Repository navigation
Harden APL/GLK hardware support #87
Description
The current driver does not implement all the recommended hardware setups
- 1. firmware download needs to be retried
- 2. [APL][GLK] harden: power-gating and clock gating needs to be disabled during firmware download #776 power-gating and clock gating needs to be disabled during firmware download
- 3. some bits are not reprogrammed on resume
- 4. watchdog and recovery needed
- 5. [APL][GLK] harden: delays after programming registers (and/or loops to check status changes) #777 delays after programming registers (and/or loops to check status changes)
- 6. [APL][GLK] harden: set I2S as slave before enabling DSP #778 set I2S as slave before enabling DSP
- 7. avoid DMA resume capability on capture
Adding a blanket bug to cover all these improvements. Filed as a bug since the current code is not robust enough.
Activity
mengdonglin commentedon Sep 18, 2018
This is an enhancement, so change label from bug to 'enhancement', still P2 to get attention.
pppoe: fix reception of frames with no mac header
ip6_tunnel: be careful when accessing the inner header
net: Fix rtl8192cu build errors on other platforms
ranj063 commentedon Jan 1, 2019
tipc: fix uninit-value in tipc_nl_compat_doit
plbossart commentedon Mar 24, 2019
@wenqingfu can we get someone assigned to this? We know our programming sequences don't fully follow recommended guidelines, let's fix those known opens, at least the opens 2, 5, 6 and keep watchdog/recovery and DAM resume capability aside.
plbossart commentedon Jun 10, 2019
@zhuyingjiang can you clarify what remains to be done for this issue?
zhuyingjiang commentedon Jun 11, 2019
@plbossart The remains are:
4. watchdog and recovery needed
7. avoid DMA resume capability on capture
The 5. #777 delays after programming registers (and/or loops to check status changes)
is almost been done after my PR #1020 #997 is merged.
Still there may need one more PR for "set DUM bit on device power up"
plbossart commentedon Jan 9, 2020
we are not going to support DMA resume, so forking the firmware recovery as a separate issue #1675
Merge pull request thesofproject#87 from adamrk/rust-int-params
reset: zynqmp: replace spaces with tabs
reset: whitespace fixes
bpf: Fix tail_call_reachable rejection for interpreter when jit failed
smb: prevent use-after-free due to open_cached_dir error paths
selftests/bpf: Extend crypto_sanity selftest with invalid dst buffer
btrfs: do proper folio cleanup when run_delalloc_nocow() failed
ipv6: Restore fib6_config validation for SIOCADDRT.
net/sched: sch_codel: Do not call qdisc_tree_reduce_backlog during pe…