Le 22/05/2026 à 13:12, David Laight a écrit :
On Fri, 22 May 2026 11:56:02 +0200
"Christophe Leroy (CS GROUP)" <[email protected]> wrote:
Call setup_tm_sigcontexts() before opening user access to avoid
having to close and open again.
Signed-off-by: Christophe Leroy (CS GROUP) <[email protected]>
---
arch/powerpc/kernel/signal_64.c | 22 +++++++++-------------
1 file changed, 9 insertions(+), 13 deletions(-)
diff --git a/arch/powerpc/kernel/signal_64.c b/arch/powerpc/kernel/signal_64.c
index 86bb5bb4c143..3849af21e1d8 100644
--- a/arch/powerpc/kernel/signal_64.c
+++ b/arch/powerpc/kernel/signal_64.c
@@ -873,6 +873,15 @@ int handle_rt_signal64(struct ksignal *ksig, sigset_t *set,
if (!MSR_TM_ACTIVE(msr))
prepare_setup_sigcontext(tsk);
+#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
+ if (MSR_TM_ACTIVE(msr))
Can't that be done without the ugly #ifdef?
I assume MSR_TM_ACTIVE() will be zero - so it will all get optimised away.
Yes but struct rt_sigframe field uc_transact only exists when
CONFIG_PPC_TRANSACTIONAL_MEM is defined.
And that would also require a stub setup_tm_sigcontexts()
Chistophe
-- David
+ err |= setup_tm_sigcontexts(&frame->uc.uc_mcontext,
+ &frame->uc_transact.uc_mcontext,
+ tsk, ksig->sig, NULL,
+ (unsigned
long)ksig->ka.sa.sa_handler,
+ msr);
+
+#endif
if (!user_write_access_begin(frame, sizeof(*frame)))
goto badframe;
@@ -889,19 +898,6 @@ int handle_rt_signal64(struct ksignal *ksig, sigset_t *set,
* ucontext_t (for transactional state) with its uc_link ptr.
*/
unsafe_put_user(&frame->uc_transact, &frame->uc.uc_link,
badframe_block);
-
- user_write_access_end();
-
- err |= setup_tm_sigcontexts(&frame->uc.uc_mcontext,
- &frame->uc_transact.uc_mcontext,
- tsk, ksig->sig, NULL,
- (unsigned
long)ksig->ka.sa.sa_handler,
- msr);
-
- if (!user_write_access_begin(&frame->uc.uc_sigmask,
- sizeof(frame->uc.uc_sigmask)))
- goto badframe;
-
#endif
} else {
unsafe_put_user(0, &frame->uc.uc_link, badframe_block);
