luo980/libuv
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

unix: disable io_uring close on selected kernels (#4141)

Browse Source 
Specifically on non-longterm kernels between 5.16.0 (non-longterm) and
6.1.0 (longterm). Starting with longterm 6.1.0, the issue is solved.
This commit is contained in:
Santiago Gimeno 2023-09-17 22:09:00 +02:00 committed by GitHub
parent 4e416266f6
commit c811169f91
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 10 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/unix/linux.c
View File

@ -815,7 +815,9 @@ static void uv__iou_submit(struct uv__iou* iou) {
int uv__iou_fs_close(uv_loop_t* loop, uv_fs_t* req) {
struct uv__io_uring_sqe* sqe;
struct uv__iou* iou;
int kv;
kv = uv__kernel_version();
/* Work around a poorly understood bug in older kernels where closing a file
* descriptor pointing to /foo/bar results in ETXTBSY errors when trying to
* execve("/foo/bar") later on. The bug seems to have been fixed somewhere
@ -823,10 +825,17 @@ int uv__iou_fs_close(uv_loop_t* loop, uv_fs_t* req) {
* but good candidates are the several data race fixes. Interestingly, it
* seems to manifest only when running under Docker so the possibility of
* a Docker bug can't be completely ruled out either. Yay, computers.
* Also, disable on non-longterm versions between 5.16.0 (non-longterm) and
* 6.1.0 (longterm). Starting with longterm 6.1.x, the issue seems to be
* solved.
*/
if (uv__kernel_version() < /* 5.15.90 */ 0x050F5A)
if (kv < /* 5.15.90 */ 0x050F5A)
return 0;
if (kv >= /* 5.16.0 */ 0x050A00 && kv < /* 6.1.0 */ 0x060100)
return 0;
iou = &uv__get_internal_fields(loop)->iou;
sqe = uv__iou_get_sqe(iou, loop, req);