Skip to content

[lwp][rv64] fix potential signal handler infinite loop #10500

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 23, 2025
Merged

[lwp][rv64] fix potential signal handler infinite loop #10500

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 12 additions & 2 deletions components/lwp/arch/risc-v/rv64/lwp_gcc.S
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,12 +75,20 @@ arch_crt_start_umode:
* we handle following things here:
* 1. restoring user mode debug state (not support yet)
* 2. handling thread's exit request
* 3. handling POSIX signal
* 3. handling POSIX signal (skipped for signal quit path)
* 4. restoring user context
* 5. jump to user mode
*/
.global arch_ret_to_user
arch_ret_to_user:
li s0, 1 // flag=1 (normal path)
j arch_ret_to_user_impl

.global arch_signal_quit_ret_to_user
arch_signal_quit_ret_to_user:
li s0, 0 // flag=0 (signal quit path)

arch_ret_to_user_impl:
// TODO: we don't support kernel gdb server in risc-v yet
// so we don't check debug state here and handle debugging bussiness

Expand All @@ -90,6 +98,8 @@ arch_ret_to_user:
call sys_exit

1:
// Skip signal handling if coming from arch_signal_quit
beqz s0, ret_to_user_exit
mv a0, sp
call lwp_thread_signal_catch

Expand Down Expand Up @@ -119,7 +129,7 @@ arch_signal_quit:

RESTORE_ALL
SAVE_ALL
j arch_ret_to_user
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

可以结合下aarch64的情况，查看下这部分如何处理比较合适

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

您好，aarch64和riscv64的处理都是类似的，避免在arch_signal_quit->arch_ret_to_user->lwp_thread_signal_catch 这条调用链上使用lwp_thread_signal_catch 处理信号即可，其余部分保持不变。我尝试了三种可能的修改方案：

1 通过给arch_ret_to_user传入一个额外参数已指示是否需要调用lwp_thread_signal_catch，这种方法可以让线程通过统一的arch_ret_to_user返回用户态。但是所有调用arch_ret_to_user的地方都需要额外传入参数，实现示例链接：https://github.com/RT-Thread/rt-thread/compare/master...eatvector:rt-thread:demo0?expand=1

2 单独针对arch_signal_quit实现一个独立的返回用户态调用接口，与之前提交的方案类似，不过复用了arch_ret_to_user的代码已减少重复代码，这种方案应该是在riscv64和aarch64上最清晰，修改也最少的实现，示例链接：https://github.com/RT-Thread/rt-thread/compare/master...eatvector:rt-thread:demo1?expand=1

3 在线程结构体内部增加额外标记已指示线程是否通过arch_signal_quit返回用户态，如果该标志被设置，则在lwp_thread_signal_catch 不进行任何处理，这种方法所需修改的架构相关内容最少，但是得修改相关的c文件，以及相关结构体，实现示例：https://github.com/RT-Thread/rt-thread/compare/master...eatvector:rt-thread:demo3?expand=1

不知道您认为哪种修改方案更好，或者有其他更好的建议吗？

j arch_signal_quit_ret_to_user

/**
* rt_noreturn
Expand Down