summaryrefslogtreecommitdiff
path: root/arch/powerpc/kernel/irq.c
diff options
context:
space:
mode:
authorBenjamin Herrenschmidt <benh@kernel.crashing.org>2012-07-10 18:36:40 +1000
committerBenjamin Herrenschmidt <benh@kernel.crashing.org>2012-07-10 19:16:07 +1000
commitbe2cf20a5ad31ebb13562c1c866ecc626fbd721e (patch)
tree939537de7ac05fc80b7b8c3588227c2b557b1687 /arch/powerpc/kernel/irq.c
parent2437fccfbfc83bcb868ccc7fdfe2b5310bf07835 (diff)
powerpc: More fixes for lazy IRQ vs. idle
Looks like we still have issues with pSeries and Cell idle code vs. the lazy irq state. In fact, the reset fixes that went upstream are exposing the problem more by causing BUG_ON() to trigger (which this patch turns into a WARN_ON instead). We need to be careful when using a variant of low power state that has the side effect of turning interrupts back on, to properly set all the SW & lazy state to look as if everything is enabled before we enter the low power state with MSR:EE off as we will return with MSR:EE on. If not, we have a discrepancy of state which can cause things to go very wrong later on. This patch moves the logic into a helper and uses it from the pseries and cell idle code. The power4/970 idle code already got things right (in assembly even !) so I'm not touching it. The power7 "bare metal" idle code is subtly different and correct. Remains PA6T and some hypervisor based Cell platforms which have questionable code in there, but they are mostly dead platforms so I'll fix them when I manage to get final answers from the respective maintainers about how the low power state actually works on them. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> CC: stable@vger.kernel.org [v3.4]
Diffstat (limited to 'arch/powerpc/kernel/irq.c')
-rw-r--r--arch/powerpc/kernel/irq.c46
1 files changed, 46 insertions, 0 deletions
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 1b415027ec0e..9270a399c9d6 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -286,6 +286,52 @@ void notrace restore_interrupts(void)
__hard_irq_enable();
}
+/*
+ * This is a helper to use when about to go into idle low-power
+ * when the latter has the side effect of re-enabling interrupts
+ * (such as calling H_CEDE under pHyp).
+ *
+ * You call this function with interrupts soft-disabled (this is
+ * already the case when ppc_md.power_save is called). The function
+ * will return whether to enter power save or just return.
+ *
+ * In the former case, it will have notified lockdep of interrupts
+ * being re-enabled and generally sanitized the lazy irq state,
+ * and in the latter case it will leave with interrupts hard
+ * disabled and marked as such, so the local_irq_enable() call
+ * in cpu_idle() will properly re-enable everything.
+ */
+bool prep_irq_for_idle(void)
+{
+ /*
+ * First we need to hard disable to ensure no interrupt
+ * occurs before we effectively enter the low power state
+ */
+ hard_irq_disable();
+
+ /*
+ * If anything happened while we were soft-disabled,
+ * we return now and do not enter the low power state.
+ */
+ if (lazy_irq_pending())
+ return false;
+
+ /* Tell lockdep we are about to re-enable */
+ trace_hardirqs_on();
+
+ /*
+ * Mark interrupts as soft-enabled and clear the
+ * PACA_IRQ_HARD_DIS from the pending mask since we
+ * are about to hard enable as well as a side effect
+ * of entering the low power state.
+ */
+ local_paca->irq_happened &= ~PACA_IRQ_HARD_DIS;
+ local_paca->soft_enabled = 1;
+
+ /* Tell the caller to enter the low power state */
+ return true;
+}
+
#endif /* CONFIG_PPC64 */
int arch_show_interrupts(struct seq_file *p, int prec)