x86/MCE/AMD: Fix the thresholding machinery initialization order

Currently, the code sets up the thresholding interrupt vector and only
then goes about initializing the thresholding banks. Which is wrong,
because an early thresholding interrupt would cause a NULL pointer
dereference when accessing those banks and prevent the machine from
booting.

Therefore, set the thresholding interrupt vector only *after* having
initialized the banks successfully.

Fixes: 18807ddb7f88 ("x86/mce/AMD: Reset Threshold Limit after logging error")
Reported-by: Rafał Miłecki <rafal@milecki.pl>
Reported-by: John Clemens <clemej@gmail.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Rafał Miłecki <rafal@milecki.pl>
Tested-by: John Clemens <john@deater.net>
Cc: Aravind Gopalakrishnan <aravindksg.lkml@gmail.com>
Cc: linux-edac@vger.kernel.org
Cc: stable@vger.kernel.org
Cc: Tony Luck <tony.luck@intel.com>
Cc: x86@kernel.org
Cc: Yazen Ghannam <Yazen.Ghannam@amd.com>
Link: https://lkml.kernel.org/r/20181127101700.2964-1-zajec5@gmail.com
Link: https://bugzilla.kernel.org/show_bug.cgi?id=201291

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index dd33c357548f11c0ac21c367d0edc20b34671218..e12454e21b8a5decc64af5e9143d2b469d6ecfbb 100644 (file)
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -56,7 +56,7 @@
 /* Threshold LVT offset is at MSR0xC0000410[15:12] */
 #define SMCA_THR_LVT_OFF       0xF000
 
-static bool thresholding_en;
+static bool thresholding_irq_en;
 
 static const char * const th_names[] = {
        "load_store",
@@ -534,9 +534,8 @@ prepare_threshold_block(unsigned int bank, unsigned int block, u32 addr,
 
 set_offset:
        offset = setup_APIC_mce_threshold(offset, new);
-
-       if ((offset == new) && (mce_threshold_vector != amd_threshold_interrupt))
-               mce_threshold_vector = amd_threshold_interrupt;
+       if (offset == new)
+               thresholding_irq_en = true;
 
 done:
        mce_threshold_block_init(&b, offset);
@@ -1357,9 +1356,6 @@ int mce_threshold_remove_device(unsigned int cpu)
 {
        unsigned int bank;
 
-       if (!thresholding_en)
-               return 0;
-
        for (bank = 0; bank < mca_cfg.banks; ++bank) {
                if (!(per_cpu(bank_map, cpu) & (1 << bank)))
                        continue;
@@ -1377,9 +1373,6 @@ int mce_threshold_create_device(unsigned int cpu)
        struct threshold_bank **bp;
        int err = 0;
 
-       if (!thresholding_en)
-               return 0;
-
        bp = per_cpu(threshold_banks, cpu);
        if (bp)
                return 0;
@@ -1408,9 +1401,6 @@ static __init int threshold_init_device(void)
 {
        unsigned lcpu = 0;
 
-       if (mce_threshold_vector == amd_threshold_interrupt)
-               thresholding_en = true;
-
        /* to hit CPUs online before the notifier is up */
        for_each_online_cpu(lcpu) {
                int err = mce_threshold_create_device(lcpu);
@@ -1419,6 +1409,9 @@ static __init int threshold_init_device(void)
                        return err;
        }
 
+       if (thresholding_irq_en)
+               mce_threshold_vector = amd_threshold_interrupt;
+
        return 0;
 }
 /*