x86/pm: Fix false positive kmemleak report in msr_build_context()

Since

  e2a1256 ("x86/speculation: Restore speculation related MSRs during S3 resume")

kmemleak reports this issue:

  unreferenced object 0xffff888009cedc00 (size 256):
    comm "swapper/0", pid 1, jiffies 4294693823 (age 73.764s)
    hex dump (first 32 bytes):
      00 00 00 00 00 00 00 00 48 00 00 00 00 00 00 00  ........H.......
      00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    backtrace:
      msr_build_context (include/linux/slab.h:621)
      pm_check_save_msr (arch/x86/power/cpu.c:520)
      do_one_initcall (init/main.c:1298)
      kernel_init_freeable (init/main.c:1370)
      kernel_init (init/main.c:1504)
      ret_from_fork (arch/x86/entry/entry_64.S:304)

Reproducer:

  - boot the VM with a debug kernel config (see
    #268)
  - wait ~1 minute
  - start a kmemleak scan

The root cause here is alignment within the packed struct saved_context
(from suspend_64.h). Kmemleak only searches for pointers that are
aligned (see how pointers are scanned in kmemleak.c), but pahole shows
that the saved_msrs struct member and all members after it in the
structure are unaligned:

  struct saved_context {
    struct pt_regs             regs;                 /*     0   168 */
    /* --- cacheline 2 boundary (128 bytes) was 40 bytes ago --- */
    u16                        ds;                   /*   168     2 */

    ...

    u64                        misc_enable;          /*   232     8 */
    bool                       misc_enable_saved;    /*   240     1 */

   /* Note below odd offset values for the remainder of this struct */

    struct saved_msrs          saved_msrs;           /*   241    16 */
    /* --- cacheline 4 boundary (256 bytes) was 1 bytes ago --- */
    long unsigned int          efer;                 /*   257     8 */
    u16                        gdt_pad;              /*   265     2 */
    struct desc_ptr            gdt_desc;             /*   267    10 */
    u16                        idt_pad;              /*   277     2 */
    struct desc_ptr            idt;                  /*   279    10 */
    u16                        ldt;                  /*   289     2 */
    u16                        tss;                  /*   291     2 */
    long unsigned int          tr;                   /*   293     8 */
    long unsigned int          safety;               /*   301     8 */
    long unsigned int          return_address;       /*   309     8 */

    /* size: 317, cachelines: 5, members: 25 */
    /* last cacheline: 61 bytes */
  } __attribute__((__packed__));

Move misc_enable_saved to the end of the struct declaration so that
saved_msrs fits in before the cacheline 4 boundary.

The comment above the saved_context declaration says to fix wakeup_64.S
file and __save/__restore_processor_state() if the struct is modified:
it looks like all the accesses in wakeup_64.S are done through offsets
which are computed at build-time. Update that comment accordingly.

At the end, the false positive kmemleak report is due to a limitation
from kmemleak but it is always good to avoid unaligned members for
optimisation purposes.

Please note that it looks like this issue is not new, e.g.

  https://lore.kernel.org/all/[email protected]/
  https://lore.kernel.org/all/[email protected]/

  [ bp: Massage + cleanup commit message. ]

Fixes: 7a9c2dd ("x86/pm: Introduce quirk framework to save/restore extra MSR registers around suspend/resume")
Suggested-by: Mat Martineau <[email protected]>
Signed-off-by: Matthieu Baerts <[email protected]>
Signed-off-by: Borislav Petkov <[email protected]>
Reviewed-by: Rafael J. Wysocki <[email protected]>
Link: https://lore.kernel.org/r/[email protected]

arch/x86/include/asm/suspend_32.h

@@ -19,7 +19,6 @@ struct saved_context {
 	u16 gs;
 	unsigned long cr0, cr2, cr3, cr4;
 	u64 misc_enable;
-	bool misc_enable_saved;
 	struct saved_msrs saved_msrs;
 	struct desc_ptr gdt_desc;
 	struct desc_ptr idt;
@@ -28,6 +27,7 @@ struct saved_context {
 	unsigned long tr;
 	unsigned long safety;
 	unsigned long return_address;
+	bool misc_enable_saved;
 } __attribute__((packed));
 
 /* routines for saving/restoring kernel state */

arch/x86/include/asm/suspend_64.h

@@ -14,9 +14,13 @@
  * Image of the saved processor state, used by the low level ACPI suspend to
  * RAM code and by the low level hibernation code.
  *
- * If you modify it, fix arch/x86/kernel/acpi/wakeup_64.S and make sure that
- * __save/__restore_processor_state(), defined in arch/x86/kernel/suspend_64.c,
- * still work as required.
+ * If you modify it, check how it is used in arch/x86/kernel/acpi/wakeup_64.S
+ * and make sure that __save/__restore_processor_state(), defined in
+ * arch/x86/power/cpu.c, still work as required.
+ *
+ * Because the structure is packed, make sure to avoid unaligned members. For
+ * optimisation purposes but also because tools like kmemleak only search for
+ * pointers that are aligned.
  */
 struct saved_context {
 	struct pt_regs regs;
@@ -36,7 +40,6 @@ struct saved_context {
 
 	unsigned long cr0, cr2, cr3, cr4;
 	u64 misc_enable;
-	bool misc_enable_saved;
 	struct saved_msrs saved_msrs;
 	unsigned long efer;
 	u16 gdt_pad; /* Unused */
@@ -48,6 +51,7 @@ struct saved_context {
 	unsigned long tr;
 	unsigned long safety;
 	unsigned long return_address;
+	bool misc_enable_saved;
 } __attribute__((packed));
 
 #define loaddebug(thread,register) \