Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1493784 > unrolled thread
| Started by | Stephen Rothwell <sfr@canb.auug.org.au> |
|---|---|
| First post | 2016-09-30 01:30 +0200 |
| Last post | 2016-10-11 15:20 +0200 |
| Articles | 4 — 4 participants |
Back to article view | Back to linux.kernel
This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by
below is the oldest one visible, not the original post.
Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user Stephen Rothwell <sfr@canb.auug.org.au> - 2016-09-30 01:30 +0200
Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user Mathieu OTHACEHE <m.othacehe@gmail.com> - 2016-10-09 11:50 +0200
Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user Michal Marek <mmarek@suse.cz> - 2016-10-11 15:20 +0200
Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user Michal Marek <mmarek@suse.com> - 2016-10-11 15:20 +0200
| From | Stephen Rothwell <sfr@canb.auug.org.au> |
|---|---|
| Date | 2016-09-30 01:30 +0200 |
| Subject | Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user |
| Message-ID | <smSW5-2lb-9@gated-at.bofh.it> |
Hi all,
On Tue, 20 Sep 2016 10:55:32 +0800 kernel test robot <xiaolong.ye@intel.com> wrote:
>
> FYI, we noticed the following commit:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> commit 784d5699eddc55878627da20d3fe0c8542e2f1a2 ("x86: move exports to actual definitions")
>
> in testcase: boot
>
> on test machine: qemu-system-x86_64 -enable-kvm -cpu host -smp 2 -m 4G
>
> caused below changes:
>
> [ 4.386609] ip_tables: no symbol version for _copy_from_user
> [ 4.387902] ip_tables: Unknown symbol _copy_from_user (err -22)
> [ 4.389219] ip_tables: no symbol version for __put_user_2
> [ 4.390473] ip_tables: Unknown symbol __put_user_2 (err -22)
> [ 4.391760] ip_tables: no symbol version for __sw_hweight32
> [ 4.393031] ip_tables: Unknown symbol __sw_hweight32 (err -22)
> [ 4.394332] ip_tables: no symbol version for memcpy
> [ 4.395508] ip_tables: Unknown symbol memcpy (err -22)
> [ 4.396729] ip_tables: no symbol version for __fentry__
> [ 4.397937] ip_tables: Unknown symbol __fentry__ (err -22)
> [ 4.399234] ip_tables: no symbol version for _copy_to_user
> [ 4.400488] ip_tables: Unknown symbol _copy_to_user (err -22)
> [ 4.405835] systemd[1]: Failed to insert module 'ip_tables': Invalid argument
>
>
>
> Thanks,
> Xiaolong
Do we have any resolution of this?
--
Cheers,
Stephen Rothwell
[toc] | [next] | [standalone]
| From | Mathieu OTHACEHE <m.othacehe@gmail.com> |
|---|---|
| Date | 2016-10-09 11:50 +0200 |
| Subject | Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user |
| Message-ID | <sqiU2-2sU-7@gated-at.bofh.it> |
| In reply to | #1493784 |
Hi, > Do we have any resolution of this? I got the exact same problem if CONFIG_MODVERSIONS is enabled. Here is my understanding. All the symbols exported using assembler macro EXPORT_SYMBOL will need a __crc_<symbol> to be defined later on (include/asm-generic/export.h). This is what genksyms should do but it only runs on *.c files. So all the symbols exported from *.S files won't get crc. modpost is complaining about missing crc and it will not include those symbols in *.mod.c modversion_info struct. Then, we got "Unknown symbol" errors when modules are inserted at runtime. I fixed it locally by setting __crc_<symbol> to a constant value in export.h. But it may be a better idea to have genksyms run on *.S files ? Thanks, Mathieu
[toc] | [prev] | [next] | [standalone]
| From | Michal Marek <mmarek@suse.cz> |
|---|---|
| Date | 2016-10-11 15:20 +0200 |
| Subject | Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user |
| Message-ID | <sr58l-7aD-7@gated-at.bofh.it> |
| In reply to | #1493784 |
Dne 30.9.2016 v 01:26 Stephen Rothwell napsal(a):
> Hi all,
>
> On Tue, 20 Sep 2016 10:55:32 +0800 kernel test robot <xiaolong.ye@intel.com> wrote:
>>
>> FYI, we noticed the following commit:
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
>> commit 784d5699eddc55878627da20d3fe0c8542e2f1a2 ("x86: move exports to actual definitions")
>>
>> in testcase: boot
>>
>> on test machine: qemu-system-x86_64 -enable-kvm -cpu host -smp 2 -m 4G
>>
>> caused below changes:
>>
>> [ 4.386609] ip_tables: no symbol version for _copy_from_user
>> [ 4.387902] ip_tables: Unknown symbol _copy_from_user (err -22)
>> [ 4.389219] ip_tables: no symbol version for __put_user_2
>> [ 4.390473] ip_tables: Unknown symbol __put_user_2 (err -22)
>> [ 4.391760] ip_tables: no symbol version for __sw_hweight32
>> [ 4.393031] ip_tables: Unknown symbol __sw_hweight32 (err -22)
>> [ 4.394332] ip_tables: no symbol version for memcpy
>> [ 4.395508] ip_tables: Unknown symbol memcpy (err -22)
>> [ 4.396729] ip_tables: no symbol version for __fentry__
>> [ 4.397937] ip_tables: Unknown symbol __fentry__ (err -22)
>> [ 4.399234] ip_tables: no symbol version for _copy_to_user
>> [ 4.400488] ip_tables: Unknown symbol _copy_to_user (err -22)
>> [ 4.405835] systemd[1]: Failed to insert module 'ip_tables': Invalid argument
>>
>>
>>
>> Thanks,
>> Xiaolong
>
> Do we have any resolution of this?
With which .config is it happening? In my testing, the undefined CRCs
show up as 0x00000000 in Module.symvers and they are also 0 in the
___kcrctab* sections in vmlinux. So the loader matches the requested and
provided symbol's CRC of 0 and is happy. But I guess I'm missing something.
Thanks,
Michal
[toc] | [prev] | [next] | [standalone]
| From | Michal Marek <mmarek@suse.com> |
|---|---|
| Date | 2016-10-11 15:20 +0200 |
| Subject | Re: [lkp] [x86] 784d5699ed: kmsg.ip_tables:no_symbol_version_for_copy_from_user |
| Message-ID | <sr58l-7aD-5@gated-at.bofh.it> |
| In reply to | #1498874 |
Dne 11.10.2016 v 15:10 Michal Marek napsal(a):
> Dne 30.9.2016 v 01:26 Stephen Rothwell napsal(a):
>> Hi all,
>>
>> On Tue, 20 Sep 2016 10:55:32 +0800 kernel test robot <xiaolong.ye@intel.com> wrote:
>>>
>>> FYI, we noticed the following commit:
>>>
>>> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
>>> commit 784d5699eddc55878627da20d3fe0c8542e2f1a2 ("x86: move exports to actual definitions")
>>>
>>> in testcase: boot
>>>
>>> on test machine: qemu-system-x86_64 -enable-kvm -cpu host -smp 2 -m 4G
>>>
>>> caused below changes:
>>>
>>> [ 4.386609] ip_tables: no symbol version for _copy_from_user
>>> [ 4.387902] ip_tables: Unknown symbol _copy_from_user (err -22)
>>> [ 4.389219] ip_tables: no symbol version for __put_user_2
>>> [ 4.390473] ip_tables: Unknown symbol __put_user_2 (err -22)
>>> [ 4.391760] ip_tables: no symbol version for __sw_hweight32
>>> [ 4.393031] ip_tables: Unknown symbol __sw_hweight32 (err -22)
>>> [ 4.394332] ip_tables: no symbol version for memcpy
>>> [ 4.395508] ip_tables: Unknown symbol memcpy (err -22)
>>> [ 4.396729] ip_tables: no symbol version for __fentry__
>>> [ 4.397937] ip_tables: Unknown symbol __fentry__ (err -22)
>>> [ 4.399234] ip_tables: no symbol version for _copy_to_user
>>> [ 4.400488] ip_tables: Unknown symbol _copy_to_user (err -22)
>>> [ 4.405835] systemd[1]: Failed to insert module 'ip_tables': Invalid argument
>>>
>>>
>>>
>>> Thanks,
>>> Xiaolong
>>
>> Do we have any resolution of this?
>
> With which .config is it happening?
Or maybe with which toolchain is the kernel built.
Thanks,
Michal
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web