fastrpc: fixes for 7.2
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest5 patches 0 repliesNew
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/5 | misc: fastrpc: Fix initial memory allocation for Audio PD memory pooldrivers/ | +4 | 0 replies | |
| 2/5 | misc: fastrpc: Remove buffer from list prior to unmap operationdrivers/ | +10 | 0 replies | |
| 3/5 | misc: fastrpc: take fl->lock when moving mmaps on interrupted invokedrivers/ | +2 | 0 replies | |
| 4/5 | misc: fastrpc: fix channel ctx ref leak when session alloc failsdrivers/ | +1 | 0 replies | |
| 5/5 | misc: fastrpc: fix memory leak in fastrpc_channel_ctx_freedrivers/ | +1 | 0 replies |
Reviews trailers per patch
All 5 patches have a Reviewed-by or Acked-by.
| Patch | Dmitry Baryshkov | Ekansh Gupta |
|---|---|---|
| 1/5 | R-b | |
| 2/5 | R-b | |
| 3/5 | R-b | |
| 4/5 | R-b | |
| 5/5 | R-b |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH 0/5] fastrpc: fixes for 7.2
From: Srinivas Kandagatla <[email protected]>
Hi Greg,
This series collects a handful of fastrpc fixes that have accumulated on the list. All patches are small, target long-standing issues, and are tagged for stable.
The fixes fall into two buckets:
Memory leaks:
- Audio PD memory pool never registered its initial buffer because pageslen was left at zero, so the pool was always empty and every allocation fell back to the remote heap (patch 1).
- fastrpc_device_open() takes a channel ctx reference before allocating a session; the -EBUSY path on session-alloc failure never dropped it (patch 4).
- fastrpc_channel_ctx_free() never destroyed ctx_idr, leaking the IDR backing storage on channel teardown (patch 5).
Locking / list corruption:
- fastrpc_req_munmap() removed the buffer from fl->mmaps only after the DSP unmap returned, allowing two concurrent unmaps to race on the same entry. Detach the buffer under fl->lock first and re-add it if the DSP call fails (patch 2).
- The -ERESTARTSYS path in fastrpc_internal_invoke() walked fl->mmaps and spliced it onto cctx->invoke_interrupted_mmaps without holding fl->lock, racing with every other mmaps accessor (patch 3).
Please queue for 7.2.
Thanks,
Srini
Anandu Krishnan E (1):
misc: fastrpc: fix channel ctx ref leak when session alloc fails
Eddie Lin (1):
misc: fastrpc: fix memory leak in fastrpc_channel_ctx_free
Ekansh Gupta (2):
misc: fastrpc: Fix initial memory allocation for Audio PD memory pool
misc: fastrpc: Remove buffer from list prior to unmap operation
Junrui Luo (1):
misc: fastrpc: take fl->lock when moving mmaps on interrupted invoke
drivers/misc/fastrpc.c | 27 ++++++++++++++++++--------- 1 file changed, 18 insertions(+), 9 deletions(-)
--
2.53.0