Skip to contentSkip to conversation

Revert "x86/HVM: also dump stacks from show_execution_state()"

PATCH In review Roger Pau Monne 3 participants 3 messages Nov 4, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Revert "x86/HVM: also dump stacks from show_execution_state()"

This reverts commit adb715db698bc8ec3b88c24eb88b21e9da5b6c07.

The dumping of stacks for HVM guests is problematic, since it requires taking the p2m lock in order to walk the guest page tables and the p2m.

The suggested solution to the issue is to introduce and use a lockless p2m walker, that relies on being executed with interrupts disabled in order to prevent any p2m pages from being freed while doing the walk.

Note that modifications of p2m entries are already done atomically in order to prevent the hardware walker from seeing partially updated values.

---
 xen/arch/x86/include/asm/processor.h |  1 +
 xen/arch/x86/traps.c                 | 35 ++++++++--------------------
 2 files changed, 11 insertions(+), 25 deletions(-)
xen/arch/x86/include/asm/processor.h +1 -0
index 8e2816fae9..4cdafe2c4d 100644
@@ -493,6 +493,7 @@ static always_inline void rep_nop(void)
 #define cpu_relax() rep_nop()
 
 void show_code(const struct cpu_user_regs *regs);
+void show_stack(const struct cpu_user_regs *regs);
 void show_stack_overflow(unsigned int cpu, const struct cpu_user_regs *regs);
 void show_registers(const struct cpu_user_regs *regs);
 void show_execution_state(const struct cpu_user_regs *regs);
xen/arch/x86/traps.c +10 -25
index 5c0aabe8a3..c1e5ef1cc4 100644
@@ -278,6 +278,10 @@ static void show_guest_stack(struct vcpu *v, const struct cpu_user_regs *regs)
     unsigned long mask = STACK_SIZE;
     void *stack_page = NULL;
 
+    /* Avoid HVM as we don't know what the stack looks like. */
+    if ( is_hvm_vcpu(v) )
+        return;
+
     if ( is_pv_32bit_vcpu(v) )
     {
         compat_show_guest_stack(v, regs, debug_stack_lines);
@@ -586,11 +590,14 @@ static void show_trace(const struct cpu_user_regs *regs)
     printk("\n");
 }
 
-static void show_stack(const struct cpu_user_regs *regs)
+void show_stack(const struct cpu_user_regs *regs)
 {
     unsigned long *stack = ESP_BEFORE_EXCEPTION(regs), *stack_bottom, addr;
     int i;
 
+    if ( guest_mode(regs) )
+        return show_guest_stack(current, regs);
+
     printk("Xen stack trace from "__OP"sp=%p:\n  ", stack);
 
     stack_bottom = _p(get_stack_dump_bottom(regs->rsp));
@@ -655,30 +662,8 @@ void show_execution_state(const struct cpu_user_regs *regs)
     unsigned long flags = console_lock_recursive_irqsave();
 
     show_registers(regs);
-
-    if ( guest_mode(regs) )
-    {
-        struct vcpu *curr = current;
-
-        if ( is_hvm_vcpu(curr) )
-        {
-            /*
-             * Stop interleaving prevention: The necessary P2M lookups
-             * involve locking, which has to occur with IRQs enabled.
-             */
-            console_unlock_recursive_irqrestore(flags);
-
-            show_hvm_stack(curr, regs);
-            return;
-        }
-
-        show_guest_stack(curr, regs);
-    }
-    else
-    {
-        show_code(regs);
-        show_stack(regs);
-    }
+    show_code(regs);
+    show_stack(regs);
 
     console_unlock_recursive_irqrestore(flags);
 }
Signature · 1 line
2.37.3
Acked-by on v1 1/1
#1 lore raw
Quoted 11 lines · Roger Pau Monne

This reverts commit adb715db698bc8ec3b88c24eb88b21e9da5b6c07.

The dumping of stacks for HVM guests is problematic, since it requires taking the p2m lock in order to walk the guest page tables and the p2m.

The suggested solution to the issue is to introduce and use a lockless p2m walker, that relies on being executed with interrupts disabled in order to prevent any p2m pages from being freed while doing the walk.

Quoted 5 lines

Note that modifications of p2m entries are already done atomically in order to prevent the hardware walker from seeing partially updated values.

x86 maintainers call.

~Andrew

#2 lore raw
In reply to Roger Pau Monne

Hi Roger,

Quoted 16 lines

-----Original Message-----
From: Roger Pau Monne <[email protected]>
Subject: [PATCH for-4.17] Revert "x86/HVM: also dump stacks from show_execution_state()"

This reverts commit adb715db698bc8ec3b88c24eb88b21e9da5b6c07.

The dumping of stacks for HVM guests is problematic, since it requires taking the p2m lock in order to walk the guest page tables and the p2m.

The suggested solution to the issue is to introduce and use a lockless p2m walker, that relies on being executed with interrupts disabled in order to prevent any p2m pages from being freed while doing the walk.

Quoted 5 lines

Note that modifications of p2m entries are already done atomically in order to prevent the hardware walker from seeing partially updated values.

Kind regards,
Henry