'archive' concatenates files into a single binary blob. Files are
indexed by the base names. See archive.h for the format description.
BUG=chromium:502066
BRANCH=tot
TEST=Tested on Glados
Change-Id: Iea108160e65c8c7bd34c02af824a77cb075ee64b
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 21a9ba860f29599ac029f8d49d32399c4e3a73a8
Original-Change-Id: I46b4efb339e3a1e05772ae752f2861026ca09cfc
Original-Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/311200
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://review.coreboot.org/12925
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
On Apollolake CPU memory mapping is similar to previous SoC, and
we place CBMEM right under TSEG.
Change-Id: I606f690449ba98af6e9fc3074d677c7287892164
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13883
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
UDELAY_IO is defined in src/cpu/x86/Kconfig, so it does
not need to be redefined in the AMD cpu or board Kconfigs.
Change-Id: I6676881c0ba5d1634230fc3d3c37da3afbc6fceb
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/13780
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The current MTRR API doesn't allow one to detect variable MTRRs
along with handling fixed MTRRs in one function call. Therefore,
add x86_setup_mtrrs_with_detect() to perform the same actions
as x86_setup_mtrrs() but always do the dynamic detection.
Change-Id: I443909691afa28ce11882e2beab12e836e5bcb3d
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13935
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins)
Add the routines to handle the UPDs for SiliconInit. Currently no
support is required.
Testing on Galileo:
* Edit the src/mainboard/intel/galileo/Makefile.inc file:
* Add "select ADD_FSP_PDAT_FILE"
* Add "select ADD_FSP_RAW_BIN"
* Add "select ADD_RMU_FILE"
* Place the FSP.bin file in the location specified by CONFIG_FSP_FILE
* Place the pdat.bin files in the location specified by
CONFIG_FSP_PDAT_FILE
* Place the rmu.bin file in the location specified by CONFIG_RMU_FILE
* Build EDK2 CorebootPayloadPkg/CorebootPayloadPkgIa32.dsc to generate
UEFIPAYLOAD.fd
* Edit .config file and add the following lines:
* CONFIG_DISPLAY_UPD_DATA=y
* Testing successful if coreboot calls SiliconInit
Change-Id: I5176ab4b1ea7681c3095f102a86f4b614366c0fc
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/13897
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Build make with the rest of the toolchain, since the targets using
a Chromium EC need make 4.x
Change-Id: I7efb0c25f605f16c2d9a1e7c4b203f3bcdae671b
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: https://review.coreboot.org/13923
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This romstage is minimalistic. Its goal is to set up some BARs
that FSP expects to be set and then invoke FSP driver to train
memory.
Change-Id: I3fa56aafe99cf6cf062a46dece3a0febeafdbfad
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13805
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
On Apollo Lake SPI flash is memory mapped. The mapping is different
to previous platforms. Only "BIOS" region is mapped in contrast to
whole flash. Also, the 128 KiB right below 4 GiB are being decoded by
readonly SRAM. Fail accesses to those regions, rather than returning
false data.
Change-Id: Iac3fa74cd221a5a46ceb34c2a79470290bcc2d84
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13706
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This adds a few helper functions that are intended to assist setting
up framebuffer.
Change-Id: Id8ed4de1f9de32e9222b0120c15a6d33676346e7
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13802
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
FSP creates hand-off-blocks (HOBs) to exchange information with
coreboot. This adds a set of utilities to parse HOBs and extract
some useful information from them.
Change-Id: If55dbfaa021cd68c312813a5532a36c68806dbbc
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13801
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This adds Notify Phase API. This is an important call that is used
to inform FSP runtimes of different stages of SoC initializations
by the coreboot.
Change-Id: Icec770d0c1c4d239adb2ef342bf6cc9c35666e4d
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13800
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This adds SiliconInit API that is needed to be called after memory
has been trained. This call is needed to let the blob do various
initialisations of IP blocks.
Change-Id: I35e02f22174c8392e55ac869265a19c4309932e5
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13799
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This adds implementation of fsp_memory_init() that is used to train
memory.
Change-Id: I72268aaa91eea7e4d4f072d70a47871d74c2b979
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13798
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Remove dependency on early_serial.c and instead use the
Super I/O's header to access the functions needed.
Also re-organize some of the superio code in order
to succesfully compile the rom.
Change-Id: I85a6f1352ae3b91c3c98e4d3fa0b90b87e02babc
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/13925
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
The E3800 datasheet only lists 2K and 20K Pull Strength for the GPIOs.
The 10K and 40K values map to 'reserved'.
This brings the code closer to the non-FSP baytrail.
Change-Id: I77078bdbbccc00976525dc43fb98f5b2e79eae03
Signed-off-by: Ben Gardner <gardner.ben@gmail.com>
Reviewed-on: https://review.coreboot.org/13907
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Split out the MTRR support into a new module: mtrr.c.
TEST=Build and run on Galileo
Change-Id: Ib9ec479d171dbbc062509e14fbe246f6d90e903a
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/13895
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Turn on the SD controller to allow it to claim resources.
Testing on Galileo:
* Edit the src/mainboard/intel/galileo/Makefile.inc file:
* Add "select ADD_FSP_PDAT_FILE"
* Add "select ADD_FSP_RAW_BIN"
* Add "select ADD_RMU_FILE"
* Place the FSP.bin file in the location specified by CONFIG_FSP_FILE
* Place the pdat.bin files in the location specified by
CONFIG_FSP_PDAT_FILE
* Place the rmu.bin file in the location specified by CONFIG_RMU_FILE
* Build EDK2 CorebootPayloadPkg/CorebootPayloadPkgIa32.dsc to generate
UEFIPAYLOAD.fd
* Edit .config file and add the following lines:
* CONFIG_PAYLOAD_ELF=y
* CONFIG_PAYLOAD_FILE="path to UEFIPAYLOAD.fd"
* Testing successful when at the UEFI shell prompt:
* After issuing:
* "connect -r"
* "map -r"
* The "dir" command displays the contents of the SD flash card
* The "drivers" command shows an SD host and SD media connection
Change-Id: I883dc87270045786ddb931bea83fc36646a128e6
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/13894
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Add a link to the "Driver Writer's Guide" and a link to the "EDK II
firmware for Intel Quark SoC X1000" document.
TEST=None
Change-Id: I8d629d06accfe24a0b8971b5b5868849587c3db7
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/13893
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
Add a link to "Making a bootable SD card"
TEST=None
Change-Id: I5682fdd51a4ba37f97ad35475e11d9843f1498fb
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/13892
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This allows memtest86+ to be added to CBFS as a 'secondary'
payload on x86 systems, to be loaded by the main payload
if desired.
Selecting this option, which defaults to no, builds the memtest86+
payload and adds it to CBFS as `img/memtest` which can then be
loaded by for example SeaBIOS or GRUB.
Change-Id: Iecf876aaf588ba1df7abdf6668cb26f089bf5f42
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13858
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Ben Gardner <gardner.ben@gmail.com>
If the TPM code isn't getting built in, the Kconfig symbol
CONFIG_TPM_TIS_BASE_ADDRESS doesn't exist. This ends up creating
an invalid operating region in the ACPI tables, causing a bluescreen
in windows.
This should fix this issue:
https://ticket.coreboot.org/issues/35
"commit 85a255fb (acpi/tpm: Gracefully handle missing TPM module)
breaks Windows"
Change-Id: I32e0e09c1f61551a40f4842168f556d5e1940d28
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13890
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This adds a few assembly lines that are generic enough to be shared
between romstage and verstage that are ran in CAR. The GDT reload
is bypassed and the stack is reloaded with the CAR stack defined
in car.ld. The entry point for all those stages is car_stage_entry().
Change-Id: Ie7ef6a02f62627f29a109126d08c68176075bd67
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13861
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Attempt to better document the symbol usage in car.ld for
cache-as-ram usage. Additionally, add _car_region_[start|end]
that completely covers the entire cache-as-ram region. The
_car_data_[start|end] symbols were renamed to
_car_relocatable_data_[start|end] in the hopes of making it
clearer that objects within there move. Lastly, all these
symbols were added to arch/symbols.h.
Change-Id: I1f1af4983804dc8521d0427f43381bde6d23a060
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13804
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Because the perl error messages go to stderr, we were not catching these
on the build server. If the script has an issue, we want to know
immediately, so change the bash script that calls into the perl lint
tool to pipe these to stdout.
Change-Id: Ieeec9ccbd59177cfd1859a9738a4ee1fab803d28
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13877
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Remove dependency on early_serial.c and instead use the
Super I/O's header to access the functions needed.
Change-Id: I9edf7fc2501aa832106dda9213e702dbcc1200b4
Signed-off-by: Antonello Dettori <dev@dettori.io>
Reviewed-on: https://review.coreboot.org/13887
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins)
The PLL multiplier value is off by one for DDR3-1866 due to a
wrong TCK value, resulting in DDR3-1600 being used by the PLL.
Needs test on real hardware !
Change-Id: I657b813889945f0d9990dd11680a3d3a25b53467
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/13613
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Sandy and Ivy Bridge processors use the same socket, and a mainboard
with the socket can support both types of CPUs. However, they use
different native graphics init code for LVDS and cause a crash if
running the wrong code.
This change detects the CPU type and then selects the right code to
run. It will add some more code in ramstage. It also merges the
{SANDY,IVY}BRIDGE_LVDS symbol to one SANDYBRIDGE_IVYBRIDGE_LVDS.
Tested on a Lenovo T520 with i7-2630qm and i7-3720qm
Signed-off-by: Iru Cai <mytbk920423@gmail.com>
Change-Id: I4624759f9c92d56d547db1ab4b9a1d611a182a91
Reviewed-on: https://review.coreboot.org/12087
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Tested-by: build bot (Jenkins)
__asm__ is more robust to compilation flags.
Change-Id: Ic7ca6e38ddd439dcfc4a62ef272ecea62416b4be
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: https://review.coreboot.org/13905
Tested-by: build bot (Jenkins)
Reviewed-by: Julius Werner <jwerner@chromium.org>
Those options have no effect or lead to compile error on ARM due
to fundamental incompatibilities. Add proper "depends on" clauses
to hide them.
Change-Id: I860fbd331439c25efd8aa92023195fda3add2e2c
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-on: https://review.coreboot.org/13904
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Test that the coreboot toolchain version of IASL is being used by
looking for the string 'coreboot toolchain' instead of a specific
version number. While this may cause people to have to rebuild
their toolchains again now, it helps to prevent toolchain failures
when bisecting in the future.
Change-Id: I9913eeae8f29ddc3ec8c70077c05d898595eb283
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/12847
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The what-jenkins-does build runs distclean when building the utilities.
It doesn't fail the build if distclean fails, but it generates a
scary warning.
Change-Id: Iac90958951976ed326a89ef2b5f2d9f17f9f2d6b
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13888
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Instead of keeping track of all the combinations of entry points
depending on the stage and other options just use _start. That way,
there's no need to update the arch/header.ld for complicated cases
as _start is always the entry point for a stage.
Change-Id: I7795a5ee1caba92ab533bdb8c3ad80294901a48b
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13882
Tested-by: build bot (Jenkins)
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
In order to align the entry points for the various stages
on x86 to _start one needs to rename the reset_vector symbol.
The section is the same; it's just a symbol change.
Change-Id: I0e6bbf1da04a6e248781a9c222a146725c34268a
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13881
Tested-by: build bot (Jenkins)
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Until recently x86 romstage used to be linked at some default
address. The address itself is not meaningful because the code
was normally relocated at address calculated during insertion
in CBFS. Since some newer SoC run romstage at CAR it became
useful to link romstage code at some address in CAR and avoid
relocation during build/run time altogether.
Change-Id: I11bec142ab204633da0000a63792de7057e2eeaf
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13860
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
In order to avoid collisions with other _start symbols while
grepping and future ones be explicit about which _start this
one is: the 16-bit one only used by the reset vector in the
bootblock.
Change-Id: I6d7580596c0e6602a87fb158633ce9d45910cec2
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13880
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
It's helpful to see the reset vector in objdump output. Without
it being marked executable it doesn't get displayed.
Change-Id: I85cb72ea0727d3f3c2186ae20b9c5cfe5d23aeed
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13879
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins)
Patrick at least indicated this jump after the reset
vector jump was a remnant from some construct used long
ago in the project. It's not longer used (nor could I find
where it was). Therefore, remove it.
Change-Id: I31512c66a9144267739b08d5f9659c4fcde1b794
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13878
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Tested-by: build bot (Jenkins)
This adds a set of utility functions that help load and identify
FSP blobs.
Change-Id: I1d23f60fd1dc8de7966142bcd793289220a1fa5e
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13797
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This adds important header files that specify calling interface between
coreboot and FSP.
Change-Id: I393601c91e3c3f630e0fc899f1140ecefed8ecba
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13796
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
In Kconfig files, the 'if' and 'endif' statements need to match up. A
file can't start an if statement that's completed in the next file.
Add a check as the files are being parsed to make sure that they match
up correctly.
Change-Id: If51207ea037089ab84c768e5a868270468cf4c4f
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/13876
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Parse manufacturer id and ASCII serial.
Required for SMBIOS type 17 field.
Change-Id: I710de1a6822e4777c359d0bfecc6113cb2a5ed8e
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/13862
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
1. Change the function which integrated one firmware, to the function
which pushes the whole group. Use fw_table as a parameter instead
of using the global table name.
2. Let PSP2 and PSP1 not dependent on the other. It turns out PSP2
can exist without PSP1. For some APU, the PSP directory has to be
put in PSP2 field (ROMSIG 0x14).
3. Reserve 32 more bytes in PSP2 header. It is defined by spec. It
is tested, and it is true.
These above changes are overlapping, hard to split them. Sorry.
Change-Id: I834630d9596d7fb941e2cad5d00ac3af04a537b5
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: https://review.coreboot.org/13808
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
In e820entry struct, the members are defined using
standard types. This can lead to different structure size
when compiling on 32 bit vs. 64 bit environment. This in turn
will affect the size of the struct linux_params.
Using the fixed width types resolves this issue and ensures
that the size of the structures will have the same length
on both 32 and 64 bit systems.
Change-Id: I1869ff2090365731e79b34950446f1791a083d0f
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/13875
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
When linux is used as payload, the parameters to the kernel are build
when cbfstool includes bzImage into the image. Since not all
parameters are used, the unused will stay uninitialized.
There is a chance, that the uninitialized parameters contain
random values. That in turn can lead to early kernel panic.
To avoid it, initialize all parameters with 0 at the beginning.
The ones that are used will be set up as needed and the rest
will contain 0 for sure. This way, kernel can deal with the
provided parameter list the right way.
Change-Id: Id081c24351ec80375255508378b5e1eba2a92e48
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/13874
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Daisuke Nojiri