x86/show_trace_log_lvl: ensure stack pointer is aligned, again
Subsystems (2)
PATCH v3 1/1 x86/show_trace_log_lvl: ensure stack pointer is aligned, again
From: Vernon Lovejoy <[email protected]>
The commit e335bb51cc15 ("x86/unwind: Ensure stack pointer is aligned") tried to align the stack pointer in show_trace_log_lvl(), otherwise the "stack < stack_info.end" check can't guarantee that the last read does not go past the end of the stack.
However, we have the same problem with the initial value of the stack pointer, it can also be unaligned. So without this patch this trivial kernel module
#include <linux/module.h>
static int init(void)
{
asm volatile("sub $0x4,%rsp");
dump_stack();
asm volatile("add $0x4,%rsp");
return -EAGAIN;
}
module_init(init);
MODULE_LICENSE("GPL");crashes the kernel.
--- arch/x86/kernel/dumpstack.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
arch/x86/kernel/dumpstack.c +5 -2
@@ -195,7 +195,6 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,printk ("%sCall Trace:\n" , log_lvl);unwind_start (&state, task, regs, stack); - stack = stack ? :get_stack_pointer (task, regs); regs =unwind_get_entry_regs (&state, &partial);/*
@@ -214,9 +213,13 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,* - hardirq stack * - entry stack */ -for ( ; stack; stack =PTR_ALIGN (stack_info.next_sp,sizeof (long ))) { +for (stack = stack ?:get_stack_pointer (task, regs); + stack; + stack = stack_info.next_sp) {const char *stack_name; + stack =PTR_ALIGN (stack,sizeof (long )); +if (get_stack_info (stack, task, &stack_info, &visit_mask)) {/* * We weren't on a valid stack. It's possible that
Signature · 1 line
2.25.1.362.g51ebf55
Josh, this is boring.
We are changing this simple fix for the second time, precisely following your recommendations. And for the second time, you mysteriously disappear afterward.
Please tell us if you finally agree with this version or if you think we should make changes again.
Thanks.
Quoted 65 lines · Oleg Nesterov
From: Vernon Lovejoy <[email protected]>
The commit
e335bb51cc15("x86/unwind: Ensure stack pointer is aligned") tried to align the stack pointer inshow_trace_log_lvl(), otherwise the "stack < stack_info.end" check can't guarantee that the last read does not go past the end of the stack.However, we have the same problem with the initial value of the stack pointer, it can also be unaligned. So without this patch this trivial kernel module
#include <linux/module.h> static int init(void) { asm volatile("sub $0x4,%rsp"); dump_stack(); asm volatile("add $0x4,%rsp"); return -EAGAIN; } module_init(init); MODULE_LICENSE("GPL");crashes the kernel.
Fixes e335bb51cc15 ("x86/unwind: Ensure stack pointer is aligned")Signed-off-by Vernon Lovejoy <[email protected]>Signed-off-by Oleg Nesterov <[email protected]>--- arch/x86/kernel/dumpstack.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)arch/x86/kernel/dumpstack.c +5
-2@@ -195,7 +195,6 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,printk ("%sCall Trace:\n" , log_lvl);unwind_start (&state, task, regs, stack); - stack = stack ? :get_stack_pointer (task, regs); regs =unwind_get_entry_regs (&state, &partial);/* @@ -214,9 +213,13 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,* - hardirq stack * - entry stack */ -for ( ; stack; stack =PTR_ALIGN (stack_info.next_sp,sizeof (long ))) { +for (stack = stack ?:get_stack_pointer (task, regs); + stack; + stack = stack_info.next_sp) {const char *stack_name; + stack =PTR_ALIGN (stack,sizeof (long )); +if (get_stack_info (stack, task, &stack_info, &visit_mask)) {/* * We weren't on a valid stack. It's possible that Signature · 1 line
2.25.1.362.g51ebf55
Quoted 5 lines · Oleg Nesterov
Josh, this is boring.
We are changing this simple fix for the second time, precisely following your recommendations. And for the second time, you mysteriously disappear afterward.
Patience, please. I'm not just sitting around waiting for your patches.
Quoted 2 lines
Please tell us if you finally agree with this version or if you think we should make changes again.
Looks fine.
Signature · 1 line
Josh
[tip: objtool/urgent] x86/show_trace_log_lvl: Ensure stack pointer is aligned, again
In reply to Oleg NesterovThe following commit has been merged into the objtool/urgent branch of tip:
The following commit has been merged into the objtool/urgent branch of tip:
Commit-ID: 2e4be0d011f21593c6b316806779ba1eba2cd7e0 Gitweb: https://git.kernel.org/tip/2e4be0d011f21593c6b316806779ba1eba2cd7e0 Author: Vernon Lovejoy <[email protected]> AuthorDate: Fri, 12 May 2023 12:42:32 +02:00 Committer: Josh Poimboeuf <[email protected]> CommitterDate: Tue, 16 May 2023 06:31:04 -07:00
x86/show_trace_log_lvl: Ensure stack pointer is aligned, again
The commit e335bb51cc15 ("x86/unwind: Ensure stack pointer is aligned") tried to align the stack pointer in show_trace_log_lvl(), otherwise the "stack < stack_info.end" check can't guarantee that the last read does not go past the end of the stack.
However, we have the same problem with the initial value of the stack pointer, it can also be unaligned. So without this patch this trivial kernel module
#include <linux/module.h>
static int init(void)
{
asm volatile("sub $0x4,%rsp");
dump_stack();
asm volatile("add $0x4,%rsp");
return -EAGAIN;
}
module_init(init);
MODULE_LICENSE("GPL");crashes the kernel.
--- arch/x86/kernel/dumpstack.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
arch/x86/kernel/dumpstack.c +5 -2
@@ -195,7 +195,6 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,printk ("%sCall Trace:\n" , log_lvl);unwind_start (&state, task, regs, stack); - stack = stack ? :get_stack_pointer (task, regs); regs =unwind_get_entry_regs (&state, &partial);/*
@@ -214,9 +213,13 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,* - hardirq stack * - entry stack */ -for ( ; stack; stack =PTR_ALIGN (stack_info.next_sp,sizeof (long ))) { +for (stack = stack ?:get_stack_pointer (task, regs); + stack; + stack = stack_info.next_sp) {const char *stack_name; + stack =PTR_ALIGN (stack,sizeof (long )); +if (get_stack_info (stack, task, &stack_info, &visit_mask)) {/* * We weren't on a valid stack. It's possible that