The designware spi driver unconditionally uses polling.
The comment to spi_hw_init() also states that the function should disable
interrupts.
According to the DesignWare DW_apb_ssi Databook, value 0xff in IMR enables
all interrupts. Since we want to mask all interrupts write 0x0 instead.
On the canaan k210 board, pressing the reset button twice to reset the
board will run u-boot. If u-boot boots Linux without having SPI interrupts
masked, Linux will hang as soon as interrupts are enabled, because of an
interrupt storm.
Properly masking the SPI interrupts in u-boot allows us to successfully
boot Linux, even after resetting the board.
Fixes: 5bef6fd79f ("spi: Add designware master SPI DM driver used on SoCFPGA")
Signed-off-by: Sean Anderson <seanga2@gmail.com>
[Niklas: rewrite commit message]
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
The Canaan Kendryte K210 SoC DW apb_ssi v4 spi controller is documented
to have a 32 word deep TX and RX FIFO, which spi_hw_init() detects.
However, when the RX FIFO is filled up to 32 entries (RXFLR = 32), an
RX FIFO overrun error occurs. Avoid this problem by force setting
fifo_len to 31.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
"kendryte" is the marketing name for the K210 RISC-V SoC produced by
Canaan Inc. Rather than "kendryte,k210", use the usual "canaan,k210"
vendor,SoC compatibility string format in the device tree files and
use the SoC name for file names.
With these changes, the device tree files are more in sync with the
Linux kernel DTS and drivers, making uboot device tree usable by the
kernel.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
The driver is currently using sizeof(op->cmd.opcode) in the op_len
calculation. Commit d15de62301 ("spi: spi-mem: allow specifying a
command's extension") changed op->cmd.opcode from one byte to two.
Instead, a new struct member op->cmd.nbytes is supposed to be used.
For regular commands op->cmd.nbytes will be one.
Commit d15de62301 ("spi: spi-mem: allow specifying a command's
extension") did update some drivers that overload the generic mem_ops()
implementation, but forgot to update dw_spi_mem_ops().
Calculating op_len incorrectly causes dw_spi_mem_ops() to misbehave, since
op_len is used to determine how many bytes that should be read/written.
On the canaan k210 board, this causes the probe of the SPI flash to fail.
Fix the op_len calculation in dw_spi_mem_ops(). Doing so results in
working SPI flash on the canaan k210 board.
Fixes: d15de62301 ("spi: spi-mem: allow specifying a command's extension")
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Pratyush Yadav <p.yadav@ti.com>
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Driver model: Rename U_BOOT_DEVICE et al.
dtoc: Tidy up and add more tests
ns16550 code clean-up
x86 and sandbox minor fixes for of-platdata
dtoc prepration for adding build-time instantiation
-----BEGIN PGP SIGNATURE-----
iQFFBAABCgAvFiEEslwAIq+Gp8wWVbYnfxc6PpAIreYFAl/09LURHHNqZ0BjaHJv
bWl1bS5vcmcACgkQfxc6PpAIrebjwwf/fHRjYsAY/Yj/+y1xgo3L3sphIvQUqTDF
KkLl+kHdV5r8W/HJULxLQcF2r7pcPEI6TAQxuj3qQ5SUvm2HviS8GHGPawDEwyht
HgBp9VD56+HUadMfnbG//DVS73ycbL4XSKlYqpkINEejtnlttsCIawUXX5cTyGM/
59VkgnKrKvJQRUXvYLa8MTugTs4fkPJGDqhActBk/7SP1SImj+rfalNSqA2/dx6y
2RnPCSzB1x2231KSj+B1NgGlR3Xb8P8zgh20ijcEU/hrlXBTZyi7K7f4SJR30Efu
LYkkuj4VbxcV/25RozR0fmknqCs0QyAI+/dql6TNtbTSPC/jAfj0jQ==
=9kN3
-----END PGP SIGNATURE-----
Merge tag 'dm-pull-5jan21' of git://git.denx.de/u-boot-dm into next
Driver model: make some udevice fields private
Driver model: Rename U_BOOT_DEVICE et al.
dtoc: Tidy up and add more tests
ns16550 code clean-up
x86 and sandbox minor fixes for of-platdata
dtoc prepration for adding build-time instantiation
Most drivers use these access methods but a few do not. Update them.
In some cases the access is not permitted, so mark those with a FIXME tag
for the maintainer to check.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Pratyush Yadav <p.yadav@ti.com>
The designware ssi device has "broken" chip select behaviour [1], and needs
specific manipulation to use the built-in chip select. The existing fix is
to use an external GPIO for chip select, but typically the K210 has SPI3
directly connected to a flash chip with dedicated pins. This makes it
impossible to use the spi_xfer function to use spi, since the CS is
de-asserted in between calls. This patch adds an implementation of
exec_op, which gives correct behaviour when reading/writing spi flash.
This patch also rearranges the headers to conform to U-Boot style.
[1] https://lkml.org/lkml/2015/12/23/132
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
CTRLR0 can have several different layouts depending on the specific device
(dw-apb-ssi vs dwc-ssi), and specific parameters set during synthesis.
Update the driver to support three specific configurations: dw-apb-ssi with
SSI_MAX_XFER_SIZE=16, dw-apb-ssi with SSI_MAX_XFER_SIZE=32, and dwc-ssi.
dw-apb-ssi is the version of the device on Altera/Intel SoCFPGAs, MSCC
SoCs, and Canaan Kendryte K210 SoCs. This is the only version this driver
supported before this change. The register layout before version 3.23a is:
| 31 .. 16 |
| other stuff |
| 15 .. 10 | 9 .. 8 | 7 .. 6 | 5 .. 4 | 3 .. 0 |
| other stuff | TMOD | MODE | FRF | DFS |
Note that DFS (Data Frame Size) is only 4 bits, limiting transfers to data
frames of 16 bits or less.
In version 3.23a, the SSI_MAX_XFER_SIZE parameter was introduced. This
parameter defaults to 16 (resulting in the same layout as prior versions),
but may also be set to 32. To allow setting longer data frame sizes, a new
DFS_32 register was introduced:
| 31 .. 21 | 20 .. 16 |
| other stuff | DFS_32 |
| 15 .. 10 | 9 .. 8 | 7 .. 6 | 5 .. 4 | 3 .. 0 |
| other stuff | TMOD | MODE | FRF | all zeros |
The old DFS field no longer controls the data frame size. To detect this
layout, we try writing 0xF to DFS. If we read back 0x0, then this device
has SSI_MAX_XFER_SIZE=32.
dwc-ssi is the version of the device on Intel Keem Bay SoCs and Canaan
Kendryte K210 SoCs. The layout of ctrlr0 is:
| 31 .. 16 |
| other stuff |
| 15 .. 12 | 11 .. 10 | 9 .. 8 | 7 .. 6 | 4 .. 0 |
| other stuff | TMOD | MODE | FRF | DFS_32 |
The semantics of the fields have not changed since the previous version.
However, SSI_MAX_XFER_SIZE is effectively always 32.
To support these different layouts, we model our approach on the one
which the Linux kernel has taken. During probe, the driver calls an init
function stored in driver_data. This init function is responsible for
determining the layout of CTRLR0, and supplying the update_cr0 function.
The style of and information behind this commit is based on the Linux MMIO
driver for these devices. Specific reference was made to the series adding
support for Intel Keem Bay SoCs [1].
[1] https://lore.kernel.org/linux-spi/20200505130618.554-1-wan.ahmad.zainie.wan.mohamad@intel.com/
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
This should reduce the size of the struct, and also groups more similar
fields together.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
A few registers had slightly different names from what is in the datasheet.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Using an fdt-specific function causes problems when compiled with a live
tree.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
This property is named differently than other SPI drivers with the same
property, as well as the property as used in Linux.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
This allows different log levels to be enabled or disabled depending on the
desired level of verbosity. In particular, it allows for general debug
information to be printed while excluding more verbose logging which may
interfere with timing.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
The resting state of MOSI is high when nothing is driving it. If we drive
it low while recieving, it looks like we are transmitting 0x00 instead of
transmitting nothing. This can confuse slaves (like SD cards) which allow
new commands to be sent over MOSI while they are returning data over MISO.
The return of MOSI from 0 to 1 at the end of recieving a byte can look like
a start bit and a transmission bit to an SD card. This will cause the card
to become out-of-sync with the SPI device, as it thinks the device has
already started transmitting two bytes of a new command. The mmc-spi driver
will not detect the R1 response from the SD card, since it is sent too
early, and offset by two bits. This patch fixes transfer errors when using
SD cards with dw spi.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
This name is far too long. Rename it to remove the 'data' bits. This makes
it consistent with the platdata->plat rename.
Signed-off-by: Simon Glass <sjg@chromium.org>
We use 'priv' for private data but often use 'platdata' for platform data.
We can't really use 'pdata' since that is ambiguous (it could mean private
or platform data).
Rename some of the latter variables to end with 'plat' for consistency.
Signed-off-by: Simon Glass <sjg@chromium.org>
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>
Use the _ptr suffixed variant instead of casting. Also, convert it to
dev_read_addr_ptr(), which is safe to CONFIG_OF_LIVE.
One curious part is an error check like follows in
drivers/watchdog/omap_wdt.c:
priv->regs = (struct wd_timer *)devfdt_get_addr(dev);
if (!priv->regs)
return -EINVAL;
devfdt_get_addr() returns FDT_ADDR_T_NONE (i.e. -1) on error.
So, this code does not catch any error in DT parsing.
dev_read_addr_ptr() returns NULL on error, so this error check
will work.
I generated this commit by the following command:
$ find . -name .git -prune -o -name '*.[ch]' -type f -print | \
xargs sed -i -e 's/([^*)]*\*)devfdt_get_addr(/dev_read_addr_ptr(/'
I manually fixed drivers/usb/host/ehci-mx6.c
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Use the _ptr suffixed variant instead of casting. Also, convert it to
dev_read_addr_ptr(), which is safe to CONFIG_OF_LIVE.
One curious part is an error check like follows in
drivers/watchdog/omap_wdt.c:
priv->regs = (struct wd_timer *)devfdt_get_addr(dev);
if (!priv->regs)
return -EINVAL;
devfdt_get_addr() returns FDT_ADDR_T_NONE (i.e. -1) on error.
So, this code does not catch any error in DT parsing.
dev_read_addr_ptr() returns NULL on error, so this error check
will work.
I generated this commit by the following command:
$ find . -name .git -prune -o -name '*.[ch]' -type f -print | \
xargs sed -i -e 's/([^*)]*\*)devfdt_get_addr(/dev_read_addr_ptr(/'
I manually fixed drivers/usb/host/ehci-mx6.c
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
At present dm/device.h includes the linux-compatible features. This
requires including linux/compat.h which in turn includes a lot of headers.
One of these is malloc.h which we thus end up including in every file in
U-Boot. Apart from the inefficiency of this, it is problematic for sandbox
which needs to use the system malloc() in some files.
Move the compatibility features into a separate header file.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present if CONFIG_SPL_GPIO_SUPPORT is enabled then the GPIO uclass
is included in SPL/TPL without any control for boards. Some boards may
want to disable this to reduce code size where GPIOs are not needed in
SPL or TPL.
Add a new Kconfig option to permit this. Default it to 'y' so that
existing boards work correctly.
Change existing uses of CONFIG_DM_GPIO to CONFIG_IS_ENABLED(DM_GPIO) to
preserve the current behaviour. Also update the 74x164 GPIO driver since
it cannot build with SPL.
This allows us to remove the hacks in config_uncmd_spl.h and
Makefile.uncmd_spl (eventually those files should be removed).
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Disable and free clock when remove driver.
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Convert 'dw_spi_ofdata_to_platdata' to use 'dev_read_u32_default'
instead of 'fdtdec_get_int' and get rid of DECLARE_GLOBAL_DATA_PTR.
Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
With current order of include files, the file designware_spi.c
can't see that the struct global_data has the member
board_type when CONFIG_BOARD_TYPES is defined. By not seeing this
then all the members are shifted in the struct global_data.
So when the driver is trying to read from device tree blob, it
would pass the wrong address to the function 'fdtdev_get_int'.
This will make to use the default frequency 500000.
The fix consists of changing the order of include files in
designware_spi.c to include first common.h file.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
On some platforms, as the Ocelot ones, when wanting to control the CS
through software, it is not possible to do it through the GPIO
controller. Indeed, this signal is managed through a dedicated range of
registers inside the SoC..
By declaring the external_cs_manage function as weak, it is possible to
manage the CS at platform level and then using the appropriate registers.
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Add code to reset all reset signals as in SPI DT node. A reset property
is an optional feature, so only print out a warning and do not fail if a
reset property is not present.
If a reset property is discovered, then use it to deassert, thus
bringing the IP out of reset.
Release reset when _remove().
Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagan Teki <jagan@openedev.com>
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 a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
While switching to readl_poll_timeout macros from custom code
the waiting condition was accidently inverted, so it was pure
luck that this code works at least in some conditions.
Fix that by inverting exit condition for readl_poll_timeout.
Fixes: c6b4f031d9 ("DW SPI: fix tx data loss on FIFO flush")
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
Current DW SPI driver uses 32 bit access for some registers and
16 bit access for others. So if DW SPI IP is connected via bus
which doesn't support 16 bit access we will get bus error.
Fix that by switching to 32 bit access only instead of 16 and 32 bit mix
Additional Documentation to Support this Change:
The DW_apb_ssi databook states:
"All registers in the DW_apb_ssi are addressed at 32-bit boundaries
to remain consistent with the AHB bus. Where the physical size of
any register is less than 32-bits wide, the upper unused bits of
the 32-bit boundary are reserved. Writing to these bits has no
effect; reading from these bits returns 0." [1]
[1] Section 6.1 of dw_apb_ssi.pdf (version 3.22a)
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
DW SPI internal chip select management has limitation:
it hold CS line in active state only when the FIFO is not
empty. If the FIFO freed before we add new data the SPI transaction will
be broken.
So add option to use external gpio for chip select. Gpio can be added
via device tree using standard gpio bindings.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
There is no sense in waiting for RX data in dw_reader function:
there is no chance that RX data will appear in RX FIFO if
RX FIFO is empty after previous TX write in dw_writer function.
So get rid of this waiting. After that we can get rid of dw_reader
return value and make it returning void. After that we can get rid
of dw_reader return value check in poll_transfer function.
With these changes we're getting closer to Linux DW SPI driver.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
In current implementation we get -ETIMEDOUT error when we try to use
transmit only mode (SPI_TMOD_TO)
This happens because in transmit only mode input FIFO never gets any data
which breaks our logic in dw_reader(): we are waiting until RX data will be
ready in dw_reader, but this newer happens, so we return with error.
Fix that by using SPI_TMOD_TR instead of SPI_TMOD_TO which allows to use
RX FIFO.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
In current implementation if some data still exists in Tx FIFO it
can be silently flushed, i.e. dropped on disabling of the controller,
which happens when writing 0 to DW_SPI_SSIENR (it happens in the
beginning of new transfer)
So add wait for current transmit operation to complete to be sure
that current transmit operation is finished before new one.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
Add option to set spi controller clock frequency via device tree
using standard clock bindings.
Define dw_spi_get_clk function as 'weak' as some targets
(like SOCFPGA_GEN5 and SOCFPGA_ARRIA10) don't use standard clock API
and implement dw_spi_get_clk their own way in their clock manager.
Get rid of clock_manager.h include as we don't use
cm_get_spi_controller_clk_hz function anymore. (we use redefined
dw_spi_get_clk in SOCFPGA clock managers instead)
Reviewed-by: Marek Vasut <marex@denx.de>
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
These support the flat device tree. We want to use the dev_read_..()
prefix for functions that support both flat tree and live tree. So rename
the existing functions to avoid confusion.
In the end we will have:
1. dev_read_addr...() - works on devices, supports flat/live tree
2. devfdt_get_addr...() - current functions, flat tree only
3. of_get_address() etc. - new functions, live tree only
All drivers will be written to use 1. That function will in turn call
either 2 or 3 depending on whether the flat or live tree is in use.
Note this involves changing some dead code - the imx_lpi2c.c file.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present devices use a simple integer offset to record the device tree
node associated with the device. In preparation for supporting a live
device tree, which uses a node pointer instead, refactor existing code to
access this field through an inline function.
Signed-off-by: Simon Glass <sjg@chromium.org>
Replace numeric mask hexcodes with GENMASK macro
in designware_spi
Cc: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Jagan Teki <jteki@openedev.com>
Replace numerical bit shift with BIT macro
in designware_spi
:%s/(1 << nr)/BIT(nr)/g
where nr = 0, 1, 2 .... 31
Cc: Stefan Roese <sr@denx.de>
Cc: Marek Vasut <marex@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Jagan Teki <jteki@openedev.com>
This is a convenient way for a driver to get the hardware address of a
device, when regmap or syscon are not being used. Change existing callers
to use it as an example to others.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
By specification the FIFO size would be in a range 2-256 bytes. From TX Level
prospective it means we can set threshold in the range 0-(FIFO size - 1) bytes.
Hence there are currently two issues:
a) FIFO size 2 bytes is actually skipped since TX Level is 1 bit and could be
either 0 or 1 byte;
b) FIFO size is incorrectly decreased by 1 which already done by meaning of
TX Level register.
Fixes: 501943696e (spi: designware_spi: Fix detecting FIFO depth)
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
This is common to all SPI drivers and specifies a structure used by the
uclass. It makes more sense to define it in the uclass.
Reviewed-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Linux now also contains SPI driver, yet the name is 'snps,dw-apb-ssi'.
Fix the naming before we have to support both names.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@opensource.altera.com>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pavel Machek <pavel@denx.de>
Cc: Stefan Roese <sr@denx.de>
Cc: Vince Bridgers <vbridger@opensource.altera.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Acked-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Jagannadha Sutradharudu Teki <jagannadh.teki@gmail.com>
