From: Marek Majtyka <marekx.majt...@intel.com>
Removal of irq_default_primary_handler usage inside kernel
api causes context switch in atomic inside cm edac driver
for 56xx. IRQ handler is not registered anymore as a threaded
interrupt by default (IRQF_ONESHOT). It is an issue as
wake_up function in RT kernel is not using raw spinlocks.
Signed-off-by: Marek Majtyka <marekx.majt...@intel.com>
---
drivers/edac/axxia_edac-cmc_56xx.c | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/drivers/edac/axxia_edac-cmc_56xx.c
b/drivers/edac/axxia_edac-cmc_56xx.c
index f99c46d..5bc8b41 100644
--- a/drivers/edac/axxia_edac-cmc_56xx.c
+++ b/drivers/edac/axxia_edac-cmc_56xx.c
@@ -182,7 +182,7 @@ static atomic64_t mc_counter = ATOMIC_INIT(0);
* one need to collect dumps for all available cs. Below given example
* for two cs0/cs1.
*
- * CMEM MC cmmon_isr cmmon_wq
+ * CMEM MC cmmon_isr_sw cmmon_wq
* | | |
* | | |
* |ALERT_N - int_status bit [30] |
@@ -679,7 +679,13 @@ error_read:
}
static irqreturn_t
-cmmon_isr(int interrupt, void *device)
+cmmon_isr_hw(int interrupt, void *device)
+{
+ return IRQ_WAKE_THREAD;
+}
+
+static irqreturn_t
+cmmon_isr_sw(int interrupt, void *device)
{
struct intel_edac_dev_info *dev_info = device;
struct cm_56xx_denali_ctl_84 denali_ctl_84;
@@ -1129,8 +1135,8 @@ static int intel_edac_mc_probe(struct platform_device
*pdev)
}
dev_info->data->irq = irq;
- rc = devm_request_irq(&pdev->dev, irq,
- cmmon_isr, IRQF_ONESHOT,
+ rc = devm_request_threaded_irq(&pdev->dev, irq,
+ cmmon_isr_hw, cmmon_isr_sw, IRQF_ONESHOT,
&dev_info->data->irq_name[0], dev_info);
if (rc) {
--
2.7.4
--
_______________________________________________
linux-yocto mailing list
linux-yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/linux-yocto
