Re: [RFC PATCH 09/14] powerpc/tm: Warn if state is transactional
From: Michael Neuling <hidden>
Date: 2018-11-15 02:51:05
On Tue, 2018-11-06 at 10:40 -0200, Breno Leitao wrote:
quoted hunk ↗ jump to hunk
Since every kernel entrance is calling TM_KERNEL_ENTRY, it is not expected to arrive at this point with a suspended transaction. If that is the case, cause a warning and reclaim the current thread in order to avoid a TM Bad Thing. Signed-off-by: Breno Leitao <leitao@debian.org> --- arch/powerpc/kernel/process.c | 7 +++---- arch/powerpc/kernel/signal.c | 2 +- 2 files changed, 4 insertions(+), 5 deletions(-)diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 73872f751b33..849591bf0881 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c@@ -1752,11 +1752,10 @@ void start_thread(struct pt_regs *regs, unsigned longstart, unsigned long sp) #ifdef CONFIG_PPC_TRANSACTIONAL_MEM /* - * Clear any transactional state, we're exec()ing. The cause is - * not important as there will never be a recheckpoint so it's not - * user visible. + * It is a bug if the transaction was not reclaimed until this + * point. Warn us and try to workaround it calling tm_reclaim(). */ - if (MSR_TM_SUSPENDED(mfmsr())) + if (WARN_ON(MSR_TM_SUSPENDED(mfmsr()))) tm_reclaim_current(0); #endif
Let's turn these into BUG_ON() Mikey
quoted hunk ↗ jump to hunk
diff --git a/arch/powerpc/kernel/signal.c b/arch/powerpc/kernel/signal.c index b3e8db376ecd..cbaccc2be0fb 100644 --- a/arch/powerpc/kernel/signal.c +++ b/arch/powerpc/kernel/signal.c@@ -203,7 +203,7 @@ unsigned long get_tm_stackpointer(struct task_struct *tsk) #ifdef CONFIG_PPC_TRANSACTIONAL_MEM BUG_ON(tsk != current); - if (MSR_TM_ACTIVE(tsk->thread.regs->msr)) { + if (WARN_ON(MSR_TM_ACTIVE(mfmsr()))) { tm_reclaim_current(TM_CAUSE_SIGNAL); if (MSR_TM_TRANSACTIONAL(tsk->thread.regs->msr)) return tsk->thread.ckpt_regs.gpr[1];