EDAC: Remove edac_stub.c
Subsystems (8)
Borislav Petkov
maintainerPATCH 8/11 EDAC: Remove EDAC_MM_EDAC
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Move all the EDAC core functionality behind CONFIG_EDAC and get rid of that indirection. Update defconfigs which had it.
While at it, fix dependencies such that EDAC depends on RAS for the tracepoints.
--- arch/arm/configs/multi_v7_defconfig | 1 - arch/arm/configs/pxa_defconfig | 3 +- arch/powerpc/configs/85xx-hw.config | 3 +- arch/powerpc/configs/85xx/ge_imp3a_defconfig | 1 - arch/powerpc/configs/85xx/xes_mpc85xx_defconfig | 1 - arch/powerpc/configs/cell_defconfig | 1 - arch/powerpc/configs/pasemi_defconfig | 1 - arch/powerpc/configs/ppc64_defconfig | 1 - arch/powerpc/configs/ppc64e_defconfig | 1 - arch/powerpc/configs/ppc6xx_defconfig | 3 +- arch/tile/configs/tilegx_defconfig | 1 - arch/tile/configs/tilepro_defconfig | 1 - drivers/acpi/Kconfig | 1 - drivers/edac/Kconfig | 101 ++++++++++-------------- drivers/edac/Makefile | 3 +- drivers/edac/edac_stub.c | 2 +- 16 files changed, 48 insertions(+), 77 deletions(-)
arch/arm/configs/multi_v7_defconfig +0 -1
@@ -748,7 +748,6 @@ CONFIG_LEDS_TRIGGER_DEFAULT_ON=y CONFIG_LEDS_TRIGGER_TRANSIENT=y CONFIG_LEDS_TRIGGER_CAMERA=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_HIGHBANK_MC=y CONFIG_EDAC_HIGHBANK_L2=y CONFIG_RTC_CLASS=y
arch/arm/configs/pxa_defconfig +1 -2
@@ -635,8 +635,7 @@ CONFIG_LEDS_TRIGGER_GPIO=m CONFIG_LEDS_TRIGGER_DEFAULT_ON=m CONFIG_LEDS_TRIGGER_TRANSIENT=m CONFIG_LEDS_TRIGGER_CAMERA=m -CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=m +CONFIG_EDAC=m CONFIG_RTC_CLASS=y CONFIG_RTC_DEBUG=y CONFIG_RTC_DRV_DS1307=m
arch/powerpc/configs/85xx-hw.config +1 -2
@@ -16,9 +16,8 @@ CONFIG_DAVICOM_PHY=y CONFIG_DMADEVICES=y CONFIG_E1000E=y CONFIG_E1000=y -CONFIG_EDAC_MM_EDAC=y -CONFIG_EDAC_MPC85XX=y CONFIG_EDAC=y +CONFIG_EDAC_MPC85XX=y CONFIG_EEPROM_AT24=y CONFIG_EEPROM_LEGACY=y CONFIG_FB_FSL_DIU=y
arch/powerpc/configs/85xx/ge_imp3a_defconfig +0 -1
@@ -155,7 +155,6 @@ CONFIG_USB_OHCI_HCD_PPC_OF_BE=y CONFIG_USB_OHCI_HCD_PPC_OF_LE=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_MPC85XX=y CONFIG_RTC_CLASS=y # CONFIG_RTC_INTF_PROC is not set
arch/powerpc/configs/85xx/xes_mpc85xx_defconfig +0 -1
@@ -116,7 +116,6 @@ CONFIG_LEDS_TRIGGERS=y CONFIG_LEDS_TRIGGER_TIMER=y CONFIG_LEDS_TRIGGER_HEARTBEAT=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y CONFIG_RTC_DRV_CMOS=y
arch/powerpc/configs/cell_defconfig +0 -1
@@ -179,7 +179,6 @@ CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_IPOIB=m CONFIG_INFINIBAND_IPOIB_DEBUG_DATA=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_CELL=y CONFIG_UIO=m CONFIG_EXT2_FS=y
arch/powerpc/configs/pasemi_defconfig +0 -1
@@ -142,7 +142,6 @@ CONFIG_USB_UHCI_HCD=y CONFIG_USB_SL811_HCD=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_PASEMI=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y
arch/powerpc/configs/ppc64_defconfig +0 -1
@@ -262,7 +262,6 @@ CONFIG_INFINIBAND_IPOIB_CM=y CONFIG_INFINIBAND_SRP=m CONFIG_INFINIBAND_ISER=m CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_PASEMI=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y
arch/powerpc/configs/ppc64e_defconfig +0 -1
@@ -173,7 +173,6 @@ CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_IPOIB=m CONFIG_INFINIBAND_ISER=m CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y CONFIG_FS_DAX=y
arch/powerpc/configs/ppc6xx_defconfig +1 -2
@@ -988,8 +988,7 @@ CONFIG_LEDS_TRIGGER_BACKLIGHT=m CONFIG_LEDS_TRIGGER_DEFAULT_ON=m CONFIG_ACCESSIBILITY=y CONFIG_A11Y_BRAILLE_CONSOLE=y -CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=m +CONFIG_EDAC=m CONFIG_RTC_CLASS=y # CONFIG_RTC_HCTOSYS is not set CONFIG_RTC_DRV_DS1307=m
arch/tile/configs/tilegx_defconfig +0 -1
@@ -249,7 +249,6 @@ CONFIG_USB_EHCI_HCD=y CONFIG_USB_OHCI_HCD=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_TILE=y CONFIG_EXT2_FS=y
arch/tile/configs/tilepro_defconfig +0 -1
@@ -358,7 +358,6 @@ CONFIG_WATCHDOG_NOWAYOUT=y # CONFIG_VGA_ARB is not set # CONFIG_USB_SUPPORT is not set CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_TILE=y CONFIG_EXT2_FS=y
drivers/acpi/Kconfig +0 -1
@@ -471,7 +471,6 @@ config ACPI_EXTLOG tristate "Extended Error Log support" depends on X86_MCE && X86_LOCAL_APIC && EDAC select UEFI_CPER - select RAS default n help Certain usages such as Predictive Failure Analysis (PFA) require
drivers/edac/Kconfig +43 -58
361 lines not on this page · raw message
drivers/edac/Makefile +1 -2
@@ -6,8 +6,7 @@ # GNU General Public License. # -obj-$(CONFIG_EDAC) := edac_stub.o -obj-$(CONFIG_EDAC_MM_EDAC) += edac_core.o +obj-$(CONFIG_EDAC) := edac_stub.o edac_core.o edac_core-y := edac_mc.o edac_device.o edac_mc_sysfs.o edac_core-y += edac_module.o edac_device_sysfs.o wq.o
drivers/edac/edac_stub.c +1 -1
@@ -20,7 +20,7 @@int edac_report_status =EDAC_REPORTING_ENABLED ;EXPORT_SYMBOL_GPL (edac_report_status); -static int __init edac_report_setup (char *str) +static int __init __maybe_unused edac_report_setup (char *str) {if (!str)return -EINVAL ;
Signature · 1 line
2.11.0
Borislav Petkov
maintainercover EDAC: Remove edac_stub.c
From: Borislav Petkov <[email protected]>
This was an ancient hack to hand in hw errors reported through an NMI into EDAC. And what a hack it was - it required a piece of the EDAC functionality to be always built-in and it called into it from NMI context.
We could've simply converted it to a notifier and put some lipstick on the pig but the two drivers which used it are for hardware which is pretty much obsolete now. So we can outright remove it - old hardware can stick with old software if they really really need it.
Borislav Petkov (11):
EDAC, highbank: Align Makefile directives
x86/nmi, EDAC: Get rid of DRAM error reporting thru PCI SERR NMI
EDAC: Get rid of edac_handlers
EDAC: Remove edac_err_assert
EDAC: Move edac_op_state to edac_mc.c
ACPI/extlog: Add EDAC dependency
EDAC: Issue tracepoint only when it is defined
EDAC: Remove EDAC_MM_EDAC
EDAC: Update Kconfig help text
EDAC: Delete edac_stub.c
EDAC: Rename report status accessors
arch/arm/configs/multi_v7_defconfig | 1 - arch/arm/configs/pxa_defconfig | 3 +- arch/powerpc/configs/85xx-hw.config | 3 +- arch/powerpc/configs/85xx/ge_imp3a_defconfig | 1 - arch/powerpc/configs/85xx/xes_mpc85xx_defconfig | 1 - arch/powerpc/configs/cell_defconfig | 1 - arch/powerpc/configs/pasemi_defconfig | 1 - arch/powerpc/configs/ppc64_defconfig | 1 - arch/powerpc/configs/ppc64e_defconfig | 1 - arch/powerpc/configs/ppc6xx_defconfig | 3 +- arch/tile/configs/tilegx_defconfig | 1 - arch/tile/configs/tilepro_defconfig | 1 - arch/x86/kernel/nmi.c | 11 --- drivers/acpi/Kconfig | 3 +- drivers/acpi/acpi_extlog.c | 8 +- drivers/edac/Kconfig | 119 ++++++++++-------------- drivers/edac/Makefile | 7 +- drivers/edac/edac_mc.c | 99 +++++++++++++++----- drivers/edac/edac_stub.c | 68 -------------- drivers/edac/pnd2_edac.c | 2 +- drivers/edac/sb_edac.c | 4 +- drivers/edac/skx_edac.c | 2 +- include/linux/edac.h | 30 +----- 23 files changed, 139 insertions(+), 232 deletions(-) delete mode 100644 drivers/edac/edac_stub.c
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 5/11 EDAC: Move edac_op_state to edac_mc.c
From: Borislav Petkov <[email protected]>
... as part of moving stuff away from edac_stub.c
--- drivers/edac/edac_mc.c | 3 +++ drivers/edac/edac_stub.c | 3 --- 2 files changed, 3 insertions(+), 3 deletions(-)
drivers/edac/edac_mc.c +3 -0
@@ -40,6 +40,9 @@#define edac_atomic_scrub (va, size) do { } while (0)#endif +int edac_op_state =EDAC_OPSTATE_INVAL ; +EXPORT_SYMBOL_GPL (edac_op_state); +/* lock to memory controller's control array */ static DEFINE_MUTEX (mem_ctls_mutex);static LIST_HEAD (mc_devices);
drivers/edac/edac_stub.c +0 -3
@@ -17,9 +17,6 @@#include <linux/atomic.h> #include <linux/device.h> -int edac_op_state =EDAC_OPSTATE_INVAL ; -EXPORT_SYMBOL_GPL (edac_op_state); -int edac_report_status =EDAC_REPORTING_ENABLED ;EXPORT_SYMBOL_GPL (edac_report_status);
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 4/11 EDAC: Remove edac_err_assert
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
... and the glue around it. It is not needed anymore.
--- drivers/edac/edac_mc.c | 18 +----------------- drivers/edac/edac_stub.c | 3 --- include/linux/edac.h | 1 - 3 files changed, 1 insertion(+), 21 deletions(-)
drivers/edac/edac_mc.c +1 -17
@@ -505,22 +505,6 @@ struct mem_ctl_info *find_mci_by_dev(struct device *dev)EXPORT_SYMBOL_GPL (find_mci_by_dev);/* -* handler for EDAC to check if NMI type handler has asserted interrupt -*/ -static int edac_mc_assert_error_check_and_clear (void ) -{ -int old_state; - -if (edac_op_state ==EDAC_OPSTATE_POLL ) -return 1 ; - - old_state = edac_err_assert; - edac_err_assert =0 ; - -return old_state; -} - -/* * edac_mc_workq_function * performs the operation scheduled by a workq request */
@@ -536,7 +520,7 @@ static void edac_mc_workq_function(struct work_struct *work_req)return ; } -if (edac_mc_assert_error_check_and_clear ()) +if (edac_op_state ==EDAC_OPSTATE_POLL ) mci->edac_check (mci);mutex_unlock (&mem_ctls_mutex);
drivers/edac/edac_stub.c +0 -3
@@ -20,9 +20,6 @@int edac_op_state =EDAC_OPSTATE_INVAL ;EXPORT_SYMBOL_GPL (edac_op_state); -int edac_err_assert =0 ; -EXPORT_SYMBOL_GPL (edac_err_assert); -int edac_report_status =EDAC_REPORTING_ENABLED ;EXPORT_SYMBOL_GPL (edac_report_status);
include/linux/edac.h +0 -1
@@ -28,7 +28,6 @@ struct device;#define EDAC_OPSTATE_INT 2extern int edac_op_state; -extern int edac_err_assert;extern struct bus_type *edac_get_sysfs_subsys (void );
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 6/11 ACPI/extlog: Add EDAC dependency
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
We are calling EDAC functions - make the proper dependencies explicit.
--- drivers/acpi/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/acpi/Kconfig +1 -1
@@ -469,7 +469,7 @@ config ACPI_WATCHDOG config ACPI_EXTLOG tristate "Extended Error Log support" - depends on X86_MCE && X86_LOCAL_APIC + depends on X86_MCE && X86_LOCAL_APIC && EDAC select UEFI_CPER select RAS default n
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 11/11 EDAC: Rename report status accessors
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Change them to have the edac_ prefix.
No functionality change.
--- drivers/acpi/acpi_extlog.c | 8 ++++---- drivers/edac/edac_mc.c | 8 ++++---- drivers/edac/pnd2_edac.c | 2 +- drivers/edac/sb_edac.c | 4 ++-- drivers/edac/skx_edac.c | 2 +- include/linux/edac.h | 4 ++-- 6 files changed, 14 insertions(+), 14 deletions(-)
drivers/acpi/acpi_extlog.c +4 -4
@@ -229,7 +229,7 @@ static int __init extlog_init(void)if (!(cap &MCG_ELOG_P ) || !extlog_get_l1addr ())return -ENODEV ; -if (get_edac_report_status () ==EDAC_REPORTING_FORCE ) { +if (edac_get_report_status () ==EDAC_REPORTING_FORCE ) {pr_warn ("Not loading eMCA, error reporting force-enabled through EDAC.\n" );return -EPERM ; }
@@ -285,8 +285,8 @@ static int __init extlog_init(void)* eMCA event report method has higher priority than EDAC method, * unless EDAC event report method is mandatory. */ - old_edac_report_status =get_edac_report_status (); -set_edac_report_status (EDAC_REPORTING_DISABLED ); + old_edac_report_status =edac_get_report_status (); +edac_set_report_status (EDAC_REPORTING_DISABLED );mce_register_decode_chain (&extlog_mce_dec);/* enable OS to be involved to take over management from BIOS */ ((struct extlog_l1_head *)extlog_l1_addr)->flags |=FLAG_OS_OPTIN ;
@@ -308,7 +308,7 @@ static int __init extlog_init(void) static void __exit extlog_exit(void) { -set_edac_report_status (old_edac_report_status); +edac_set_report_status (old_edac_report_status);mce_unregister_decode_chain (&extlog_mce_dec); ((struct extlog_l1_head *)extlog_l1_addr)->flags &= ~FLAG_OS_OPTIN ;if (extlog_l1_addr)
drivers/edac/edac_mc.c +4 -4
@@ -57,20 +57,20 @@ static void const *edac_mc_owner;static struct bus_type mc_bus[EDAC_MAX_MCS ]; -int get_edac_report_status (void ) +int edac_get_report_status (void ) {return edac_report; } -EXPORT_SYMBOL_GPL (get_edac_report_status); +EXPORT_SYMBOL_GPL (edac_get_report_status); -void set_edac_report_status (int new) +void edac_set_report_status (int new) {if (new ==EDAC_REPORTING_ENABLED || new ==EDAC_REPORTING_DISABLED || new ==EDAC_REPORTING_FORCE ) edac_report = new; } -EXPORT_SYMBOL_GPL (set_edac_report_status); +EXPORT_SYMBOL_GPL (edac_set_report_status);static int edac_report_set (const char *str,const struct kernel_param *kp) {
drivers/edac/pnd2_edac.c +1 -1
@@ -1349,7 +1349,7 @@ static int pnd2_mce_check_error(struct notifier_block *nb, unsigned long val, vostruct dram_addr daddr;char *type; -if (get_edac_report_status () ==EDAC_REPORTING_DISABLED ) +if (edac_get_report_status () ==EDAC_REPORTING_DISABLED )return NOTIFY_DONE ; mci = pnd2_mci;
drivers/edac/sb_edac.c +2 -2
@@ -3075,7 +3075,7 @@ static int sbridge_mce_check_error(struct notifier_block *nb, unsigned long val,struct sbridge_pvt *pvt;char *type; -if (get_edac_report_status () ==EDAC_REPORTING_DISABLED ) +if (edac_get_report_status () ==EDAC_REPORTING_DISABLED )return NOTIFY_DONE ; mci =get_mci_for_node_id (mce->socketid);
@@ -3441,7 +3441,7 @@ static int __init sbridge_init(void)if (rc >=0 ) {mce_register_decode_chain (&sbridge_mce_dec); -if (get_edac_report_status () ==EDAC_REPORTING_DISABLED ) +if (edac_get_report_status () ==EDAC_REPORTING_DISABLED )sbridge_printk (KERN_WARNING ,"Loading driver, error reporting disabled.\n" );return 0 ; }
drivers/edac/skx_edac.c +1 -1
@@ -971,7 +971,7 @@ static int skx_mce_check_error(struct notifier_block *nb, unsigned long val,struct mem_ctl_info *mci;char *type; -if (get_edac_report_status () ==EDAC_REPORTING_DISABLED ) +if (edac_get_report_status () ==EDAC_REPORTING_DISABLED )return NOTIFY_DONE ;/* ignore unless this is memory related with an address */
include/linux/edac.h +2 -2
@@ -30,8 +30,8 @@ struct device;extern int edac_op_state;struct bus_type *edac_get_sysfs_subsys (void ); -int get_edac_report_status (void ); -void set_edac_report_status (int new); +int edac_get_report_status (void ); +void edac_set_report_status (int new);enum {EDAC_REPORTING_ENABLED ,
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 9/11 EDAC: Update Kconfig help text
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Remove the old URLs.
--- drivers/edac/Kconfig | 18 ++++-------------- 1 file changed, 4 insertions(+), 14 deletions(-)
drivers/edac/Kconfig +4 -14
@@ -13,23 +13,13 @@ menuconfig EDAC tristate "EDAC (Error Detection And Correction) reporting" depends on HAS_IOMEM && EDAC_SUPPORT && RAS help - EDAC is designed to report errors in the core system. - These are low-level errors that are reported in the CPU or - supporting chipset or other subsystems: + EDAC is a subsystem along with hardware-specific drivers designed to + report hardware errors. These are low-level errors that are reported + in the CPU or supporting chipset or other subsystems: memory errors, cache errors, PCI errors, thermal throttling, etc.. If unsure, select 'Y'. - If this code is reporting problems on your system, please - see the EDAC project web pages for more information at: - - <http://bluesmoke.sourceforge.net/> - - and: - - <http://buttersideup.com/edacwiki> - - There is also a mailing list for the EDAC project, which can - be found via the sourceforge page. + The mailing list for the EDAC project is [email protected]. if EDAC
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 7/11 EDAC: Issue tracepoint only when it is defined
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
... and this happens only when CONFIG_RAS is enabled.
--- drivers/edac/edac_mc.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/edac/edac_mc.c +7 -4
@@ -1180,10 +1180,13 @@ void edac_mc_handle_error(const enum hw_event_mc_err_type type,/* Report the error via the trace interface */ grain_bits =fls_long (e->grain) +1 ; -trace_mc_event (type, e->msg, e->label, e->error_count, - mci->mc_idx, e->top_layer, e->mid_layer, e->low_layer, - (e->page_frame_number <<PAGE_SHIFT ) | e->offset_in_page, - grain_bits, e->syndrome, e->other_detail); + +if (IS_ENABLED (CONFIG_RAS )) +trace_mc_event (type, e->msg, e->label, e->error_count, + mci->mc_idx, e->top_layer, e->mid_layer, + e->low_layer, + (e->page_frame_number <<PAGE_SHIFT ) | e->offset_in_page, + grain_bits, e->syndrome, e->other_detail);edac_raw_mc_handle_error (type, mci, e); }
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 10/11 EDAC: Delete edac_stub.c
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Move the remaining functionality to edac_mc.c. Convert "edac_report=" to a module parameter.
--- drivers/edac/Makefile | 2 +- drivers/edac/edac_mc.c | 61 ++++++++++++++++++++++++++++++++++++++++++++++++ drivers/edac/edac_stub.c | 37 ----------------------------- include/linux/edac.h | 26 +++------------------ 4 files changed, 65 insertions(+), 61 deletions(-) delete mode 100644 drivers/edac/edac_stub.c
drivers/edac/Makefile +1 -1
@@ -6,7 +6,7 @@ # GNU General Public License. # -obj-$(CONFIG_EDAC) := edac_stub.o edac_core.o +obj-$(CONFIG_EDAC) := edac_core.o edac_core-y := edac_mc.o edac_device.o edac_mc_sysfs.o edac_core-y += edac_module.o edac_device_sysfs.o wq.o
drivers/edac/edac_mc.c +61 -0
73 lines not on this page · raw message
drivers/edac/edac_stub.c deleted +0 -37
@@ -1,37 +0,0 @@ -/* -* common EDAC components that must be in kernel -* -* Author: Dave Jiang <[email protected]> -* -* 2007 (c) MontaVista Software, Inc. -* 2010 (c) Advanced Micro Devices Inc. -* Borislav Petkov <[email protected]> -* -* This file is licensed under the terms of the GNU General Public -* License version 2. This program is licensed "as is" without any -* warranty of any kind, whether express or implied. -* -*/ -#include <linux/module.h> -#include <linux/edac.h> -#include <linux/atomic.h> -#include <linux/device.h> - -int edac_report_status =EDAC_REPORTING_ENABLED ; -EXPORT_SYMBOL_GPL (edac_report_status); - -static int __init __maybe_unused edac_report_setup (char *str) -{ -if (!str) -return -EINVAL ; - -if (!strncmp (str,"on" ,2 )) -set_edac_report_status (EDAC_REPORTING_ENABLED ); -else if (!strncmp (str,"off" ,3 )) -set_edac_report_status (EDAC_REPORTING_DISABLED ); -else if (!strncmp (str,"force" ,5 )) -set_edac_report_status (EDAC_REPORTING_FORCE ); - -return 0 ; -} -__setup ("edac_report=" , edac_report_setup);
include/linux/edac.h +3 -23
@@ -29,7 +29,9 @@ struct device;extern int edac_op_state; -extern struct bus_type *edac_get_sysfs_subsys (void ); +struct bus_type *edac_get_sysfs_subsys (void ); +int get_edac_report_status (void ); +void set_edac_report_status (int new);enum {EDAC_REPORTING_ENABLED ,
@@ -37,28 +39,6 @@ enum { EDAC_REPORTING_FORCE }; -extern int edac_report_status; -#ifdef CONFIG_EDAC -static inline int get_edac_report_status (void ) -{ -return edac_report_status; -} - -static inline void set_edac_report_status (int new) -{ - edac_report_status = new; -} -#else -static inline int get_edac_report_status (void ) -{ -return EDAC_REPORTING_DISABLED ; -} - -static inline void set_edac_report_status (int new) -{ -} -#endif -static inline void opstate_init (void ) {switch (edac_op_state) {
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 2/11 x86/nmi, EDAC: Get rid of DRAM error reporting thru PCI SERR NMI
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Apparently, some machines used to report DRAM errors through a PCI SERR NMI. This is why we have a call into EDAC in the NMI handler. See
c0d121720220 ("drivers/edac: add new nmi rescan").
Quoted 1 line
From looking at the patch above, that's two drivers:
e752x_edac.cand
e7xxx_edac.c. Now, I wanna say those are old machines which are probably decommissioned already.
Tony says that "[t]the newest CPU supported by either of those drivers is the Xeon E7520 (a.k.a. "Nehalem") released in Q1'2010. Possibly some folks are still using these ... but people that hold onto h/w for 7 years generally cling to old s/w too ... so I'd guess it unlikely that we will get complaints for breaking these in upstream."
So even if there is a small number still in use, we did load EDAC with edac_op_state == EDAC_OPSTATE_POLL by default (we still do, in fact) which means a default EDAC setup without any parameters supplied on the command line or otherwise would never even log the error in the NMI handler because we're polling by default:
inline int edac_handler_set(void)
{
if (edac_op_state == EDAC_OPSTATE_POLL)
return 0;return atomic_read(&edac_handlers);
}
So, long story short, I'd like to get rid of that nastiness called edac_stub.c and confine all the EDAC drivers solely to drivers/edac/. If we ever have to do stuff like that again, it should be notifiers we're using and not some insanity like this one.
--- arch/x86/kernel/nmi.c | 11 ----------- drivers/edac/edac_stub.c | 22 ---------------------- include/linux/edac.h | 2 -- 3 files changed, 35 deletions(-)
arch/x86/kernel/nmi.c +0 -11
@@ -224,17 +224,6 @@ pci_serr_error(unsigned char reason, struct pt_regs *regs)pr_emerg ("NMI: PCI system error (SERR) for reason %02x on CPU %d.\n" , reason,smp_processor_id ()); -/* -* On some machines, PCI SERR line is used to report memory -* errors. EDAC makes use of it. -*/ -#if defined(CONFIG_EDAC ) -if (edac_handler_set ()) { -edac_atomic_assert_error (); -return ; - } -#endif -if (panic_on_unrecovered_nmi)nmi_panic (regs,"NMI: Not continuing" );
drivers/edac/edac_stub.c +0 -22
@@ -44,25 +44,3 @@ static int __init edac_report_setup(char *str)return 0 ; }__setup ("edac_report=" , edac_report_setup); - -/* -* called to determine if there is an EDAC driver interested in -* knowing an event (such as NMI) occurred -*/ -int edac_handler_set (void ) -{ -if (edac_op_state ==EDAC_OPSTATE_POLL ) -return 0 ; - -return atomic_read (&edac_handlers); -} -EXPORT_SYMBOL_GPL (edac_handler_set); - -/* -* handler for NMI type of interrupts to assert error -*/ -void edac_atomic_assert_error (void ) -{ - edac_err_assert++; -} -EXPORT_SYMBOL_GPL (edac_atomic_assert_error);
include/linux/edac.h +0 -2
@@ -31,8 +31,6 @@ extern int edac_op_state;extern int edac_err_assert;extern atomic_t edac_handlers; -extern int edac_handler_set (void ); -extern void edac_atomic_assert_error (void );extern struct bus_type *edac_get_sysfs_subsys (void );enum {
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 3/11 EDAC: Get rid of edac_handlers
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Use mc_devices list instead to check whether we have EDAC driver instances successfully registered with EDAC core.
--- drivers/edac/edac_mc.c | 6 ++---- drivers/edac/edac_stub.c | 3 --- include/linux/edac.h | 1 - 3 files changed, 2 insertions(+), 8 deletions(-)
drivers/edac/edac_mc.c +2 -4
@@ -601,7 +601,6 @@ static int add_mc_to_global_list(struct mem_ctl_info *mci) }list_add_tail_rcu (&mci->link, insert_before); -atomic_inc (&edac_handlers);return 0 ; fail0:
@@ -619,7 +618,6 @@ static int add_mc_to_global_list(struct mem_ctl_info *mci)static int del_mc_from_global_list (struct mem_ctl_info *mci) { -int handlers =atomic_dec_return (&edac_handlers);list_del_rcu (&mci->link);/* these are for safe removal of devices from global list while
@@ -628,7 +626,7 @@ static int del_mc_from_global_list(struct mem_ctl_info *mci)synchronize_rcu ();INIT_LIST_HEAD (&mci->link); -return handlers; +return list_empty (&mc_devices); }struct mem_ctl_info *edac_mc_find (int idx)
@@ -763,7 +761,7 @@ struct mem_ctl_info *edac_mc_del_mc(struct device *dev)/* mark MCI offline: */ mci->op_state =OP_OFFLINE ; -if (!del_mc_from_global_list (mci)) +if (del_mc_from_global_list (mci)) edac_mc_owner =NULL ;mutex_unlock (&mem_ctls_mutex);
drivers/edac/edac_stub.c +0 -3
@@ -20,9 +20,6 @@int edac_op_state =EDAC_OPSTATE_INVAL ;EXPORT_SYMBOL_GPL (edac_op_state); -atomic_t edac_handlers =ATOMIC_INIT (0 ); -EXPORT_SYMBOL_GPL (edac_handlers); -int edac_err_assert =0 ;EXPORT_SYMBOL_GPL (edac_err_assert);
include/linux/edac.h +0 -1
@@ -29,7 +29,6 @@ struct device;extern int edac_op_state;extern int edac_err_assert; -extern atomic_t edac_handlers;extern struct bus_type *edac_get_sysfs_subsys (void );
Signature · 1 line
2.11.0
Borislav Petkov
maintainerPATCH 1/11 EDAC, highbank: Align Makefile directives
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
... like the rest of the file.
--- drivers/edac/Makefile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/edac/Makefile +2 -2
@@ -67,8 +67,8 @@ obj-$(CONFIG_EDAC_AMD8131) += amd8131_edac.o obj-$(CONFIG_EDAC_TILE) += tile_edac.o -obj-$(CONFIG_EDAC_HIGHBANK_MC) += highbank_mc_edac.o -obj-$(CONFIG_EDAC_HIGHBANK_L2) += highbank_l2_edac.o +obj-$(CONFIG_EDAC_HIGHBANK_MC) += highbank_mc_edac.o +obj-$(CONFIG_EDAC_HIGHBANK_L2) += highbank_l2_edac.o obj-$(CONFIG_EDAC_OCTEON_PC) += octeon_edac-pc.o obj-$(CONFIG_EDAC_OCTEON_L2C) += octeon_edac-l2c.o
Signature · 1 line
2.11.0
Re: [PATCH 02/11] x86/nmi, EDAC: Get rid of DRAM error reporting thru PCI SERR NMI
In reply to Borislav PetkovQuoted 8 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
Apparently, some machines used to report DRAM errors through a PCI SERR NMI. This is why we have a call into EDAC in the NMI handler. See
c0d121720220("drivers/edac: add new nmi rescan").Quoted 1 line
From looking at the patch above, that's two drivers:
e752x_edac.cand
Stray '>'
Quoted 20 lines
e7xxx_edac.c. Now, I wanna say those are old machines which are probably decommissioned already.Tony says that "[t]the newest CPU supported by either of those drivers is the Xeon E7520 (a.k.a. "Nehalem") released in Q1'2010. Possibly some folks are still using these ... but people that hold onto h/w for 7 years generally cling to old s/w too ... so I'd guess it unlikely that we will get complaints for breaking these in upstream."
So even if there is a small number still in use, we did load EDAC with
edac_op_state==EDAC_OPSTATE_POLLby default (we still do, in fact) which means a default EDAC setup without any parameters supplied on the command line or otherwise would never even log the error in the NMI handler because we're polling by default:inline int edac_handler_set(void) { if (edac_op_state == EDAC_OPSTATE_POLL) return 0;
Quoted 6 lines
return atomic_read(&edac_handlers);}
So, long story short, I'd like to get rid of that nastiness called
edac_stub.cand confine all the EDAC drivers solely todrivers/edac/. If we ever have to do stuff like that again, it should be notifiers we're
Notifiers? You mean a proper NMI handler, right?
Other than that: Acked-by: Thomas Gleixner <[email protected]>
Thanks,
tglx
Borislav Petkov
maintainerRe: [PATCH 02/11] x86/nmi, EDAC: Get rid of DRAM error reporting thru PCI SERR NMI
Quoted 1 line · Thomas Gleixner
Notifiers? You mean a proper NMI handler, right?
I mean notifiers in the sense, the NMI handler should queue the logged error on the decoding chain x86_mce_decoder_chain like the rest of RAS code does instead of calling straight into EDAC which can be a module and which causes all this ugliness I'm removing now :-)
Quoted 1 line
Other than that: Acked-by: Thomas Gleixner <[email protected]>
Thanks!
Signature · 4 lines
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.