StarMirror/sandboxed-api
1
0
Fork 0
mirror of https://github.com/google/sandboxed-api.git synced 2024-03-22 13:11:30 +08:00
Code Issues Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
Federico Stazi 2020-09-22 15:44:39 +01:00 committed by Federico Stazi
parent 22d21cd1c3
commit 192e443b7c
1 changed files with 0 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
oss-internship-2020/libuv/README.md
View File

@ -60,15 +60,3 @@ Each example and test has an ad-hoc policy implemented on its own file. These po
## Callbacks
The `callbacks.h` and `callbacks.cc` files in the `callbacks` folder implement all the callbacks used by examples and tests.
## Performance
A LibUV program looks like this:
- Setup the loop
- Run the loop
- Close the loop
And the second step (running the loop) is the only performance critical part, because this is were actual I/0 and events handling happens. But since this only involves a single call (to `sapi_uv_run`), the overhead from Sandboxed API is constant with respect to the number of handles and requests.
In other words, after Sandboxed API has called the `uv_run` method on the sandboxed process, whatever happens there is the same as what would have happened in a regular process running LibUV with no sandbox.
This means that, because of LibUV's nature, Sandboxed API has a totally negligible impact on performance. However, there is and overhead caused by `seccomp-bpf` for each syscall.