5.15-stable review patch. If anyone has any objections, please let me know.
------------------
From: Benjamin Gray <bgray@xxxxxxxxxxxxx>
[ Upstream commit cc879ab3ce39bc39f9b1d238b283f43a5f6f957d ]
thread_change_pc() uses CPU local data, so must be protected from
swapping CPUs while it is reading the breakpoint struct.
The error is more noticeable after 1e60f3564bad ("powerpc/watchpoints:
Track perf single step directly on the breakpoint"), which added an
unconditional __this_cpu_read() call in thread_change_pc(). However the
existing __this_cpu_read() that runs if a breakpoint does need to be
re-inserted has the same issue.
Signed-off-by: Benjamin Gray <bgray@xxxxxxxxxxxxx>
Signed-off-by: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
Link: https://msgid.link/20230829063457.54157-2-bgray@xxxxxxxxxxxxx
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
arch/powerpc/kernel/hw_breakpoint.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/kernel/hw_breakpoint.c b/arch/powerpc/kernel/hw_breakpoint.c
index 91a3be14808b1..63fec0602af22 100644
--- a/arch/powerpc/kernel/hw_breakpoint.c
+++ b/arch/powerpc/kernel/hw_breakpoint.c
@@ -478,11 +478,13 @@ void thread_change_pc(struct task_struct *tsk, struct pt_regs *regs)
struct arch_hw_breakpoint *info;
int i;
+ preempt_disable();
+
for (i = 0; i < nr_wp_slots(); i++) {
if (unlikely(tsk->thread.last_hit_ubp[i]))
goto reset;
}
- return;
+ goto out;
reset:
regs_set_return_msr(regs, regs->msr & ~MSR_SE);
@@ -491,6 +493,9 @@ void thread_change_pc(struct task_struct *tsk, struct pt_regs *regs)
__set_breakpoint(i, info);
tsk->thread.last_hit_ubp[i] = NULL;
}
+
+out:
+ preempt_enable();
}
static bool is_larx_stcx_instr(int type)
--
2.40.1
