Support MediaTek mt8188 platform #6796
Conversation
Add xtensa headers for mtk mt8188 platform. Signed-off-by: Tinghan Shen <[email protected]>
tinghan-shen
requested review from
yaochunhung,
kuanhsuncheng and
wenliangwu
tinghan-shen
requested review from
akloniex,
marc-hb,
aborisovich,
lgirdwood,
plbossart,
mmaka1,
lbetlej and
dbaluta
as code owners
tinghan-shen
force-pushed
the
mt8188-to-sof
branch
2 times, most recently
from
d4a9fca
to
44087f2
Compare
|
Getting rid of the SPDX / C99 warnings is tricky. Last time I tried I think I found a way: |
scripts/xtensa-build-all.sh
Outdated
| @@ -11,7 +11,7 @@ DEFAULT_PLATFORMS=( | |||
| tgl tgl-h | |||
| imx8 imx8x imx8m imx8ulp | |||
| rn rmb | |||
| mt8186 mt8195 | |||
| mt8186 mt8188 mt8195 | |||
There was a problem hiding this comment.
The container is not updated automatically (more container info in #5073). In the mean time try this instead to avoid CI failure like the one below
--- a/scripts/xtensa-build-all.sh
+++ b/scripts/xtensa-build-all.sh
@@ -18,11 +18,17 @@ DEFAULT_PLATFORMS=(
# the -a option for everyone.
SUPPORTED_PLATFORMS=( "${DEFAULT_PLATFORMS[@]}" )
+# Waiting for container work in progress
+SUPPORTED_PLATFORMS=( mt8188 )
+
+
# Not actually "supported" in the main branch anymore (go to stable-v2.3
https://github.com/thesofproject/sof/actions/runs/3692264763/jobs/6250986683
xtensa-mt8188-elf-gcc
is not a full path and was not found in the PATH.
| @@ -0,0 +1,16 @@ | |||
| CONFIG_MT8188=y | |||
| CONFIG_CORE_COUNT=1 | |||
| CONFIG_XT_VIRTUAL_OPS=1 | |||
There was a problem hiding this comment.
Don't need this configuration "CONFIG_XT_VIRTUAL_OPS" because new code support virtual ops by default.
94db5c7#diff-1ee8aded2f65c24944261dcab296c3ff94e0d3bacd5909895db9c63f029113c5R1085
src/platform/mt8188/lib/clk.c
Outdated
|
|
||
| DECLARE_TR_CTX(clkdrv_tr, SOF_UUID(clkdrv_uuid), LOG_LEVEL_INFO); | ||
|
|
||
| /* default voltage is 0.8V */ |
There was a problem hiding this comment.
fixed 0.75V for mt8188.
src/platform/mt8188/platform.c
Outdated
| * for more details. | ||
| */ | ||
| const struct xthal_MPU_entry __xt_mpu_init_table[] __section(".ResetVector.text") = { | ||
| XTHAL_MPU_ENTRY(0x00000000, 1, XTHAL_AR_RWXrwx, XTHAL_MEM_DEVICE), // unused |
There was a problem hiding this comment.
Is it really unused because memory attribute is not XTHAL_AR_NONE?
tinghan-shen
force-pushed
the
mt8188-to-sof
branch
from
44087f2
to
5dbd425
Compare
Add memory layout and register address for mtk mt8188. [Cache] I-Cache: 64KB, 4-way Associativity D-Cache: 128KB, 4-way Associativity [Memory] DRAM: DSP can access DRAM shared with CPU L2TCM: 512KB DSP SRAM POOL Signed-off-by: Tinghan Shen <[email protected]>
Reuse mt8186 interrupt, timer, and ipc drivers for mt8188. mt8188 DSP has 32 interrupts, 4 internal timers, and 5 mailbox for IPC. Signed-off-by: Tinghan Shen <[email protected]>
tinghan-shen
force-pushed
the
mt8188-to-sof
branch
from
5dbd425
to
9052080
Compare
tinghan-shen
force-pushed
the
mt8188-to-sof
branch
from
9052080
to
a37bca7
Compare
mt8188 platform integrates a single-core HIFI5 DSP. The highest DSP operation frequency is 800MHz which requires 0.75v working voltage. otherwise, it should switch to 26M which can operate at the lowest working voltage 0.55v. Signed-off-by: Tinghan Shen <[email protected]>
Add default config for mt8188 platform. Signed-off-by: Tinghan Shen <[email protected]>
Add CMakefile for MediaTek mt8188 platform to build it. Signed-off-by: Tinghan Shen <[email protected]>
Add mt8188 to Kconfig. Signed-off-by: Tinghan Shen <[email protected]>
The company name should be MediaTek instead of Mediatek.
^ ^
Signed-off-by: Tinghan Shen <[email protected]>
tinghan-shen
force-pushed
the
mt8188-to-sof
branch
from
a37bca7
to
9535aa7
Compare
|
@wszypelt can you check CI, not expecting this PR to fail. Thanks! |
|
@lgirdwood |
|
@greg-intel @marc-hb CI sync appears sluggish, seeing blob not found with Jenkins. |
|
@lgirdwood The problem was on the side of the internal service, it has already been fixed |
|
checkpatch looks at each commit individually. So fixing checkpatch warnings in a separate commit is pointless. |
|
Hi marc, The SPDX checkpatch warnings is fixed in each commit. The commit (9535aa7) only includes files fixed the company name. |
|
Oh I see now that commit c916f3f (= 9535aa7 after rebase and merge) cleaned up old existing files not added by this PR, sorry I got confused. So it looks like no commit in this PR has SPDX or comment style warning, my bad. That cleanup is good for consistency but it wasn't really necessary, at least not in this PR. I wasn't asking for it! As I wrote checkpatch is focused on incoming changes, checkpatch does not care about existing code. |
Pleasing checkpatch is hard when adding new files. This is tricky and comes up every time someone adds new files, examples in thesofproject#6284, thesofproject#6796, thesofproject#6931 , etc. Signed-off-by: Marc Herbert <[email protected]>
Pleasing checkpatch is hard when adding new files. This is tricky and comes up every time someone adds new files, examples in #6284, #6796, #6931 , etc. Signed-off-by: Marc Herbert <[email protected]>
Add mt8188 platform support in SOF project.
mt8188: Cadence HiFi-5 DSP, single core, 4 dais , 4 dmas and 48k 16 bits.
Please let me know if there is any questions.
Related PRs for mt8188,
rimage: thesofproject/rimage#122
xtensa-overlay: thesofproject/xtensa-overlay#16
crosstool-ng gcc-10x: thesofproject/crosstool-ng#17
crosstool-ng gcc-11x: thesofproject/crosstool-ng#18
Thank you!