linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [RFC PATCH v2] ACPI: Move sdei_init and ghes_init ahead
@ 2021-11-14 12:33 Shuai Xue
  0 siblings, 0 replies; only message in thread
From: Shuai Xue @ 2021-11-14 12:33 UTC (permalink / raw)
  To: linux-kernel, linux-acpi, linux-arm-kernel, linux-pci
  Cc: bp, tony.luck, james.morse, lenb, rjw, bhelgaas, xueshuai,
	zhangliguang, zhuo.song

On an ACPI system, ACPI is initialised very early from a
subsys_initcall(), while SDEI is not ready until a subsys_initcall().
More seriously, the kernel is able to handle and report errors until the
GHES is initialised by device_initcall().

Consequently, when an error occurs during the kernel booting, the
phyiscal sdei dispatcher in firmware fails to dispatch error events. All
errors that occurred before GHES initialization are missed and there is
no chance to report and find them again.

In this patch, move sdei_init and ghes_init as far ahead as possible,
right after acpi_hest_init().

Signed-off-by: Shuai Xue <xueshuai@linux.alibaba.com>

---
Changelog v1 -> v2:
Fix compile error without CONFIG_ACPI_APEI enabled
Reported-by: kernel test robot<lkp@intel.com>
---
 drivers/acpi/apei/ghes.c    | 3 +--
 drivers/acpi/pci_root.c     | 2 ++
 drivers/firmware/arm_sdei.c | 9 +--------
 include/acpi/apei.h         | 4 ++++
 4 files changed, 8 insertions(+), 10 deletions(-)

diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
index 0c8330ed1ffd..4200369503b8 100644
--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -1457,7 +1457,7 @@ static struct platform_driver ghes_platform_driver = {
 	.remove		= ghes_remove,
 };
 
-static int __init ghes_init(void)
+int __init ghes_init(void)
 {
 	int rc;
 
@@ -1499,4 +1499,3 @@ static int __init ghes_init(void)
 err:
 	return rc;
 }
-device_initcall(ghes_init);
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index ab2f7dfb0c44..658b6e536b60 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -946,6 +946,8 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
 void __init acpi_pci_root_init(void)
 {
 	acpi_hest_init();
+	sdei_init();
+	ghes_init();
 	if (acpi_pci_disabled)
 		return;
 
diff --git a/drivers/firmware/arm_sdei.c b/drivers/firmware/arm_sdei.c
index a7e762c352f9..606520be326e 100644
--- a/drivers/firmware/arm_sdei.c
+++ b/drivers/firmware/arm_sdei.c
@@ -1059,7 +1059,7 @@ static bool __init sdei_present_acpi(void)
 	return true;
 }
 
-static int __init sdei_init(void)
+int __init sdei_init(void)
 {
 	struct platform_device *pdev;
 	int ret;
@@ -1080,13 +1080,6 @@ static int __init sdei_init(void)
 	return ret;
 }
 
-/*
- * On an ACPI system SDEI needs to be ready before HEST:GHES tries to register
- * its events. ACPI is initialised from a subsys_initcall(), GHES is initialised
- * by device_initcall(). We want to be called in the middle.
- */
-subsys_initcall_sync(sdei_init);
-
 int sdei_event_handler(struct pt_regs *regs,
 		       struct sdei_registered_event *arg)
 {
diff --git a/include/acpi/apei.h b/include/acpi/apei.h
index ece0a8af2bae..12909c96ef89 100644
--- a/include/acpi/apei.h
+++ b/include/acpi/apei.h
@@ -33,8 +33,12 @@ extern bool ghes_disable;
 
 #ifdef CONFIG_ACPI_APEI
 void __init acpi_hest_init(void);
+int __init sdei_init(void);
+int __init ghes_init(void);
 #else
 static inline void acpi_hest_init(void) { return; }
+static inline void sdei_init(void) { return; }
+static inline void ghes_init(void) { return; }
 #endif
 
 int erst_write(const struct cper_record_header *record);
-- 
2.20.1.12.g72788fdb


^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2021-11-14 12:34 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-11-14 12:33 [RFC PATCH v2] ACPI: Move sdei_init and ghes_init ahead Shuai Xue

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).