Re: [linux-next:pending-fixes] [mm/hugetlb] d3fe1a0a02: WARNING:at_include/linux/hugetlb.h:#huge_pte_lockptr

[David Hildenbrand <david@redhat.com>]

On 31.07.24 11:28, David Hildenbrand wrote:
> On 31.07.24 10:16, David Hildenbrand wrote:
>> On 31.07.24 09:20, kernel test robot wrote:
>>>
>>>
>>> Hello,
>>>
>>> kernel test robot noticed "WARNING:at_include/linux/hugetlb.h:#huge_pte_lockptr" on:
>>>
>>> commit: d3fe1a0a0272723474056e3000717d69923ad901 ("mm/hugetlb: fix hugetlb vs. core-mm PT locking")
>>> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git pending-fixes
>>>
>>> [test failed on linux-next/master cd19ac2f903276b820f5d0d89de0c896c27036ed]
>>>
>>> in testcase: trinity
>>> version: trinity-i386-abe9de86-1_20230429
>>> with following parameters:
>>>
>>> 	runtime: 300s
>>> 	group: group-00
>>> 	nr_groups: 5
>>>
>>>
>>>
>>> compiler: gcc-11
>>> test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
>>>
>>> (please refer to attached dmesg/kmsg for entire log/backtrace)
>>>
>>>
>>> +------------------------------------------------------+------------+------------+
>>> |                                                      | e98970a1d2 | d3fe1a0a02 |
>>> +------------------------------------------------------+------------+------------+
>>> | WARNING:at_include/linux/hugetlb.h:#huge_pte_lockptr | 0          | 12         |
>>> | EIP:huge_pte_lockptr                                 | 0          | 12         |
>>> +------------------------------------------------------+------------+------------+
>>>
>>>
>>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>>> the same patch/commit), kindly add following tags
>>> | Reported-by: kernel test robot <oliver.sang@intel.com>
>>> | Closes: https://lore.kernel.org/oe-lkp/202407311421.95f5c759-lkp@intel.com
>>>
>>>
>>> [   39.570466][ T3479] ------------[ cut here ]------------
>>> [ 39.570921][ T3479] WARNING: CPU: 0 PID: 3479 at include/linux/hugetlb.h:948 huge_pte_lockptr (include/linux/hugetlb.h:948 (discriminator 1))
>>
>> Helpful! Likely triggers because in that kernel cfg P4D_SIZE is 0.
> 
> Hm, with !CONFIG_X86_5LEVEL we should have
> 
> #define PGDIR_SHIFT 39
> #define P4D_SHIFT PGDIR_SHIFT
> #define P4D_SIZE (1UL << P4D_SHIFT)
> 
> ... getting a hugepage size >= P4D_SIZE seems pretty much impossible.
> 
> Anyhow, taking the easy route to be done with his hugetlb crap and live
> a happy live, I'll drop the warning and just handle it explicitly.

Realizing that this is 32bit with CONFIG_PGTABLE_LEVELS=2:

So

#define PGDIR_SHIFT	22

and so should be then P4D_SHIFT,PUD_SHIFT,PMD_SHIFT

So we would get "hugepage size >= 4 MiB", which really just is the 
default PMD size and we would want to do pmd_lockptr() with "hugepage 
size == PMD_SIZE".

-- 
Cheers,

David / dhildenb