lkml.org 
[lkml]   [2022]   [Aug]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRE: [RESEND PATCH v3 4/9] EDAC/ghes: Move ghes_edac.force_load to setup parameter
Date
Hi Borislav

> -----Original Message-----
[...]
> >
> > diff --git a/Documentation/admin-guide/kernel-parameters.txt
> > b/Documentation/admin-guide/kernel-parameters.txt
> > index d7f30902fda0..a5f0ee0d7727 100644
> > --- a/Documentation/admin-guide/kernel-parameters.txt
> > +++ b/Documentation/admin-guide/kernel-parameters.txt
> > @@ -1593,6 +1593,11 @@
> > When zero, profiling data is discarded and associated
> > debugfs files are removed at module unload time.
> >
> > + ghes_edac_force= [X86] Skip the platform check and forcibly load the
>
> So there already is ghes.disable which is using the module param thing.
> Why don't you do that too?
>
> > + ghes_edac modules.
>
> "module" - singular.
>
> > + Format: <bool>
> > + default: false (0)
> > +
> > goldfish [X86] Enable the goldfish android emulator platform.
> > Don't use this when you are not running on the
> > android emulator
> > diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c index
> > 9c52183e3ad9..e17e0ee8f842 100644
> > --- a/drivers/acpi/apei/ghes.c
> > +++ b/drivers/acpi/apei/ghes.c
> > @@ -94,6 +94,26 @@
> > #define FIX_APEI_GHES_SDEI_CRITICAL __end_of_fixed_addresses
> > #endif
> >
> > +/*
> > + * "ghes_edac_force=1" forcibly loads ghes_edac and skips the
> > +platform
> > + * check.
> > + */
> > +bool __read_mostly ghes_edac_force;
> > +EXPORT_SYMBOL(ghes_edac_force);
> > +
> > +static int __init setup_ghes_edac_load(char *str) {
> > + if (str)
> > + if (!strcmp("true", str) || !strcmp("1", str))
> > + ghes_edac_force = true;
> > +
> > + if (!IS_ENABLED(CONFIG_X86))
> > + ghes_edac_force = true;
> > +
> > + return 1;
> > +}
> > +__setup("ghes_edac_force=", setup_ghes_edac_load);
>
> Why all that?
>
> Isn't specifying
>
> ghes.edac_force_load
>
> on the kernel command line enough? I.e., you don't need to parse the passed
> in option - just the presence of the parameter is enough.
>
> > +
> > static ATOMIC_NOTIFIER_HEAD(ghes_report_chain);
> >
> > static inline bool is_hest_type_generic_v2(struct ghes *ghes) @@
> > -1517,13 +1537,13 @@ static struct acpi_platform_list plat_list[] = {
> > { } /* End */
> > };
> >
> > -struct list_head *ghes_get_devices(bool force)
> > +struct list_head *ghes_get_devices(void)
> > {
> > int idx = -1;
> >
> > if (IS_ENABLED(CONFIG_X86)) {
> > idx = acpi_match_platform_list(plat_list);
> > - if (idx < 0 && !force)
> > + if (idx < 0 && !ghes_edac_force)
> > return NULL;
> > }
> >
> > diff --git a/drivers/edac/ghes_edac.c b/drivers/edac/ghes_edac.c index
> > bb3ea42ba70b..6a2b54cc7240 100644
> > --- a/drivers/edac/ghes_edac.c
> > +++ b/drivers/edac/ghes_edac.c
> > @@ -54,10 +54,6 @@ static DEFINE_MUTEX(ghes_reg_mutex);
> > */
> > static DEFINE_SPINLOCK(ghes_lock);
> >
> > -/* "ghes_edac.force_load=1" skips the platform check */ -static bool
> > __read_mostly force_load; -module_param(force_load, bool, 0);
> > -
> > static bool system_scanned;
> >
> > static struct list_head *ghes_devs;
> > @@ -437,23 +433,12 @@ static int ghes_edac_register(struct device *dev)
> > mci->ctl_name = "ghes_edac";
> > mci->dev_name = "ghes";
> >
> > - if (fake) {
> > - pr_info("This system has a very crappy BIOS: It doesn't even list the
> DIMMS.\n");
> > - pr_info("Its SMBIOS info is wrong. It is doubtful that the error report
> would\n");
> > - pr_info("work on such system. Use this driver with caution\n");
> > - } else if (force_load) {
> > - pr_info("This EDAC driver relies on BIOS to enumerate memory and
> get error reports.\n");
> > - pr_info("Unfortunately, not all BIOSes reflect the memory layout
> correctly.\n");
> > - pr_info("So, the end result of using this driver varies from vendor to
> vendor.\n");
> > - pr_info("If you find incorrect reports, please contact your hardware
> vendor\n");
> > - pr_info("to correct its BIOS.\n");
> > - pr_info("This system has %d DIMM sockets.\n",
> ghes_hw.num_dimms);
> > - }
> > -
> > if (!fake) {
> > struct dimm_info *src, *dst;
> > int i = 0;
> >
> > + pr_info("This system has %d DIMM sockets.\n",
> ghes_hw.num_dimms);
> > +
> > mci_for_each_dimm(mci, dst) {
> > src = &ghes_hw.dimms[i];
> >
>
> This hunk...
>
> > @@ -478,6 +463,17 @@ static int ghes_edac_register(struct device *dev)
> > } else {
> > struct dimm_info *dimm = edac_get_dimm(mci, 0, 0, 0);
> >
> > + pr_info("This system has a very crappy BIOS: It doesn't even list the
> DIMMS.\n");
> > + pr_info("Its SMBIOS info is wrong. It is doubtful that the error report
> would\n");
> > + pr_info("work on such system. Use this driver with caution\n");
> > +
> > + if (ghes_edac_force) {
> > + pr_info("This EDAC driver relies on BIOS to enumerate memory
> and get\n");
> > + pr_info("error reports. Unfortunately, not all BIOSes reflect
> the\n");
> > + pr_info("memory layout correctly. If you find incorrect reports,
> please\n");
> > + pr_info("contact your hardware vendor for its in correct
> BIOS.\n");
> > + }
> > +
> > dimm->nr_pages = 1;
> > dimm->grain = 128;
> > dimm->mtype = MEM_UNKNOWN;
>
> ... and this hunk look unrelated to what this patch is doing. What are they for?
>
Another purpose for adjusting the pr_info hunk is to suppress the warning dmesg log
on Arm server.
On Arm, the ghes_edac_force_load (module parameter in ghes) should be true unconditionally.
So I moved the following block in fake==true check, is this reasonable?
+If (!fake)
+...
+else
+ if (ghes_edac_force) {
+ pr_info("This EDAC driver relies on BIOS to enumerate memory and get\n");
+ pr_info("error reports. Unfortunately, not all BIOSes reflect the\n");
+ pr_info("memory layout correctly. If you find incorrect reports, please\n");
+ pr_info("contact your hardware vendor for its in correct BIOS.\n");
+ }

What do you think of it?

--
Cheers,
Justin (Jia He)

\
 
 \ /
  Last update: 2022-08-30 03:22    [W:1.216 / U:0.012 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site