Re: [PATCH v2 mm-hotfixes 1/5] mm: move page table sync declarations to asm/pgalloc.h

[Harry Yoo <harry.yoo@oracle.com>]

On Mon, Jul 21, 2025 at 12:38:27PM +0100, Lorenzo Stoakes wrote:
> On Mon, Jul 21, 2025 at 11:40:10AM +0800, kernel test robot wrote:
> > Hi Harry,
> >
> > kernel test robot noticed the following build warnings:
> >
> > [auto build test WARNING on akpm-mm/mm-everything]
> >
> > url:    https://github.com/intel-lab-lkp/linux/commits/Harry-Yoo/mm-move-page-table-sync-declarations-to-asm-pgalloc-h/20250721-074448
> > base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
> > patch link:    https://lore.kernel.org/r/20250720234203.9126-2-harry.yoo%40oracle.com
> > patch subject: [PATCH v2 mm-hotfixes 1/5] mm: move page table sync declarations to asm/pgalloc.h
> > config: i386-buildonly-randconfig-003-20250721 (https://download.01.org/0day-ci/archive/20250721/202507211129.Xbn2bAOg-lkp@intel.com/config)
> > compiler: gcc-12 (Debian 12.2.0-14+deb12u1) 12.2.0
> > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250721/202507211129.Xbn2bAOg-lkp@intel.com/reproduce)
> >
> > 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 <lkp@intel.com>
> > | Closes: https://lore.kernel.org/oe-kbuild-all/202507211129.Xbn2bAOg-lkp@intel.com/
> >
> > All warnings (new ones prefixed by >>):
> >
> > >> arch/x86/mm/fault.c:265:6: warning: no previous prototype for 'arch_sync_kernel_mappings' [-Wmissing-prototypes]
> >      265 | void arch_sync_kernel_mappings(unsigned long start, unsigned long end)
> >          |      ^~~~~~~~~~~~~~~~~~~~~~~~~
> >
> 
> Looks like arch/x86/mm/fault.c, which includes linux/vmalloc.h (such an odd
> place for this decl!) needs to:
> 
> #include <asm/pgalloc.h>

But in x86-32, ARCH_PAGE_TABLE_SYNC_MASK is defined in
arch/x86/include/asm/pgtable-2,3level_types.h, which can be included
via <asm/pgtable.h> or <linux/pgtable.h>.

I think it was a mistake to move the declarations to
<asm-generic/pgalloc.h> because if a file includes <asm/pgalloc.h> but
forgets to include <linux/vmalloc.h> or <linux/pgtable.h>, then
arch_sync_kernel_mappings() will be optimized out even when it's needed.

I'll move them to <linux/pgtable.h> and let architectures
override ARCH_PAGE_TABLE_SYNC_MASK by defining their own in <asm/pgtable.h>.

> This seems to be a 32-bit build thing, as your series builds locally on my
> x86-64 machine.

Yeah, I only tested it on x86-64 (with 4 & 5 level paging)..
I was unware that I was breaking x86-32.

Thanks!

-- 
Cheers,
Harry / Hyeonggon

> > vim +/arch_sync_kernel_mappings +265 arch/x86/mm/fault.c
> >
> > 4819e15f740ec88 Joerg Roedel        2020-09-02  264
> > 1e15d374bb1cb95 Alexander Potapenko 2023-01-11 @265  void arch_sync_kernel_mappings(unsigned long start, unsigned long end)
> > f2f13a8535174db Ingo Molnar         2009-02-20  266  {
> > 86cf69f1d893d48 Joerg Roedel        2020-06-01  267  	unsigned long addr;
> > f2f13a8535174db Ingo Molnar         2009-02-20  268
> > 86cf69f1d893d48 Joerg Roedel        2020-06-01  269  	for (addr = start & PMD_MASK;
> > 86cf69f1d893d48 Joerg Roedel        2020-06-01  270  	     addr >= TASK_SIZE_MAX && addr < VMALLOC_END;
> > 86cf69f1d893d48 Joerg Roedel        2020-06-01  271  	     addr += PMD_SIZE) {
> > f2f13a8535174db Ingo Molnar         2009-02-20  272  		struct page *page;
> > f2f13a8535174db Ingo Molnar         2009-02-20  273
> > a79e53d85683c6d Andrea Arcangeli    2011-02-16  274  		spin_lock(&pgd_lock);
> > f2f13a8535174db Ingo Molnar         2009-02-20  275  		list_for_each_entry(page, &pgd_list, lru) {
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  276  			spinlock_t *pgt_lock;
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  277
> > a79e53d85683c6d Andrea Arcangeli    2011-02-16  278  			/* the pgt_lock only for Xen */
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  279  			pgt_lock = &pgd_page_get_mm(page)->page_table_lock;
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  280
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  281  			spin_lock(pgt_lock);
> > 86cf69f1d893d48 Joerg Roedel        2020-06-01  282  			vmalloc_sync_one(page_address(page), addr);
> > 617d34d9e5d8326 Jeremy Fitzhardinge 2010-09-21  283  			spin_unlock(pgt_lock);
> > f2f13a8535174db Ingo Molnar         2009-02-20  284  		}
> > a79e53d85683c6d Andrea Arcangeli    2011-02-16  285  		spin_unlock(&pgd_lock);
> > f2f13a8535174db Ingo Molnar         2009-02-20  286  	}
> > f2f13a8535174db Ingo Molnar         2009-02-20  287  }
> > f2f13a8535174db Ingo Molnar         2009-02-20  288
> >
> > --
> > 0-DAY CI Kernel Test Service
> > https://github.com/intel/lkp-tests/wiki