# Re: \[PATCH 6/9\] kernel/api: add API specification for sys\_close

- From: Sasha Levin \<sashal@kernel.org\>
- Date: 2026-03-13 16:46:02 
- List: linux-api
- Page: https://ratatoskr.run/linux-api/2026/03/343400
- Thread: https://ratatoskr.run/linux-api/2026/03/343369/t.md

```text
On Fri, Mar 13, 2026 at 04:49:28PM +0100, Greg Kroah-Hartman wrote:
>On Fri, Mar 13, 2026 at 11:09:16AM -0400, Sasha Levin wrote:
>> + *   Calling close() on a file descriptor while another thread is using it
>> + *   (e.g., in a blocking read() or write()) has implementation-defined
>> + *   behavior. On Linux, the blocked operation continues on the underlying
>> + *   file and may complete even after close() returns.
>
>I'm guessing this came from the man pages?  This is Linux, so we are the
>"implementation" here :)

Right :)

I was just trying to make it a comparison to posix. I'll clarify the docs here.

-- 
Thanks,
Sasha
```