ram: stm32mp1: add parameter addr in test FrequencySelectivePattern
Add a parameter addr in test FrequencySelectivePattern to select
the base address used to execute the tests.

Default value (when the parameter is absent) is STM32_DDR_BASE,
selected in get_addr() function.

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-07-07 16:01:23 +02:00
Patrick Delaunay
1a5be5a416 ram: stm32mp1: protect minimum value in get_bufsize
Add protection on minimum value for result of get_bufsize
and check the alignment of buffer size: only multiple min_size
is allowed; only 4 bytes alignment was checked previously
(value & 0x3).

For example the "Random" test raises an issue when size is not 8 bytes
aligned because address for buffer = address + size / 2 is not word
aligned.

This patch avoid test error for unsupported size value.

Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2020-07-07 16:01:23 +02:00
Simon Glass
cd93d625fd common: Drop linux/bitops.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:23 -04:00
Simon Glass
c05ed00afb common: Drop linux/delay.h from common header
Move this uncommon header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:23 -04:00
Simon Glass
f7ae49fc4f common: Drop log.h from common header
Move this header out of the common header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 21:19:18 -04:00
Simon Glass
0914011310 command: Remove the cmd_tbl_t typedef
We should not use typedefs in U-Boot. They cannot be used as forward
declarations which means that header files must include the full header to
access them.

Drop the typedef and rename the struct to remove the _s suffix which is
now not useful.

This requires quite a few header-file additions.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 18:36:55 -04:00
Simon Glass
90526e9fba common: Drop net.h from common header
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.

Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-05-18 17:33:31 -04:00
Marek Vasut
0c27c16495 ram: stm32mp1: Add support for multiple configs
Add support for multiple DRAM configuration subnodes, while retaining
the support for a single flat DRAM configuration node. This is useful
on systems which can be manufactured in multiple configurations and
where the DRAM configuration can be determined at runtime.

The code is augmented by a function which can be overridden on board
level, allowing a match on the configuration node name, very much like
the fitImage configuration node name matching works. The default match
is on the single top-level DRAM configuration, if matching on subnodes
is required, then this board_stm32mp1_ddr_config_name_match() must be
overridden.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
2020-05-14 09:02:12 +02:00
Patrick Delaunay
654706be84 configs: stm32mp1: replace STM32MP1_TRUSTED by TFABOOT
Activate ARCH_SUPPORT_TFABOOT and replace the arch stm32mp
specific config CONFIG_STM32MP1_TRUSTED by the generic CONFIG_TFABOOT
introduced by the commit 535d76a121 ("armv8: layerscape: Add TFABOOT
support").
This config CONFIG_TFABOOT is activated for the trusted boot chain,
when U-Boot is loaded by TF-A.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
2020-04-15 09:08:37 +02:00
Patrick Delaunay
9368bdfebd ram: stm32mp1: the property st, phy-cal becomes optional
This parameter "st,phy-cal" becomes optional and when it is
absent the built-in PHY calibration is done.

It is the case in the helper dtsi file "stm32mp15-ddr.dtsi"
except if DDR_PHY_CAL_SKIP is defined.

This patch also impact the ddr interactive mode
- the registers of the param 'phy.cal' are initialized to 0 when
  "st,phy-cal" is not present in device tree (default behavior when
  DDR_PHY_CAL_SKIP is not activated)
- the info 'cal' field can be use to change the calibration behavior
  - cal=1 => use param phy.cal to initialize the PHY, built-in training
             is skipped
  - cal=0 => param phy.cal is absent, built-in training is used (default)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:23:30 +01:00
Patrick Delaunay
d424e6786f ram: stm32mp1: reduce delay after BIST reset for tuning
Reduce the delay after BIST delay, from 1ms to 10us
which is enough accoriding datasheet.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:23:26 +01:00
Patrick Delaunay
b604a41c6b ram: stm32mp1_ddr: fix self refresh disable during DQS training
DDRCTRL_PWRCTL.SELFREF_EN needs to be reset before DQS training step, not
to enter in self refresh mode during the execution of this phase.
Depending on settings, it can be set after the DQS training.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:23:18 +01:00
Patrick Delaunay
8c9ce08075 ram: stm32mp1: update BIST config for tuning
Update the BIST config to compute the real use mask for the real
bank, row and col of the used DDR. The values are get from addrmap
register value.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Patrick Delaunay
27e7b4edea ram: stm32mp1: tuning: deactivate derating during BIST test
The derating (timing parameter derating using MR4 read value)
can't be activated during BIST test, as the MR4 read answer will
be not understood by BIST (BISTGSR.BDONE bit stay at 0,
BISTWCSR.DXWCNT = 0x206 instead of BISTWCR.BWCNT = 0x200).

This patch only impacts the tuning on LPDDR2/LPDDR3,
if derateen.derate_enable = 1.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Patrick Delaunay
f711d1f080 ram: stm32mp1: tuning: add timeout for polling BISTGSR.BDDONE
Avoid to block the tuning procedure on BIST error (not finished
BIST procedure) by adding a 1000us timeout on the polling of
BISTGSR.BDDONE executed to detect the end of BIST.

The normal duration of the BIST test is around 5us.

This patch also cleanup comments.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Patrick Delaunay
1c55a91b9d ram: stm32mp1: don't display the prompt two times
Remove one "DDR>" display on command
- next
- step
- go

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Patrick Delaunay
c8eb4e038c ram: stm32mp1: display result for software read DQS gating
Display result information for software read DQS gating, the tuning 0
which be used by CubeMX DDR tuning tools.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Patrick Delaunay
e9a20f8a19 ram: stm32mp1: increase vdd2_ddr: buck2 for 32bits LPDDR
Need to increase the LPDDR2/LPDDR3 the voltage vdd2_ddr: buck2
form 1.2V to 1.25V for 32bits configuration.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
2020-03-24 14:20:50 +01:00
Simon Glass
9b4a205f45 common: Move RAM-sizing functions to init.h
These functions relate to memory init so move them into the init
header.

Signed-off-by: Simon Glass <sjg@chromium.org>
2020-01-17 14:02:35 -05:00
Patrick Delaunay
757bca8d19 stm32mp1: ram: add pattern parameter in infinite write test
Add pattern for infinite test_read and test_write, that
allow to change the pattern to test without recompilation;
default pattern is 0xA5A5AA55.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-08-27 11:19:23 +02:00
Patrick Delaunay
25331ae1c1 stm32mp1: ram: reload watchdog during ddr test
Avoid watchdog during infinite DDR test.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-08-27 11:19:23 +02:00
Patrick Delaunay
37f41ae900 stm32mp1: ram: update loop management in infinite test
Reduce verbosity of the infinite tests to avoid CubeMX issue.
test and display loop by 1024*1024 accesses: read or write.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-08-27 11:19:23 +02:00
Patrick Delaunay
4b0496fe79 stm32mp1: ram: fix address issue in 2 tests
If user choose to test memory size is 1GByte (0x40000000),
memory address would overflow in test "Random" and
test "FrequencySelectivePattern".
Thus the system would hangs up when running DDR test.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Bossen WU <bossen.wu@st.com>
2019-08-27 11:19:23 +02:00
Patrick Delaunay
375c28ac76 stm32mp1: ram: cosmetic: remove unused prototype
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-08-27 11:19:23 +02:00
Patrick Delaunay
588448517f ram: stm32mp1_ram: Fix warnings when compiling with W=1
This patch solves the following warnings:

drivers/ram/stm32mp1/stm32mp1_ram.c: In function 'stm32mp1_ddr_clk_enable':
warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
  for (idx = 0; idx < ARRAY_SIZE(clkname); idx++) {
                    ^
drivers/ram/stm32mp1/stm32mp1_ram.c: In function 'stm32mp1_ddr_setup':
warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
  for (idx = 0; idx < ARRAY_SIZE(param); idx++) {
                    ^
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-07-12 11:50:56 +02:00
Patrick Delaunay
187c41d783 stm32mp1: ram: add tuning in DDR interactive mode
Add command tuning for DDR interactive mode, used during
board bring-up or with CubeMX DDR tools to execute software
tuning for the DDR configuration:
- software read DQS Gating (replace the built-in one)
- Bit de-skew
- Eye Training or DQS training

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:11 +02:00
Patrick Delaunay
0d44752442 stm32mp1: ram: add tests in DDR interactive mode
Add command tests for DDR interactive mode, used during
board bring-up or with CubeMX DDR tools to verify the
DDR configuration.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:11 +02:00
Patrick Delaunay
01a7510849 stm32mp1: ram: add interactive mode for DDR configuration
This debug mode is used by CubeMX DDR tuning tools
or manualy for tests during board bring-up.
It is simple console used to change DDR parameters and check
initialization.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:11 +02:00
Patrick Delaunay
1767ac2d1f stm32mp1: ram: add support for LPDDR2/LPDDR3
Manage power supply configuration for board using stpmic1
with LPDDR2 or with LPDDR3:
+ VDD_DDR1 = 1.8V with BUCK3 (bypass if possible)
+ VDD_DDR2 = 1.2V with BUCK2

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:11 +02:00
Patrick Delaunay
53bb831658 stm32mp1: ram: update parameter array initialization
Force alignment of the size of parameters array with
the expected value in the binding, that allows compilation
error when the array size change.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:10 +02:00
Patrick Delaunay
c60fed14f6 stm32mp1: ram: change ddr speed to kHz
Allow fractional support in DDR tools.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:10 +02:00
Patrick Delaunay
0cb1aa9409 stm32mp1: ram: increase the delay after reset to 128 cycles
Component Notification DDR controller errata (3.00a):9001313030
Synchronization Time Waited After De-assertion of presetn is
128 pclk Cycles.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:10 +02:00
Patrick Delaunay
c3ec370aed stm32mp1: ram: update mask for operating mode
Regression introduced by rebase, when loop
was replaced by readl_poll_timeout() function.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-05-23 11:38:10 +02:00
Patrick Delaunay
abf2678f0f stm32mp1: add trusted boot with TF-A
Add support of trusted boot, using TF-A as first stage bootloader,
The boot sequence is
  BootRom >=> TF-A.stm32 (clock & DDR) >=> U-Boot.stm32

The TF-A monitor provides secure monitor with support of SMC
- proprietary to manage secure devices (BSEC for example)
- PSCI for power

The same device tree is used for STMicroelectronics boards with
basic boot and with trusted boot.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
2019-04-12 16:09:13 +02:00
Masahiro Yamada
d35812368a regmap: change regmap_init_mem() to take ofnode instead udevice
Currently, regmap_init_mem() takes a udevice. This requires the node
has already been associated with a device. It prevents syscon/regmap
from behaving like those in Linux.

Change the first argumenet to take a device node.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2018-05-07 15:49:51 -04:00
Tom Rini
4549e789c1 SPDX: Convert all of our multiple license tags to Linux Kernel style
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from.  So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry.  Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.

In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.

This commit changes all instances where we have multiple licenses (in
these cases, dual license) declared in the SPDX-License-Identifier tag.
In this case we change from listing "LICENSE-A LICENSE-B" or "LICENSE-A
or LICENSE-B" or "(LICENSE-A OR LICENSE-B)" to "LICENSE-A OR LICENSE-B"
as per the Linux Kernel style document.  Note that parenthesis are
allowed so when they were used before we continue to use them.

Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
2018-05-07 10:24:31 -04:00
Tom Rini
d024236e5a Remove unnecessary instances of DECLARE_GLOBAL_DATA_PTR
We have a large number of places where while we historically referenced
gd in the code we no longer do, as well as cases where the code added
that line "just in case" during development and never dropped it.

Signed-off-by: Tom Rini <trini@konsulko.com>
2018-04-27 14:54:48 -04:00