Hi Lijiang,
On Thu, Jul 25, 2024 at 6:27 PM lijiang <lijiang@redhat.com> wrote:
>
> On Thu, Jul 25, 2024 at 7:59 AM Tao Liu <ltao@redhat.com> wrote:
>>
>> Hi Lijiang,
>>
>> On Wed, Jul 24, 2024 at 6:37 PM lijiang <lijiang@redhat.com> wrote:
>> >
>> > Hi, Tao
>> >
>> > On Tue, Jul 23, 2024 at 2:24 PM lijiang <lijiang@redhat.com> wrote:
>> >>
>> >> On Tue, Jul 23, 2024 at 1:55 PM Tao Liu <ltao@redhat.com> wrote:
>> >>>
>> >>> Hi Aditya & Lianbo,
>> >>>
>> >>> On Mon, Jun 24, 2024 at 11:46 PM Aditya Gupta <adityag@linux.ibm.com> wrote:
>> >>> >
>> >>> > Hello Lianbo,
>> >>> >
>> >>> > On 24/06/24 05:32PM, lijiang wrote:
>> >>> > > > <...snip...>
>> >>> > > >
>> >>> > > > Before:
>> >>> > > > crash> gdb bt
>> >>> > > > #0 0xffffffff816a8f65 in context_switch ...
>> >>> > > > #1 __schedule () ...
>> >>> > > > #2 0xffffffff816a94e9 in schedule ...
>> >>> > > > #3 0xffffffff816a86fd in schedule_hrtimeout_range_clock ...
>> >>> > > > #4 0xffffffff816a8733 in schedule_hrtimeout_range ...
>> >>> > > > #5 0xffffffff8124bb7e in ep_poll ...
>> >>> > > > #6 0xffffffff8124d00d in SYSC_epoll_wait ...
>> >>> > > > #7 SyS_epoll_wait ...
>> >>> > > > #8 <signal handler called>
>> >>> > > > #9 0x00007f0449407923 in ?? ()
>> >>> > > > #10 0xffff880100000001 in ?? ()
>> >>> > > > #11 0xffff880169b3c010 in ?? ()
>> >>> > > > #12 0x0000000000000040 in irq_stack_union ()
>> >>> > > > #13 0xffff880169b3c058 in ?? ()
>> >>> > > > #14 0xffff880169b3c048 in ?? ()
>> >>> > > > #15 0xffff880169b3c050 in ?? ()
>> >>> > > > #16 0x0000000000000000 in ?? ()
>> >>> > > >
>> >>> > > > After:
>> >>> > > > crash> gdb bt
>> >>> > > > #0 0xffffffff816a8f65 in context_switch ...
>> >>> > > > #1 __schedule () ...
>> >>> > > > #2 0xffffffff816a94e9 in schedule () ...
>> >>> > > > #3 0xffffffff816a86fd in schedule_hrtimeout_range_clock ...
>> >>> > > > #4 0xffffffff816a8733 in schedule_hrtimeout_range ...
>> >>> > > > #5 0xffffffff8124bb7e in ep_poll ...
>> >>> > > > #6 0xffffffff8124d00d in SYSC_epoll_wait ...
>> >>> > > > #7 SyS_epoll_wait ...
>> >>> > > > #8 <signal handler called>
>> >>> > > > #9 0x00007f0449407923 in ?? ()
>> >>> > > >
>> >>> > > >
>> >>> > > It seems that there are still some non-kernel addresses that do not get
>> >>> > > filtered. Can you help double check?
>> >>>
>> >>> Yes, it is a non-kernel address which does not get filtered.
>> >>>
>> >>> > >
>> >>> > >
>> >>> > > For example:
>> >>> > >
>> >>> > > crash> gdb bt
>> >>> > > #0 crash_setup_regs (newregs=0xffffb5bb4f197938, oldregs=0x0) at
>> >>> > > ./arch/x86/include/asm/kexec.h:114
>> >>> > > #1 0xffffffff8e61e32e in __crash_kexec (regs=regs@entry=0x0) at
>> >>> > > kernel/crash_core.c:122
>> >>> > > #2 0xffffffff8e51a64d in panic (fmt=fmt@entry=0xffffffff8fa51609 "sysrq
>> >>> > > triggered crash\n") at kernel/panic.c:366
>> >>> > > #3 0xffffffff8ec21f86 in sysrq_handle_crash (key=<optimized out>) at
>> >>> > > drivers/tty/sysrq.c:154
>> >>> > > #4 0xffffffff8ec22550 in __handle_sysrq (key=<optimized out>,
>> >>> > > check_mask=check_mask@entry=false) at drivers/tty/sysrq.c:612
>> >>> > > #5 0xffffffff8ec22bf5 in write_sysrq_trigger (file=<optimized out>,
>> >>> > > buf=<optimized out>, count=2, ppos=<optimized out>) at
>> >>> > > drivers/tty/sysrq.c:1183
>> >>> > > #6 0xffffffff8e935ae5 in pde_write (ppos=<optimized out>, count=<optimized
>> >>> > > out>, buf=<optimized out>, file=0xffffb5bb4f197938, pde=0xffff98338b78e0c0)
>> >>> > > at fs/proc/inode.c:334
>> >>> > > #7 proc_reg_write (file=0xffffb5bb4f197938, buf=0x0, count=1, ppos=0x0) at
>> >>> > > fs/proc/inode.c:346
>> >>> > > #8 0xffffffff8e88d382 in vfs_write (file=file@entry=0xffff98338b789200,
>> >>> > > buf=buf@entry=0x5614d58a22c0 <error: Cannot access memory at address
>> >>> > > 0x5614d58a22c0>, count=count@entry=2, pos=pos@entry=0xffffb5bb4f197b78) at
>> >>> > > fs/read_write.c:588
>> >>> > > #9 0xffffffff8e88d9ff in ksys_write (fd=<optimized out>,
>> >>> > > buf=0x5614d58a22c0 <error: Cannot access memory at address 0x5614d58a22c0>,
>> >>> > > count=2) at fs/read_write.c:643
>> >>> > > #10 0xffffffff8f124429 in do_syscall_x64 (nr=1, regs=0xffffb5bb4f197f58) at
>> >>> > > arch/x86/entry/common.c:52
>> >>> > > #11 do_syscall_64 (regs=0xffffb5bb4f197f58, nr=1) at
>> >>> > > arch/x86/entry/common.c:83
>> >>> > > #12 0xffffffff8f20012b in entry_SYSCALL_64 () at
>> >>> > > arch/x86/entry/entry_64.S:121
>> >>> > > #13 0x00007f9a147f69e0 in ?? ()
>> >>> > >
>> >>> > > The frame #13 looks like a non-kernel address.
>> >>>
>> >>> The address usually to be the user space address before entering
>> >>> kernel, you can see it by:
>> >>>
>> >>> crash> gdb bt
>> >>> ...snip...
>> >>> #7 SyS_epoll_wait ...
>> >>> #8 <signal handler called>
>> >>> #9 0x00007f0449407923 in ?? ()
>> >>>
>> >>> crash> bt
>> >>> ...snip...
>> >>> #6 [ffff880169b3bf80] system_call_fastpath at ffffffff816b5009
>> >>> RIP: 00007f0449407923 ...
>> >>>
>> >>> So I think leaving the last frame here is useful and shouldn't be
>> >>> filtered. Though it looks like some garbage data, it can help for some
>> >>> experienced users...
>> >>
>> >>
>> >>
>> >> Hmm, normally it should be filtered, otherwise this looks weird.
>> >>
>> >> Let me dig into the details and see if that can be filtered out.
>> >
>> >
>> Thanks for your patch!
>>
>> > Can you help try this? I saw the non-kernel addresses are filtered out.
>> >
>> > + #ifdef CRASH_MERGE
>> > + extern "C" int is_kvaddr(ulong);
>> > + #endif
>> >
>> > static void
>> > print_frame (const frame_print_options &fp_opts,
>> > frame_info *frame, int print_level,
>> > enum print_what print_what, int print_args,
>> > struct symtab_and_line sal)
>> > {
>> > struct gdbarch *gdbarch = get_frame_arch (frame);
>> > struct ui_out *uiout = current_uiout;
>> > enum language funlang = language_unknown;
>> > struct value_print_options opts;
>> > struct symbol *func;
>> > CORE_ADDR pc = 0;
>> > int pc_p;
>> >
>> > pc_p = get_frame_pc_if_available (frame, &pc);
>> > + #ifdef CRASH_MERGE
>> > + if (!is_kvaddr(pc))
>> > + return;
>> > + #endif
>> > ...
>> > }
>>
>> Your code change is "stop printing for non-kernel address", however I
>> would prefer to "stop stack unwinding for non-kernel address", so I
>
>
> Yes, they are different.
>
> If "stop stack unwinding for non-kernel address" won't truncate the backtrace, I would also prefer it.
> Otherwise I would prefer another one.
OK, it won't truncate the backtrace.
>
>>
>> think it is better to modify it within the for loop, this is where
>> stack unwinding happens.
>>
>> for (fi = trailing; fi && count--; fi = get_prev_frame (fi))
>> ...
>> print_frame_info (fp_opts, fi, 1, LOCATION, 1, 0);
>>
>> I made the following code change:
>>
>> for (fi = trailing; fi && count--; fi = get_prev_frame (fi))
>> {
>> QUIT;
>> ...
>> + CORE_ADDR pc = 0;
>> + get_frame_pc_if_available (fi, &pc);
>> + if (!is_kvaddr(pc)) {
>> + fi = NULL;
>> + break;
>> + }
>> print_frame_info (fp_opts, fi, 1, LOCATION, 1, 0);
>>
>> With the change:
>>
>> crash> gdb bt
>> #0 blk_mq_rq_timed_out (req=0xffff880fdb246000,
>> reserved=reserved@entry=false) at block/blk-mq.c:640
>> #1 0xffffffff8130504c in blk_mq_check_expired
>> (hctx=hctx@entry=0xffff880fda56bc00, rq=<optimized out>,
>> priv=priv@entry=0xffff880fcf68fde8, reserved=reserved@entry=false) at
>> block/blk-mq.c:697
>> #2 0xffffffff81305eb4 in bt_for_each
>> (hctx=hctx@entry=0xffff880fda56bc00, bt=bt@entry=0xffff88014c14c310,
>> off=32, fn=fn@entry=0xffffffff81304ff0 <blk_mq_check_expired>,
>> data=data@entry=0xffff880fcf68fde8, reserved=reserved@entry=false) at
>> block/blk-mq-tag.c:431
>> #3 0xffffffff8130686e in blk_mq_queue_tag_busy_iter
>> (q=q@entry=0xffff88004912a340, fn=fn@entry=0xffffffff81304ff0
>> <blk_mq_check_expired>, priv=priv@entry=0xffff880fcf68fde8) at
>> block/blk-mq-tag.c:530
>> #4 0xffffffff81301d2b in blk_mq_timeout_work
>> (work=0xffff88004912ab68) at block/blk-mq.c:730
>> #5 0xffffffff810a881a in process_one_work
>> (worker=worker@entry=0xffff880fd86efa00, work=0xffff88004912ab68) at
>> kernel/workqueue.c:2252
>> #6 0xffffffff810a94e6 in worker_thread (__worker=0xffff880fd86efa00)
>> at kernel/workqueue.c:2380
>> #7 0xffffffff810b098f in kthread (_create=0xffff8801695f7d38) at
>> kernel/kthread.c:202
>> #8 <signal handler called>
>
I don't know, I didn't dive into gdb for this inspection. Frankly I
didn't have much knowledge on how gdb unwinded each stack frame. All I
know is, gdb has different handlers for processing each frame
unwinding, such as the one to process function inline. So I guess for
this case, gdb thinks it is some signal processing routine.
Thanks for the explanation, Tao.
One more question:
Will the backtrace('#8 <signal handler called>') be printed if my changes are applied?
Thanks
Lianbo
Personally I'm OK with this, if you see the original stack trace:
#7 SyS_epoll_wait (epfd=<optimized out>, events=140721208415648,
maxevents=29, timeout=4294967295) at fs/eventpoll.c:2008
#8 <signal handler called>
#9 0x00007f0449407923 in ?? ()
It usually happens right before the syscall, which is switching from
user space to kernel space. Maybe it looks like a signal or interrupt
handling routine. Just my guess...
> ^^^^^^^^^^^^^^^^^^^^^^^^
>
> What's this for?
>
> Thanks
> Lianbo
>
>>
>> crash>
>>
>> What do you think?
>>
>> Thanks,
>> Tao Liu
>>
>> >
>> > Thanks
>> > Lianbo
>> >
>> >>
>> >> Thanks
>> >> Lianbo
>> >>
>> >>>
>> >>> >
>> >>> > True. Though it seems to be okay for it to print the last frame with a
>> >>> > non-kernel address, as in this snippet from gdb:
>> >>> >
>> >>> > for (fi = trailing; fi && count--; fi = get_prev_frame (fi))
>> >>> > ...
>> >>> > print_frame_info (fp_opts, fi, 1, LOCATION, 1, 0);
>> >>> >
>> >>> > Seems that frame #13, fi was not NULL.
>> >>> >
>> >>> > Seeing Tao's change, it compares the current frame's NIP/PC to see if
>> >>> > it should return NULL (which I think is nice and works). Here the
>> >>> > 'this_frame' would have been frame 12, (which would have called
>> >>> > `'get_prev_frame' to get the frame 13)
>> >>> >
>> >>> > ```
>> >>> > frame_pc_p = get_frame_pc_if_available (this_frame, &frame_pc);
>> >>> > #ifdef CRASH_MERGE
>> >>> > if (!is_kvaddr(frame_pc)) {
>> >>> > return NULL;
>> >>> > }
>> >>> > #endif
>> >>> > ```
>> >>> >
>> >>> > Tao's condition will hit when 'get_prev_frame(this_frame=frame#13)' will
>> >>> > be called to get the frame #14, which will return NULL and hence break
>> >>> > out of the loop.
>> >>> >
>> >>> > This is based on what I recall and a quick look at the implementation,
>> >>> > please feel free to correct Lianbo/Tao.
>> >>>
>> >>> Thanks Aditya for the detailed inspection, which I didn't dive into.
>> >>> When I notice the last frame to be the userspace address, I just keep
>> >>> it as it is.
>> >>>
>> >>> Thanks,
>> >>> Tao Liu
>> >>>
>> >>>
>> >>> >
>> >>> > Thanks,
>> >>> > Aditya Gupta
>> >>> >
>> >>> >
>> >>> > >
>> >>> > >
>> >>> > > Thanks
>> >>> > > Lianbo
>> >>> > >
>> >>> > >
>> >>> > > > Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
>> >>> > > > Cc: Hari Bathini <hbathini@linux.ibm.com>
>> >>> > > > Cc: Mahesh J Salgaonkar <mahesh@linux.ibm.com>
>> >>> > > > Cc: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
>> >>> > > > Cc: Lianbo Jiang <lijiang@redhat.com>
>> >>> > > > Cc: HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab@nec.com>
>> >>> > > > Cc: Tao Liu <ltao@redhat.com>
>> >>> > > > Cc: Alexey Makhalov <alexey.makhalov@broadcom.com>
>> >>> > > > Signed-off-by: Tao Liu <ltao@redhat.com>
>> >>> > > > ---
>> >>> > > > defs.h | 1 +
>> >>> > > > gdb-10.2.patch | 26 ++++++++++++++++++++++++++
>> >>> > > > gdb_interface.c | 6 ++++++
>> >>> > > > 3 files changed, 33 insertions(+)
>> >>> > > >
>> >>> > > > diff --git a/defs.h b/defs.h
>> >>> > > > index 012ffdc..c0e6a29 100644
>> >>> > > > --- a/defs.h
>> >>> > > > +++ b/defs.h
>> >>> > > > @@ -7902,6 +7902,7 @@ extern unsigned char *gdb_prettyprint_arrays;
>> >>> > > > extern unsigned int *gdb_repeat_count_threshold;
>> >>> > > > extern unsigned char *gdb_stop_print_at_null;
>> >>> > > > extern unsigned int *gdb_output_radix;
>> >>> > > > +int is_kvaddr(ulong);
>> >>> > > >
>> >>> > > > /*
>> >>> > > > * gdb/top.c
>> >>> > > > diff --git a/gdb-10.2.patch b/gdb-10.2.patch
>> >>> > > > index 0bed96a..3ed40c0 100644
>> >>> > > > --- a/gdb-10.2.patch
>> >>> > > > +++ b/gdb-10.2.patch
>> >>> > > > @@ -16171,3 +16171,29 @@ exit 0
>> >>> > > > }
>> >>> > > >
>> >>> > > > /*
>> >>> > > > +--- gdb-10.2/gdb/frame.c.orig
>> >>> > > > ++++ gdb-10.2/gdb/frame.c
>> >>> > > > +@@ -2331,6 +2331,10 @@ inside_entry_func (frame_info *this_frame)
>> >>> > > > + This function should not contain target-dependent tests, such as
>> >>> > > > + checking whether the program-counter is zero. */
>> >>> > > > +
>> >>> > > > ++#ifdef CRASH_MERGE
>> >>> > > > ++extern "C" int is_kvaddr(ulong);
>> >>> > > > ++#endif
>> >>> > > > ++
>> >>> > > > + struct frame_info *
>> >>> > > > + get_prev_frame (struct frame_info *this_frame)
>> >>> > > > + {
>> >>> > > > +@@ -2353,7 +2357,11 @@ get_prev_frame (struct frame_info *this_frame)
>> >>> > > > + get_frame_id (this_frame);
>> >>> > > > +
>> >>> > > > + frame_pc_p = get_frame_pc_if_available (this_frame, &frame_pc);
>> >>> > > > +-
>> >>> > > > ++#ifdef CRASH_MERGE
>> >>> > > > ++ if (!is_kvaddr(frame_pc)) {
>> >>> > > > ++ return NULL;
>> >>> > > > ++ }
>> >>> > > > ++#endif
>> >>> > > > + /* tausq/2004-12-07: Dummy frames are skipped because it doesn't make
>> >>> > > > much
>> >>> > > > + sense to stop unwinding at a dummy frame. One place where a dummy
>> >>> > > > + frame may have an address "inside_main_func" is on HPUX. On HPUX,
>> >>> > > > the
>> >>> > > > diff --git a/gdb_interface.c b/gdb_interface.c
>> >>> > > > index b13d5fd..e76ecc6 100644
>> >>> > > > --- a/gdb_interface.c
>> >>> > > > +++ b/gdb_interface.c
>> >>> > > > @@ -947,6 +947,12 @@ gdb_lookup_module_symbol(ulong addr, ulong *offset)
>> >>> > > > }
>> >>> > > > }
>> >>> > > >
>> >>> > > > +int
>> >>> > > > +is_kvaddr(ulong addr)
>> >>> > > > +{
>> >>> > > > + return IS_KVADDR(addr);
>> >>> > > > +}
>> >>> > > > +
>> >>> > > > /*
>> >>> > > > * Used by gdb_interface() to catch gdb-related errors, if desired.
>> >>> > > > */
>> >>> > > > --
>> >>> > > > 2.40.1
>> >>> > > >
>> >>> >
>> >>>
>>