From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Xen-devel <xen-devel@lists.xen.org>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>,
Keir Fraser <keir@xen.org>, Jan Beulich <JBeulich@suse.com>,
Tim Deegan <tim@xen.org>
Subject: [xen-devel] [Patch 1/4] x86/stack: Refactor show_trace()
Date: Fri, 9 Aug 2013 20:55:55 +0100 [thread overview]
Message-ID: <1376078158-13739-2-git-send-email-andrew.cooper3@citrix.com> (raw)
In-Reply-To: <1376078158-13739-1-git-send-email-andrew.cooper3@citrix.com>
Before, show_trace() had two implementations depending on
CONFIG_FRAME_POINTER. Some parts were common, while the loops to wander up
the stack were different.
The version aided by frame pointers had a special case for function calls on
wild function pointers, but this doesn't need to be a special case.
After the refactoring, there are now two implementations of __show_trace()
which differ depending on CONFIG_FRAME_POINTER, and a single show_trace()
with the common bits, including the logic for wild function pointers.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
CC: Keir Fraser <keir@xen.org>
CC: Jan Beulich <JBeulich@suse.com>
CC: Tim Deegan <tim@xen.org>
---
The new wild pointer logic is rather larger than its pre-refactor version, but
is rather more legible. I cant think of a way to compact it without making it
substantially less legible.
---
xen/arch/x86/traps.c | 66 +++++++++++++++++++++++++++++---------------------
1 file changed, 39 insertions(+), 27 deletions(-)
diff --git a/xen/arch/x86/traps.c b/xen/arch/x86/traps.c
index 57dbd0c..b686177 100644
--- a/xen/arch/x86/traps.c
+++ b/xen/arch/x86/traps.c
@@ -192,14 +192,13 @@ static void show_guest_stack(struct vcpu *v, struct cpu_user_regs *regs)
#if !defined(CONFIG_FRAME_POINTER)
-static void show_trace(struct cpu_user_regs *regs)
+/* Stack trace from pointers found in stack, unaided by frame pointers. For
+ * caller convenience, this has the same prototype as its alternative, and
+ * simply ignores the rbp parameter.
+ */
+static void __show_trace(unsigned long sp, unsigned long __maybe_unused bp)
{
- unsigned long *stack = ESP_BEFORE_EXCEPTION(regs), addr;
-
- printk("Xen call trace:\n ");
-
- printk("[<%p>]", _p(regs->eip));
- print_symbol(" %s\n ", regs->eip);
+ unsigned long *stack = (unsigned long *)sp, addr;
while ( ((long)stack & (STACK_SIZE-BYTES_PER_LONG)) != 0 )
{
@@ -210,36 +209,22 @@ static void show_trace(struct cpu_user_regs *regs)
print_symbol(" %s\n ", addr);
}
}
-
- printk("\n");
}
#else
-static void show_trace(struct cpu_user_regs *regs)
+/* Stack trace from frames in the stack, using frame pointers */
+static void __show_trace(unsigned long sp, unsigned long bp)
{
unsigned long *frame, next, addr, low, high;
- printk("Xen call trace:\n ");
-
- /*
- * If RIP is not pointing into hypervisor code then someone may have
- * called into oblivion. Peek to see if they left a return address at
- * top of stack.
- */
- addr = is_active_kernel_text(regs->eip) ||
- !is_active_kernel_text(*ESP_BEFORE_EXCEPTION(regs)) ?
- regs->eip : *ESP_BEFORE_EXCEPTION(regs);
- printk("[<%p>]", _p(addr));
- print_symbol(" %s\n ", addr);
-
/* Bounds for range of valid frame pointer. */
- low = (unsigned long)(ESP_BEFORE_EXCEPTION(regs) - 2);
+ low = sp - 2*sizeof(unsigned long);
high = (low & ~(STACK_SIZE - 1)) +
(STACK_SIZE - sizeof(struct cpu_info) - 2*sizeof(unsigned long));
/* The initial frame pointer. */
- next = regs->ebp;
+ next = bp;
for ( ; ; )
{
@@ -272,12 +257,39 @@ static void show_trace(struct cpu_user_regs *regs)
low = (unsigned long)&frame[2];
}
-
- printk("\n");
}
#endif
+static void show_trace(const struct cpu_user_regs *regs)
+{
+ unsigned long sp = regs->rsp;
+ printk("Xen call trace:\n ");
+
+ /* If RIP looks sensible, or the top of the stack doesn't look sensible,
+ * print RIP at the top of the stack trace. */
+ if ( is_active_kernel_text(regs->rip) ||
+ !is_active_kernel_text(regs->rsp) )
+ {
+ printk("[<%p>]", _p(regs->rip));
+ print_symbol(" %s\n ", regs->rip);
+ }
+ /* else RIP looks bad but the top of the stack looks ok. Perhaps we
+ * followed a wild function pointer, so lets assume the top of the stack is
+ * a return address. Skip past it so__show_trace() doesn't print it
+ * again. */
+ else
+ {
+ printk("[<%p>]", _p(sp));
+ print_symbol(" %s\n ", sp);
+ sp += sizeof (unsigned long);
+ }
+
+ __show_trace(sp, regs->rbp);
+
+ printk("\n");
+}
+
void show_stack(struct cpu_user_regs *regs)
{
unsigned long *stack = ESP_BEFORE_EXCEPTION(regs), addr;
--
1.7.10.4
next prev parent reply other threads:[~2013-08-09 19:55 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-08-09 19:55 [xen-devel] [Patch 0/4] Xen stack trace printing improvements Andrew Cooper
2013-08-09 19:55 ` Andrew Cooper [this message]
2013-08-09 19:55 ` [xen-devel] [Patch 2/4] x86/stack: Adjust boundary conditions for printed stacks Andrew Cooper
2013-08-12 8:46 ` Jan Beulich
2013-08-12 9:43 ` Andrew Cooper
2013-08-12 9:49 ` Jan Beulich
2013-08-12 12:15 ` [xen-devel] [Patch v2 " Andrew Cooper
2013-08-12 13:44 ` Jan Beulich
2013-08-12 13:53 ` Andrew Cooper
2013-08-09 19:55 ` [xen-devel] [Patch 3/4] x86/stack: Change show_stack_overflow() to use frame pointers if available Andrew Cooper
2013-08-09 19:55 ` [xen-devel] [Patch 4/4] DO NOT APPLY: Test code for interesting stack overflows Andrew Cooper
2013-09-09 11:09 ` [xen-devel] [Patch 0/4] Xen stack trace printing improvements Keir Fraser
2013-09-09 12:28 ` Andrew Cooper
2013-09-09 12:44 ` Jan Beulich
2013-09-09 12:50 ` Andrew Cooper
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1376078158-13739-2-git-send-email-andrew.cooper3@citrix.com \
--to=andrew.cooper3@citrix.com \
--cc=JBeulich@suse.com \
--cc=keir@xen.org \
--cc=tim@xen.org \
--cc=xen-devel@lists.xen.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).