west.yml: update Zephyr to ceff785434b

[kv2019i]

Patches affecting SOF:

fb54eb5ab133 llext: symbol.h: Explicit typecast in EXPORT_SYMBOL macros
ce243944375e llext: add object test case
33c6bdab9ad1 llext: add STT_OBJECT relocation
41b3c7158687 drivers: dma: intel_adsp_hda: Fix L1 exit condition
5cac834bb651 llext: arch_elf_relocate: pass opval unmodified

[kv2019i]

Submitting this right away. This might trigger CI fails as #8936 is not yet merged, but let's see the results, we may need to merge that on an expedited schedule.

For Intel platforms, this brings potential fixes to #8770 and #8761 . But let's see the results.

[marc-hb]

ACE all green except MTLP_RVP_HDA unavailable in https://sof-ci.01.org/sofpr/PR8954/build3460/devicetest/index.html

CAVS https://sof-ci.01.org/sofpr/PR8954/build3461/devicetest/index.html has a DSP panic which is probably multicore panic #8942. But multicore was just turned off by 6bdbf71 (after this run)

Re-running these tests.

[marc-hb]

SOFCI TEST

EDIT: Known pause/resume issues in ACE https://sof-ci.01.org/sofpr/PR8954/build3481/devicetest/index.html, everything else green.

Only one suspend/resume failure TIMEOUT in https://sof-ci.01.org/sofpr/PR8954/build3482/devicetest/index.html?model=TGLU_RVP_NOCODEC-ipc4&testcase=check-suspend-resume-with-capture, everything else green.

[kv2019i]

One system pm fail in https://sof-ci.01.org/sofpr/PR8954/build3482/devicetest/index.html

In https://sof-ci.01.org/sofpr/PR8954/build3481/devicetest/index.html
There's a MOD_SET_DX time out which matches fingerprints of #8642 . So not a new issue. Proceeding with merge.