Re: [PATCH v7 05/21] x86/cea: Export API for per-CPU exception stacks for KVM

[Xin Li <xin@zytor.com>]

On 8/29/2025 8:31 AM, Xin Li (Intel) wrote:
> Convert the __this_cpu_ist_{top,bottom}_va() macros into proper functions,
> and export __this_cpu_ist_top_va() to allow KVM to retrieve the top of the
> per-CPU exception stack.
> 
> FRED introduced new fields in the host-state area of the VMCS for stack
> levels 1->3 (HOST_IA32_FRED_RSP[123]), each respectively corresponding to
> per-CPU exception stacks for #DB, NMI and #DF.  KVM must populate these
> fields each time a vCPU is loaded onto a CPU.
> 
> To simplify access to the exception stacks in struct cea_exception_stacks,
> a union is used to create an array alias, enabling array-style indexing of
> the stack entries.

After introducing array-style indexing, we can further simplify the code by
removing ESTACKS_MEMBERS() from struct cea_exception_stacks, as done in the
following patch.  However, including this change in the current patch set
may be distracting, so I plan to submit it separately at a later time.


commit b305b83ab90c77242030727139c9b2e04f4de11e
Author: Xin Li (Intel) <xin@zytor.com>
Date:   Fri Aug 29 12:22:35 2025 -0400

     x86/cea: Simplify cea_exception_stacks by removing ESTACKS_MEMBERS()

     With most accesses to cea_exception_stacks now using array-style indexing,
     the ESTACKS_MEMBERS() macro is no longer necessary in cea_exception_stacks
     and can be removed to streamline the structure and improve code 
readability.

     Remove the CEA_ESTACK_SIZE macro, which redundantly defines 
EXCEPTION_STKSZ.

     Signed-off-by: Xin Li (Intel) <xin@zytor.com>

diff --git a/arch/x86/include/asm/cpu_entry_area.h 
b/arch/x86/include/asm/cpu_entry_area.h
index 58cd71144e5e..509e52fc3a0f 100644
--- a/arch/x86/include/asm/cpu_entry_area.h
+++ b/arch/x86/include/asm/cpu_entry_area.h
@@ -52,22 +52,15 @@ struct exception_stacks {

  /* The effective cpu entry area mapping with guard pages. */
  struct cea_exception_stacks {
-	union{
-		struct {
-			ESTACKS_MEMBERS(PAGE_SIZE, EXCEPTION_STKSZ)
-		};
-		struct {
-			char stack_guard[PAGE_SIZE];
-			char stack[EXCEPTION_STKSZ];
-		} event_stacks[N_EXCEPTION_STACKS];
-	};
+	struct {
+		char stack_guard[PAGE_SIZE];
+		char stack[EXCEPTION_STKSZ];
+	} event_stacks[N_EXCEPTION_STACKS];
+	char IST_top_guard[PAGE_SIZE];
  };

-#define CEA_ESTACK_SIZE(st)					\
-	sizeof(((struct cea_exception_stacks *)0)->st## _stack)
-
  #define CEA_ESTACK_OFFS(st)					\
-	offsetof(struct cea_exception_stacks, st## _stack)
+	offsetof(struct cea_exception_stacks, event_stacks[st].stack)

  #define CEA_ESTACK_PAGES					\
  	(sizeof(struct cea_exception_stacks) / PAGE_SIZE)
diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c
index 40f51e278171..93b10b264e53 100644
--- a/arch/x86/kernel/dumpstack_64.c
+++ b/arch/x86/kernel/dumpstack_64.c
@@ -70,9 +70,9 @@ struct estack_pages {

  #define EPAGERANGE(st)							\
  	[PFN_DOWN(CEA_ESTACK_OFFS(st)) ...				\
-	 PFN_DOWN(CEA_ESTACK_OFFS(st) + CEA_ESTACK_SIZE(st) - 1)] = {	\
+	 PFN_DOWN(CEA_ESTACK_OFFS(st) + EXCEPTION_STKSZ - 1)] = {	\
  		.offs	= CEA_ESTACK_OFFS(st),				\
-		.size	= CEA_ESTACK_SIZE(st),				\
+		.size	= EXCEPTION_STKSZ,				\
  		.type	= STACK_TYPE_EXCEPTION + st, }

  /*
diff --git a/arch/x86/mm/cpu_entry_area.c b/arch/x86/mm/cpu_entry_area.c
index 595c2e03ddd5..de0deb8b824c 100644
--- a/arch/x86/mm/cpu_entry_area.c
+++ b/arch/x86/mm/cpu_entry_area.c
@@ -157,7 +157,7 @@ static void __init percpu_setup_debug_store(unsigned 
int cpu)

  #define cea_map_stack(name) do {					\
  	npages = sizeof(estacks->name## _stack) / PAGE_SIZE;		\
-	cea_map_percpu_pages(cea->estacks.name## _stack,		\
+	cea_map_percpu_pages(cea->estacks.event_stacks[name].stack,	\
  			estacks->name## _stack, npages, PAGE_KERNEL);	\
  	} while (0)