west.yml: update Zephyr to 3.4.99

[kv2019i]

First Zephyr update after 3.4.0 release.

Contains 554 commits, including following directly affecting
SOF targets:

2958a407f8f1 adsp: dmic: Add source clock selection support
0ee64528168b adsp: ace: dmic: Add missing registers definitions
fbb55d1d5e19 adsp: dmic: Moved registers definitions to a separate file
e2e3dc077118 dts: xtensa: intel: add imr entry to cavs25_tgph
6c9a360647c8 drivers: intel_adsp_gpdma: Fix typo in reg name
a8b28f13c195 soc: intel_adsp: cavs: add simple IMR functionality
339b00de11e5 soc: xtensa: intel_adsp: fix memory bank shutdown
5dfaf23f47b7 xtensa: intel_adsp: lnl: Fix dspcs struct
09085ef63c02 dts: xtensa: intel: update cavs25 sram size
b871138fae1f soc/intel_adsp: fix typo in L1EXP definition
9db44fa2813c toolchain: xtensa: enable C11 features with xt-clang
2ba7855ddf23 modules: mipi-syst: support minimal C library

[juimonen]

otherwise good but tgl-h needs this: zephyrproject-rtos/zephyr#59442

[kv2019i]

Ack @juimonen , let's wait for zephyrproject-rtos/zephyr#59442

[kv2019i]

V2:

• a new try with more commits included

  2958a407f8f1 adsp: dmic: Add source clock selection support
  0ee64528168b adsp: ace: dmic: Add missing registers definitions
  fbb55d1d5e19 adsp: dmic: Moved registers definitions to a separate file
  e2e3dc077118 dts: xtensa: intel: add imr entry to cavs25_tgph

[kv2019i]

Nice, new problems. Builds are fine now, but all capture test fail on Intel ACE targets:
https://sof-ci.01.org/sofpr/PR7836/build9930/devicetest/index.html

[marc-hb]

I'm adding a screenshot to https://sof-ci.01.org/sofpr/PR7836/build9930/devicetest/index.html because the regularity is beautiful! And of course useful, demonstrating how the test suite is laser focused on the regression

[kv2019i]

In Intel CI system test, one LNL test fail test_10_gain_peak_vol_api.py::TestGainPeakVolApi::test_01_10 with test file no silent. I think I've seen this before and not sure this is related to this PR. The test in general is passing but in the "mute test" part, there is bit of samples at start. Will rerun the test once to see.

[marc-hb]

In Intel CI system test, one LNL test fail test_10_gain_peak_vol_api.py::TestGainPeakVolApi::test_01_10 with test file no silent.

Please save the URL next time. Now it's practically impossible to find.

[kv2019i]

Rerun results now good, one PM fail in https://sof-ci.01.org/sofpr/PR7836/build9954/devicetest/index.html and a few PM fails in https://sof-ci.01.org/sofpr/PR7836/build9955/devicetest/index.html . Seems we are good to go, merging.