From: Tony Luck <tony.luck@intel.com>
To: Sohil Mehta <sohil.mehta@intel.com>
Cc: Qiuxu Zhuo <qiuxu.zhuo@intel.com>,
bp@alien8.de, tglx@linutronix.de, dave.hansen@linux.intel.com,
mingo@redhat.com, hpa@zytor.com, x86@kernel.org,
linux-edac@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 06/10] x86/mce: Convert multiple if () statements into a switch() statement
Date: Fri, 18 Oct 2024 13:14:56 -0700 [thread overview]
Message-ID: <ZxLBwO4HkkJG4WYn@agluck-desk3.sc.intel.com> (raw)
In-Reply-To: <c928d9aa-1609-4f5f-943c-fec72091e989@intel.com>
On Fri, Oct 18, 2024 at 12:44:00PM -0700, Sohil Mehta wrote:
> > diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
> > index 725c1d6fb1e5..40672fe0991a 100644
> > --- a/arch/x86/kernel/cpu/mce/core.c
> > +++ b/arch/x86/kernel/cpu/mce/core.c
> > @@ -1892,7 +1892,8 @@ static int __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c)
> > }
> >
> > /* This should be disabled by the BIOS, but isn't always */
>
> This comment is specific to the AMD and placing it before the switch
> makes it seem generic to the entire switch statement. It should probably
> be moved inside the AMD case just above the disable GART TLB check.
>
> > - if (c->x86_vendor == X86_VENDOR_AMD) {
> > + switch (c->x86_vendor) {
> > + case X86_VENDOR_AMD:
> > if (c->x86 == 15 && this_cpu_read(mce_num_banks) > 4) {
> > /*
> > * disable GART TBL walk error reporting, which
> > @@ -1925,9 +1926,9 @@ static int __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c)
> > if (c->x86 >= 0x17 && c->x86 <= 0x1A)
> > mce_flags.zen_ifu_quirk = 1;
> >
> > - }
> > + break;
> >
>
>
> Also, why not include the unknown vendor check (right above) inside the
> switch case as well?
>
> if (c->x86_vendor == X86_VENDOR_UNKNOWN) {
> pr_info("unknown CPU type - not enabling MCE support\n");
> return -EOPNOTSUPP;
> }
>
> This seems to follow the same pattern as others and can be the first
> case inside the switch.
The vendor specific bits are large enough to warrant their own
static functions (as we do elsewhere in this file).
How about this (only compile-tested) patch?
-Tony
From 967d8637ac90823f28f4612cbbac305c421b4853 Mon Sep 17 00:00:00 2001
From: Tony Luck <tony.luck@intel.com>
Date: Fri, 18 Oct 2024 13:01:02 -0700
Subject: [PATCH] x86/mce: Break up __mcheck_cpu_apply_quirks()
Split each vendor specific part into its own helper function.
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
arch/x86/kernel/cpu/mce/core.c | 172 ++++++++++++++++++---------------
1 file changed, 96 insertions(+), 76 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 2a938f429c4d..f51fb393d369 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -1880,101 +1880,121 @@ static void __mcheck_cpu_check_banks(void)
}
}
-/* Add per CPU specific workarounds here */
-static int __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c)
+static void apply_quirks_amd(struct cpuinfo_x86 *c)
{
struct mce_bank *mce_banks = this_cpu_ptr(mce_banks_array);
struct mca_config *cfg = &mca_cfg;
- if (c->x86_vendor == X86_VENDOR_UNKNOWN) {
- pr_info("unknown CPU type - not enabling MCE support\n");
- return -EOPNOTSUPP;
- }
-
/* This should be disabled by the BIOS, but isn't always */
- if (c->x86_vendor == X86_VENDOR_AMD) {
- if (c->x86 == 15 && this_cpu_read(mce_num_banks) > 4) {
- /*
- * disable GART TBL walk error reporting, which
- * trips off incorrectly with the IOMMU & 3ware
- * & Cerberus:
- */
- clear_bit(10, (unsigned long *)&mce_banks[4].ctl);
- }
- if (c->x86 < 0x11 && cfg->bootlog < 0) {
- /*
- * Lots of broken BIOS around that don't clear them
- * by default and leave crap in there. Don't log:
- */
- cfg->bootlog = 0;
- }
+ if (c->x86 == 15 && this_cpu_read(mce_num_banks) > 4) {
/*
- * Various K7s with broken bank 0 around. Always disable
- * by default.
+ * disable GART TBL walk error reporting, which
+ * trips off incorrectly with the IOMMU & 3ware
+ * & Cerberus:
*/
- if (c->x86 == 6 && this_cpu_read(mce_num_banks) > 0)
- mce_banks[0].ctl = 0;
-
+ clear_bit(10, (unsigned long *)&mce_banks[4].ctl);
+ }
+ if (c->x86 < 0x11 && cfg->bootlog < 0) {
/*
- * overflow_recov is supported for F15h Models 00h-0fh
- * even though we don't have a CPUID bit for it.
+ * Lots of broken BIOS around that don't clear them
+ * by default and leave crap in there. Don't log:
*/
- if (c->x86 == 0x15 && c->x86_model <= 0xf)
- mce_flags.overflow_recov = 1;
+ cfg->bootlog = 0;
+ }
+ /*
+ * Various K7s with broken bank 0 around. Always disable
+ * by default.
+ */
+ if (c->x86 == 6 && this_cpu_read(mce_num_banks) > 0)
+ mce_banks[0].ctl = 0;
- if (c->x86 >= 0x17 && c->x86 <= 0x1A)
- mce_flags.zen_ifu_quirk = 1;
+ /*
+ * overflow_recov is supported for F15h Models 00h-0fh
+ * even though we don't have a CPUID bit for it.
+ */
+ if (c->x86 == 0x15 && c->x86_model <= 0xf)
+ mce_flags.overflow_recov = 1;
- }
+ if (c->x86 >= 0x17 && c->x86 <= 0x1A)
+ mce_flags.zen_ifu_quirk = 1;
+}
- if (c->x86_vendor == X86_VENDOR_INTEL) {
- /*
- * SDM documents that on family 6 bank 0 should not be written
- * because it aliases to another special BIOS controlled
- * register.
- * But it's not aliased anymore on model 0x1a+
- * Don't ignore bank 0 completely because there could be a
- * valid event later, merely don't write CTL0.
- */
+static void apply_quirks_intel(struct cpuinfo_x86 *c)
+{
+ struct mce_bank *mce_banks = this_cpu_ptr(mce_banks_array);
+ struct mca_config *cfg = &mca_cfg;
- if (c->x86 == 6 && c->x86_model < 0x1A && this_cpu_read(mce_num_banks) > 0)
- mce_banks[0].init = false;
+ /*
+ * SDM documents that on family 6 bank 0 should not be written
+ * because it aliases to another special BIOS controlled
+ * register.
+ * But it's not aliased anymore on model 0x1a+
+ * Don't ignore bank 0 completely because there could be a
+ * valid event later, merely don't write CTL0.
+ */
- /*
- * All newer Intel systems support MCE broadcasting. Enable
- * synchronization with a one second timeout.
- */
- if ((c->x86 > 6 || (c->x86 == 6 && c->x86_model >= 0xe)) &&
- cfg->monarch_timeout < 0)
- cfg->monarch_timeout = USEC_PER_SEC;
+ if (c->x86 == 6 && c->x86_model < 0x1A && this_cpu_read(mce_num_banks) > 0)
+ mce_banks[0].init = false;
- /*
- * There are also broken BIOSes on some Pentium M and
- * earlier systems:
- */
- if (c->x86 == 6 && c->x86_model <= 13 && cfg->bootlog < 0)
- cfg->bootlog = 0;
+ /*
+ * All newer Intel systems support MCE broadcasting. Enable
+ * synchronization with a one second timeout.
+ */
+ if ((c->x86 > 6 || (c->x86 == 6 && c->x86_model >= 0xe)) &&
+ cfg->monarch_timeout < 0)
+ cfg->monarch_timeout = USEC_PER_SEC;
- if (c->x86_vfm == INTEL_SANDYBRIDGE_X)
- mce_flags.snb_ifu_quirk = 1;
+ /*
+ * There are also broken BIOSes on some Pentium M and
+ * earlier systems:
+ */
+ if (c->x86 == 6 && c->x86_model <= 13 && cfg->bootlog < 0)
+ cfg->bootlog = 0;
- /*
- * Skylake, Cascacde Lake and Cooper Lake require a quirk on
- * rep movs.
- */
- if (c->x86_vfm == INTEL_SKYLAKE_X)
- mce_flags.skx_repmov_quirk = 1;
+ if (c->x86_vfm == INTEL_SANDYBRIDGE_X)
+ mce_flags.snb_ifu_quirk = 1;
+
+ /*
+ * Skylake, Cascacde Lake and Cooper Lake require a quirk on
+ * rep movs.
+ */
+ if (c->x86_vfm == INTEL_SKYLAKE_X)
+ mce_flags.skx_repmov_quirk = 1;
+}
+
+static void apply_quirks_zhaoxin(struct cpuinfo_x86 *c)
+{
+ struct mca_config *cfg = &mca_cfg;
+
+ /*
+ * All newer Zhaoxin CPUs support MCE broadcasting. Enable
+ * synchronization with a one second timeout.
+ */
+ if (c->x86 > 6 || (c->x86_model == 0x19 || c->x86_model == 0x1f)) {
+ if (cfg->monarch_timeout < 0)
+ cfg->monarch_timeout = USEC_PER_SEC;
}
+}
- if (c->x86_vendor == X86_VENDOR_ZHAOXIN) {
- /*
- * All newer Zhaoxin CPUs support MCE broadcasting. Enable
- * synchronization with a one second timeout.
- */
- if (c->x86 > 6 || (c->x86_model == 0x19 || c->x86_model == 0x1f)) {
- if (cfg->monarch_timeout < 0)
- cfg->monarch_timeout = USEC_PER_SEC;
- }
+/* Add per CPU specific workarounds here */
+static int __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c)
+{
+ struct mca_config *cfg = &mca_cfg;
+
+ switch (c->x86_vendor) {
+ case X86_VENDOR_UNKNOWN:
+ pr_info("unknown CPU type - not enabling MCE support\n");
+ return -EOPNOTSUPP;
+
+ case X86_VENDOR_AMD:
+ apply_quirks_amd(c);
+ break;
+ case X86_VENDOR_INTEL:
+ apply_quirks_intel(c);
+ break;
+ case X86_VENDOR_ZHAOXIN:
+ apply_quirks_zhaoxin(c);
+ break;
}
if (cfg->monarch_timeout < 0)
--
2.47.0
>
next prev parent reply other threads:[~2024-10-18 20:14 UTC|newest]
Thread overview: 132+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-10 15:31 [PATCH 00/10] Clean up some x86/mce code Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 01/10] x86/mce/dev-mcelog: Use xchg() to get and clear the flags Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 02/10] x86/mce/intel: Use MCG_BANKCNT_MASK instead of 0xff Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 03/10] x86/mce: Make several functions return bool Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 04/10] x86/mce/threshold: Remove the redundant this_cpu_dec_return() Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 05/10] x86/mce/genpool: Make mce_gen_pool_create() return explicit error codes Qiuxu Zhuo
2024-11-02 14:31 ` Thomas Gleixner
2024-11-03 0:42 ` Zhuo, Qiuxu
2024-10-10 15:31 ` [PATCH 06/10] x86/mce: Convert multiple if () statements into a switch() statement Qiuxu Zhuo
2024-10-10 15:31 ` [PATCH 07/10] x86/mce: Remove the unnecessary {} Qiuxu Zhuo
2024-10-10 15:32 ` [PATCH 08/10] x86/mce: Remove the redundant zeroing assignments Qiuxu Zhuo
2024-10-10 15:32 ` [PATCH 09/10] x86/mce/amd: Remove unnecessary NULL pointer initializations Qiuxu Zhuo
2024-10-11 20:55 ` Luck, Tony
2024-10-10 15:32 ` [PATCH 10/10] x86/mce: Fix typos in comments Qiuxu Zhuo
2024-10-11 20:57 ` [PATCH 00/10] Clean up some x86/mce code Luck, Tony
2024-10-16 12:30 ` [PATCH v2 00/10] x86/mce: " Qiuxu Zhuo
2024-10-16 12:30 ` [PATCH v2 01/10] x86/mce/dev-mcelog: Use xchg() to get and clear the flags Qiuxu Zhuo
2024-10-16 12:30 ` [PATCH v2 02/10] x86/mce/intel: Use MCG_BANKCNT_MASK instead of 0xff Qiuxu Zhuo
2024-10-16 12:30 ` [PATCH v2 03/10] x86/mce: Make several functions return bool Qiuxu Zhuo
2024-10-18 19:02 ` Sohil Mehta
2024-10-19 3:00 ` Zhuo, Qiuxu
2024-10-16 12:30 ` [PATCH v2 04/10] x86/mce/threshold: Remove the redundant this_cpu_dec_return() Qiuxu Zhuo
2024-10-16 12:30 ` [PATCH v2 05/10] x86/mce/genpool: Make mce_gen_pool_create() return explicit error codes Qiuxu Zhuo
2024-10-18 19:20 ` Sohil Mehta
2024-10-19 3:29 ` Zhuo, Qiuxu
2024-10-16 12:30 ` [PATCH v2 06/10] x86/mce: Convert multiple if () statements into a switch() statement Qiuxu Zhuo
2024-10-18 19:44 ` Sohil Mehta
2024-10-18 20:14 ` Tony Luck [this message]
2024-10-19 0:04 ` Sohil Mehta
2024-10-19 5:46 ` Zhuo, Qiuxu
2024-10-21 16:06 ` Luck, Tony
2024-10-21 16:31 ` Dave Hansen
2024-10-23 8:20 ` Zhuo, Qiuxu
2024-10-24 13:05 ` Zhuo, Qiuxu
2024-10-24 16:42 ` Luck, Tony
2024-10-24 21:31 ` Sohil Mehta
2024-10-24 21:41 ` Luck, Tony
2024-10-24 22:18 ` Dave Hansen
2024-10-21 17:38 ` Sohil Mehta
2024-10-21 17:51 ` Luck, Tony
2024-10-21 18:26 ` Sohil Mehta
2024-10-21 18:40 ` Luck, Tony
2024-10-21 22:57 ` Sohil Mehta
2024-10-21 23:31 ` Luck, Tony
2024-10-22 0:07 ` Sohil Mehta
2024-10-22 0:17 ` Dave Hansen
2024-10-22 18:12 ` Sohil Mehta
2024-10-23 2:08 ` Zhuo, Qiuxu
2024-10-23 6:48 ` Sohil Mehta
2024-10-23 7:57 ` Zhuo, Qiuxu
2024-10-23 6:47 ` Sohil Mehta
2024-10-23 7:43 ` Zhuo, Qiuxu
2024-10-19 4:39 ` Zhuo, Qiuxu
2024-10-16 12:30 ` [PATCH v2 07/10] x86/mce: Remove the unnecessary {} Qiuxu Zhuo
2024-10-16 12:30 ` [PATCH v2 08/10] x86/mce: Remove the redundant zeroing assignments Qiuxu Zhuo
2024-10-18 20:25 ` Sohil Mehta
2024-10-19 7:37 ` Zhuo, Qiuxu
2024-10-19 7:40 ` H. Peter Anvin
2024-10-19 8:30 ` Zhuo, Qiuxu
2024-10-20 0:25 ` H. Peter Anvin
2024-10-21 6:29 ` Zhuo, Qiuxu
2024-10-16 12:30 ` [PATCH v2 09/10] x86/mce/amd: Remove unnecessary NULL pointer initializations Qiuxu Zhuo
2024-10-18 20:34 ` Sohil Mehta
2024-10-19 7:40 ` Zhuo, Qiuxu
2024-10-16 12:30 ` [PATCH v2 10/10] x86/mce: Fix typos in comments Qiuxu Zhuo
2024-10-18 20:36 ` Sohil Mehta
2024-10-19 8:02 ` Zhuo, Qiuxu
2024-10-17 15:30 ` [PATCH v2 00/10] x86/mce: Clean up some x86/mce code Nikolay Borisov
2024-10-18 20:40 ` Sohil Mehta
2024-10-25 2:45 ` [PATCH v3 " Qiuxu Zhuo
2024-10-25 2:45 ` [PATCH v3 01/10] x86/mce/dev-mcelog: Use xchg() to get and clear the flags Qiuxu Zhuo
2024-10-25 2:45 ` [PATCH v3 02/10] x86/mce/intel: Use MCG_BANKCNT_MASK instead of 0xff Qiuxu Zhuo
2024-10-25 2:45 ` [PATCH v3 03/10] x86/mce: Make several functions return bool and rename a function Qiuxu Zhuo
2024-10-28 13:06 ` Borislav Petkov
2024-10-29 3:32 ` Zhuo, Qiuxu
2024-10-30 10:24 ` Borislav Petkov
2024-10-31 3:23 ` Zhuo, Qiuxu
2024-10-25 2:45 ` [PATCH v3 04/10] x86/mce/threshold: Remove the redundant this_cpu_dec_return() Qiuxu Zhuo
2024-10-25 2:45 ` [PATCH v3 05/10] x86/mce/genpool: Make mce_gen_pool_create() return explicit error codes Qiuxu Zhuo
2024-11-02 13:13 ` Borislav Petkov
2024-10-25 2:45 ` [PATCH v3 06/10] x86/mce: Break up __mcheck_cpu_apply_quirks() Qiuxu Zhuo
2024-10-29 21:39 ` Yazen Ghannam
2024-10-30 1:39 ` Zhuo, Qiuxu
2024-10-30 14:16 ` Yazen Ghannam
2024-10-31 5:18 ` Zhuo, Qiuxu
2024-10-25 2:45 ` [PATCH v3 07/10] x86/mce: Convert family/model mixed checks to VFM-based checks Qiuxu Zhuo
2024-10-25 2:46 ` [PATCH v3 08/10] x86/mce: Remove the unnecessary {} Qiuxu Zhuo
2024-10-25 2:46 ` [PATCH v3 09/10] x86/mce/amd: Remove unnecessary NULL pointer initializations Qiuxu Zhuo
2024-10-25 2:46 ` [PATCH v3 10/10] x86/mce: Fix typos Qiuxu Zhuo
2024-11-11 6:04 ` [PATCH v4 0/8] x86/mce: Clean up some x86/mce code Qiuxu Zhuo
2024-11-11 6:04 ` [PATCH v4 1/8] x86/mce: Make several functions return bool Qiuxu Zhuo
2024-11-11 6:04 ` [PATCH v4 2/8] x86/mce/threshold: Remove the redundant this_cpu_dec_return() Qiuxu Zhuo
2024-11-11 6:04 ` [PATCH v4 3/8] x86/mce: Make four functions return bool Qiuxu Zhuo
2024-11-12 2:30 ` Sohil Mehta
2024-11-11 6:04 ` [PATCH v4 4/8] x86/mce: Break up __mcheck_cpu_apply_quirks() Qiuxu Zhuo
2024-11-12 2:31 ` Sohil Mehta
2024-11-11 6:04 ` [PATCH v4 5/8] x86/mce: Convert family/model mixed checks to VFM-based checks Qiuxu Zhuo
2024-11-12 2:31 ` Sohil Mehta
2024-11-12 15:27 ` Yazen Ghannam
2024-11-13 12:10 ` Zhuo, Qiuxu
2024-11-13 14:25 ` Yazen Ghannam
2024-11-14 1:11 ` Zhuo, Qiuxu
2024-11-11 6:04 ` [PATCH v4 6/8] x86/mce: Remove the unnecessary {} Qiuxu Zhuo
2024-11-12 15:43 ` Yazen Ghannam
2024-11-13 13:32 ` Zhuo, Qiuxu
2024-11-13 14:26 ` Yazen Ghannam
2024-11-13 16:11 ` Borislav Petkov
2024-11-14 1:23 ` Zhuo, Qiuxu
2024-11-11 6:04 ` [PATCH v4 7/8] x86/mce/amd: Remove unnecessary NULL pointer initializations Qiuxu Zhuo
2024-11-12 15:48 ` Yazen Ghannam
2024-11-11 6:04 ` [PATCH v4 8/8] x86/mce: Fix typos Qiuxu Zhuo
2024-11-12 2:36 ` Sohil Mehta
2024-11-12 15:38 ` Yazen Ghannam
2024-11-12 22:35 ` Sohil Mehta
2024-11-13 14:31 ` Yazen Ghannam
2024-11-14 1:20 ` Zhuo, Qiuxu
2024-11-13 14:48 ` Zhuo, Qiuxu
2024-12-12 14:00 ` [PATCH v5 0/7] Clean up some x86/mce code Qiuxu Zhuo
2024-12-12 14:00 ` [PATCH v5 1/7] x86/mce: Make several functions return bool Qiuxu Zhuo
2024-12-18 15:20 ` Yazen Ghannam
2024-12-19 15:19 ` Zhuo, Qiuxu
2024-12-12 14:00 ` [PATCH v5 2/7] x86/mce/threshold: Remove the redundant this_cpu_dec_return() Qiuxu Zhuo
2024-12-18 15:36 ` Yazen Ghannam
2024-12-12 14:00 ` [PATCH v5 3/7] x86/mce: Make four functions return bool Qiuxu Zhuo
2024-12-18 15:48 ` Yazen Ghannam
2024-12-12 14:01 ` [PATCH v5 4/7] x86/mce: Break up __mcheck_cpu_apply_quirks() Qiuxu Zhuo
2024-12-18 15:54 ` Yazen Ghannam
2024-12-12 14:01 ` [PATCH v5 5/7] x86/mce: Convert family/model mixed checks to VFM-based checks Qiuxu Zhuo
2024-12-12 14:01 ` [PATCH v5 6/7] x86/mce: Remove the redundant mce_hygon_feature_init() Qiuxu Zhuo
2024-12-12 18:08 ` Sohil Mehta
2024-12-18 16:02 ` Yazen Ghannam
2024-12-12 14:01 ` [PATCH v5 7/7] x86/mce/amd: Remove unnecessary NULL pointer initializations Qiuxu Zhuo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZxLBwO4HkkJG4WYn@agluck-desk3.sc.intel.com \
--to=tony.luck@intel.com \
--cc=bp@alien8.de \
--cc=dave.hansen@linux.intel.com \
--cc=hpa@zytor.com \
--cc=linux-edac@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=qiuxu.zhuo@intel.com \
--cc=sohil.mehta@intel.com \
--cc=tglx@linutronix.de \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox