Hi,Lianbo
Thank you for your feedback.
My kernel version is 5.10.59, and this issue is happened when the
ramdumps without vmcoreinfo.
I guess your scene may have vmcore,the following logic helps to
correctly locate MODULES/VMALLOC ranges with right kernel version.
279 ms->kimage_end = (sp ? sp->value : 0);
280
281 if (ms->struct_page_size && (r =
arm64_get_va_range(ms))) {------------->//If there is a vmcore, it
will go to the following codes
282 /* We can get all the
MODULES/VMALLOC/VMEMMAP ranges now.*/
283 ms->modules_vaddr = r->modules_vaddr;
284 ms->modules_end =
r->modules_end - 1;
285 ms->vmalloc_start_addr =
r->vmalloc_start_addr;
286 ms->vmalloc_end =
r->vmalloc_end - 1;
287 ms->vmemmap_vaddr = r->vmemmap_vaddr;
288 ms->vmemmap_end =
r->vmemmap_end - 1;
289 } else if (ms->VA_BITS_ACTUAL)
{------------>//unable to locate the version and other information
through vmcore
}
lijiang <lijiang@redhat.com> 于2022年6月29日周三 14:21写道：
>
> Hi, Qianli
> On Tue, Jun 28, 2022 at 9:55 AM lijiang <lijiang@redhat.com> wrote:
>>
>> Hi, Kazu and Qianli
>>
>> On Tue, Jun 28, 2022 at 9:17 AM HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab@nec.com> wrote:
>>>
>>> Hi Qianli,
>>>
>>> thanks for the patch and explanation. I was off.
>>>
>>> On 2022/06/27 11:24, qianli zhao wrote:
>>> > Hi,Kazu
>>> >
>>> > Would you like to help review this patch?
>>>
>>> Sure, I think I can review it this week.
>>>
>>> Lianbo, can you possibly reproduce and test this?
>>
>>
>> OK, I will have a look and give feedback later.
>
>
> Could you please point out the kernel version? I tried it with the latest kernel and did not reproduce this issue when disabling the kaslr feature(# CONFIG_RANDOMIZE_BASE is not set or nokaslr)
>
> crash > help -m
> ..
> vmalloc_start_addr: ffff800008000000
> vmalloc_end: fffffbffefffffff
> modules_vaddr: ffff800000000000
> modules_end: ffff800007ffffff
> vmemmap_vaddr: fffffc0000000000
> vmemmap_end: fffffdffffffffff
> ...
>
> And the following information is dumped from the kernel(commit: 941e3e791269)
> ...
> [ 0.000000] Virtual kernel memory layout:
> [ 0.000000] modules : 0xffff800000000000 - 0xffff800008000000 ( 128 MB)
> [ 0.000000] vmalloc : 0xffff800008000000 - 0xfffffbfff0000000 (126975 GB)
> ...
> [ 0.000000] vmemmap : 0xfffffc0000000000 - 0xfffffe0000000000 ( 2048 GB maximum)
> [ 0.000000] 0xfffffc000000bc00 - 0xfffffc023df40000 ( 9183 MB actual)
>
> I'm wondering if that can be only reproduced on the old kernel, right? Or did I miss anything else?
>
> Thanks.
> Lianbo
>
>>
>>>
>>> Kazu
>>>
>>> >
>>> > qianli zhao <zhaoqianligood@gmail.com> 于2022年6月24日周五 10:56写道：
>>> >
>>> >>
>>> >> Hi,all
>>> >>
>>> >> Here's some explanation for this patch
>>> >>
>>> >> Without patch:
>>> >> Consider the following scenario
>>> >> ->arm64_init(PRE_GDB)
>>> >> case PRE_GDB:
>>> >> ...
>>> >> 292 } else if (ms->VA_BITS_ACTUAL) {
>>> >> 293 ms->modules_vaddr =
>>> >> (st->_stext_vmlinux & TEXT_OFFSET_MASK) -
>>> >> ARM64_MODULES_VSIZE;-->//ms->modules_vaddr=0xfffffffff8000000
>>> >> 294 ms->modules_end =
>>> >> ms->modules_vaddr + ARM64_MODULES_VSIZE
>>> >> -1;--->//ms->modules_end=0xffffffffffffffff
>>> >> 295 ms->vmalloc_start_addr =
>>> >> ms->modules_end + 1;--->//ms->vmalloc_start_addr=0
>>> >> 296 } else {
>>> >> ....
>>> >> }
>>> >> arm64_calc_kimage_voffset();
>>> >> .....
>>> >>
>>> >> Since arm64_calc_kimage_voffset() depends on vmalloc_start_addr,
>>> >> kimage_voffset cannot be calculated correctly.
>>> >>
>>> >> st->_stext_vmlinux can be initialized in numeric_forward(),just set
>>> >> st->_stext_vmlinux to UNINITIALIZED.
>>> >>
>>> >> ============
>>> >> log as below:
>>> >>
>>> >> $ ~/crash/crash/crash vmlinux DDRCS0.bin@0x80000000 --machdep vabits_actual=48
>>> >>
>>> >> crash 8.0.1++
>>> >> Copyright (C) 2002-2022 Red Hat, Inc.
>>> >> Copyright (C) 2004, 2005, 2006, 2010 IBM Corporation
>>> >> Copyright (C) 1999-2006 Hewlett-Packard Co
>>> >> Copyright (C) 2005, 2006, 2011, 2012 Fujitsu Limited
>>> >> Copyright (C) 2006, 2007 VA Linux Systems Japan K.K.
>>> >> Copyright (C) 2005, 2011, 2020-2022 NEC Corporation
>>> >> Copyright (C) 1999, 2002, 2007 Silicon Graphics, Inc.
>>> >> Copyright (C) 1999, 2000, 2001, 2002 Mission Critical Linux, Inc.
>>> >> Copyright (C) 2015, 2021 VMware, Inc.
>>> >> This program is free software, covered by the GNU General Public License,
>>> >> and you are welcome to change it and/or distribute copies of it under
>>> >> certain conditions. Enter "help copying" to see the conditions.
>>> >> This program has absolutely no warranty. Enter "help warranty" for details.
>>> >>
>>> >> NOTE: setting vabits_actual to: 48
>>> >>
>>> >> WARNING: kimage_voffset cannot be determined from the dumpfile.
>>> >> Try using the command line option: --machdep kimage_voffset=<addr>
>>> >> GNU gdb (GDB) 10.2
>>> >> Copyright (C) 2021 Free Software Foundation, Inc.
>>> >> License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
>>> >> This is free software: you are free to change and redistribute it.
>>> >> There is NO WARRANTY, to the extent permitted by law.
>>> >> Type "show copying" and "show warranty" for details.
>>> >> This GDB was configured as "--host=x86_64-pc-linux-gnu
>>> >> --target=aarch64-elf-linux".
>>> >> Type "show configuration" for configuration details.
>>> >> Find the GDB manual and other documentation resources online at:
>>> >> <http://www.gnu.org/software/gdb/documentation/>.
>>> >>
>>> >> For help, type "help".
>>> >> Type "apropos word" to search for commands related to "word"...
>>> >>
>>> >> crash: read error: kernel virtual address: ffff80001083d4a0 type:
>>> >> "kernel_config_data"
>>> >> WARNING: cannot read kernel_config_data
>>> >> crash: read error: kernel virtual address: ffff80001170e798 type: "possible"
>>> >> WARNING: cannot read cpu_possible_map
>>> >> crash: read error: kernel virtual address: ffff80001170e7a8 type: "present"
>>> >> WARNING: cannot read cpu_present_map
>>> >> crash: read error: kernel virtual address: ffff80001170e788 type: "online"
>>> >> WARNING: cannot read cpu_online_map
>>> >> crash: read error: kernel virtual address: ffff80001170e7c0 type: "active"
>>> >> WARNING: cannot read cpu_active_map
>>> >> crash: read error: kernel virtual address: ffff8000122e00f0 type:
>>> >> "shadow_timekeeper xtime_sec"
>>> >> crash: read error: kernel virtual address: ffff80001171dc04 type: "init_uts_ns"
>>> >> crash: vmlinux and /var/tmp/ramdump_elf_m2ivkg do not match!
>>> >>
>>> >> Usage:
>>> >>
>>> >> crash [OPTION]... NAMELIST MEMORY-IMAGE[@ADDRESS] (dumpfile form)
>>> >> crash [OPTION]... [NAMELIST] (live system form)
>>> >>
>>> >> Enter "crash -h" for details.
>>> >>
>>> >> Qianli Zhao <zhaoqianligood@gmail.com> 于2022年6月24日周五 00:14写道：
>>> >>>
>>> >>> From: Qianli Zhao <qianli.zhao@horizon.ai>
>>> >>>
>>> >>> Setting st->_stext_vmlinux to UNINITIALIZED to search for "_stext" from the vmlinux
>>> >>> Without the patch, if we do not enable kaslr, will get the wrong
>>> >>> MODULES/VMALLOC ranges, cause parsing dump failure
>>> >>>
>>> >>> Signed-off-by: Qianli Zhao <qianli.zhao@horizon.ai>
>>> >>> ---
>>> >>> arm64.c | 3 +++
>>> >>> 1 file changed, 3 insertions(+)
>>> >>>
>>> >>> diff --git a/arm64.c b/arm64.c
>>> >>> index 0f615cf..4458a66 100644
>>> >>> --- a/arm64.c
>>> >>> +++ b/arm64.c
>>> >>> @@ -149,6 +149,9 @@ arm64_init(int when)
>>> >>>
>>> >>> ms = machdep->machspec;
>>> >>>
>>> >>> + if (ms->VA_BITS_ACTUAL)
>>> >>> + st->_stext_vmlinux = UNINITIALIZED;
>>> >>> +
>>> >>> if (!ms->kimage_voffset && STREQ(pc->live_memsrc, "/dev/crash"))
>>> >>> ioctl(pc->mfd, DEV_CRASH_ARCH_DATA, &ms->kimage_voffset);
>>> >>>
>>> >>> --
>>> >>> 2.17.1
>>> >>>