lkml.org 
[lkml]   [2022]   [Aug]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: [PATCH] ACPI: APEI: move edac_init ahead of ghes platform drv register
On Fri, Aug 5, 2022 at 4:32 AM Jia He <justin.he@arm.com> wrote:
>
> Commit dc4e8c07e9e2 ("ACPI: APEI: explicit init of HEST and GHES in
> apci_init()") introduced a bug that invoking ghes_edac_register()
> before edac_init(). Because at that time, the bus "edac" hasn't been
> registered, this created sysfs /devices/mc0 instead of
> /sys/devices/system/edac/mc/mc0 and caused a sysfs dup splat on an
> Ampere eMag server:
> sysfs: cannot create duplicate filename '/devices/mc0'
> CPU: 19 PID: 1 Comm: swapper/0 Not tainted 5.19.0+ #138
> random: crng init done
> Hardware name: MiTAC RAPTOR EV-883832-X3-0001/RAPTOR, BIOS 0.14 02/22/2019
> Call trace:
> sysfs_warn_dup+0x6c/0x88
> sysfs_create_dir_ns+0xec/0x108
> kobject_add_internal+0xc0/0x328
> kobject_add+0x94/0x108
> device_add+0x104/0x8b0
> pmu_dev_alloc+0xb4/0x128
> perf_pmu_register+0x308/0x438
> xgene_pmu_dev_add+0x168/0x2c8
> acpi_pmu_dev_add+0x1f0/0x370
> acpi_ns_walk_namespace+0x16c/0x1ec
> acpi_walk_namespace+0xb0/0xf8
> xgene_pmu_probe+0x6b8/0x8a0
> platform_probe+0x70/0xe0
> really_probe+0x164/0x3b0
> __driver_probe_device+0x11c/0x190
> driver_probe_device+0x44/0xf8
> __driver_attach+0xc4/0x1b8
> bus_for_each_dev+0x78/0xd0
> driver_attach+0x2c/0x38
> bus_add_driver+0x150/0x240
> driver_register+0x6c/0x128
> __platform_driver_register+0x30/0x40
> xgene_pmu_driver_init+0x24/0x30
> do_one_initcall+0x50/0x248
> kernel_init_freeable+0x284/0x328
> kernel_init+0x2c/0x140
> ret_from_fork+0x10/0x20
> kobject_add_internal failed for mc0 with -EEXIST, don't try to register things with the same name in the same
>
> This patch fixes it by moving edac_init() into acpi_ghes_init() and ahead of
> platform_driver_register().
>
> Fixes: dc4e8c07e9e2 ("ACPI: APEI: explicit init of HEST and GHES in apci_init()")
> Signed-off-by: Jia He <justin.he@arm.com>
> Cc: Shuai Xue <xueshuai@linux.alibaba.com>
> ---
> drivers/acpi/apei/ghes.c | 1 +
> drivers/edac/edac_module.c | 3 +--
> include/linux/edac.h | 1 +
> 3 files changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
> index d91ad378c00d..1127dfffeeb0 100644
> --- a/drivers/acpi/apei/ghes.c
> +++ b/drivers/acpi/apei/ghes.c
> @@ -1462,6 +1462,7 @@ void __init acpi_ghes_init(void)
> int rc;
>
> sdei_init();
> + edac_init();
>
> if (acpi_disabled)
> return;
> diff --git a/drivers/edac/edac_module.c b/drivers/edac/edac_module.c
> index 32a931d0cb71..34ada2064b36 100644
> --- a/drivers/edac/edac_module.c
> +++ b/drivers/edac/edac_module.c
> @@ -99,7 +99,7 @@ EXPORT_SYMBOL_GPL(edac_get_sysfs_subsys);
> * edac_init
> * module initialization entry point
> */
> -static int __init edac_init(void)
> +int __init edac_init(void)
> {
> int err = 0;
>
> @@ -160,7 +160,6 @@ static void __exit edac_exit(void)
> /*
> * Inform the kernel of our entry and exit points
> */
> -subsys_initcall(edac_init);

This effectively makes EDAC depend on GHES which may not be always valid AFAICS.

> module_exit(edac_exit);
>
> MODULE_LICENSE("GPL");
> diff --git a/include/linux/edac.h b/include/linux/edac.h
> index e730b3468719..104b22c2c177 100644
> --- a/include/linux/edac.h
> +++ b/include/linux/edac.h
> @@ -30,6 +30,7 @@ struct device;
>
> extern int edac_op_state;
>
> +int __init edac_init(void);
> struct bus_type *edac_get_sysfs_subsys(void);
>
> static inline void opstate_init(void)
> --

\
 
 \ /
  Last update: 2022-08-08 20:18    [W:0.074 / U:2.044 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site