Konrad Sztyber 10cb404a1a bdev/compress: limit operations performed on reduce_thread ... Previously, it was possible to execute spdk_bdev_io_get_buf() on a different thread than the one that the IO was submitted on, which is unsafe. Now, the buffers are always allocated on the correct thread and we do spdk_thread_send_msg() only for the reduce library functions. Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com> Change-Id: I50bf9aa3054073636204ec3a147d464f194a8c8f Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15667 Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com> Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot		2022-12-05 09:51:26 +00:00
..
Makefile	update Intel copyright notices	2022-11-10 08:28:53 +00:00
vbdev_compress_rpc.c	update Intel copyright notices	2022-11-10 08:28:53 +00:00
vbdev_compress.c	bdev/compress: limit operations performed on reduce_thread	2022-12-05 09:51:26 +00:00
vbdev_compress.h	update Intel copyright notices	2022-11-10 08:28:53 +00:00