lkml.org 
[lkml]   [2018]   [Oct]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
Subject[RFC v1 24/31] kunit: improved sigsegv stack trace printing
From
Sacrificed the control of printing stack trace within the crash handler
in the test runner for getting a better stack trace; this is still not
ideal, but much better than before.

Signed-off-by: Brendan Higgins <brendanhiggins@google.com>
---
arch/um/kernel/trap.c | 11 ++++++++++-
kunit/test.c | 13 +++++++++----
2 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/arch/um/kernel/trap.c b/arch/um/kernel/trap.c
index 9b97712daf14f..c3ff8346800c4 100644
--- a/arch/um/kernel/trap.c
+++ b/arch/um/kernel/trap.c
@@ -226,8 +226,17 @@ unsigned long segv(struct faultinfo fi, unsigned long ip, int is_user,
current->thread.segv_regs = container_of(regs, struct pt_regs, regs);

catcher = current->thread.fault_catcher;
- if (catcher && current->thread.is_running_test)
+ if (catcher && current->thread.is_running_test) {
+ /*
+ * TODO(b/77223210): Right now we don't have a way to store a
+ * copy of the stack, or a copy of information from the stack,
+ * so we need to print it now; otherwise, the stack will be
+ * destroyed by segv_run_catcher which works by popping off
+ * stack frames.
+ */
+ show_stack(NULL, NULL);
segv_run_catcher(catcher, (void *) address);
+ }
else if (!is_user && (address >= start_vm) && (address < end_vm)) {
flush_tlb_kernel_vm();
goto out;
diff --git a/kunit/test.c b/kunit/test.c
index 6ea60059b4918..5d78f76b421af 100644
--- a/kunit/test.c
+++ b/kunit/test.c
@@ -210,12 +210,17 @@ static void test_handle_test_crash(struct test *test,
struct test_module *module,
struct test_case *test_case)
{
- test_err(test, "%s crashed", test_case->name);
/*
- * TODO(brendanhiggins@google.com): This prints the stack trace up
- * through this frame, not up to the frame that caused the crash.
+ * TODO(brendanhiggins@google.com): Right now we don't have a way to
+ * store a copy of the stack, or a copy of information from the stack,
+ * so we need to print it in the "trap" handler; otherwise, the stack
+ * will be destroyed when it returns to us by popping off the
+ * appropriate stack frames (see longjmp).
+ *
+ * Ideally we would print the stack trace here, but we do not have the
+ * ability to do so with meaningful information at this time.
*/
- show_stack(NULL, NULL);
+ test_err(test, "%s crashed", test_case->name);

test_case_internal_cleanup(test);
}
--
2.19.1.331.ge82ca0e54c-goog
\
 
 \ /
  Last update: 2018-10-17 01:56    [W:0.664 / U:0.348 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site