Re: [PATCH v3 03/13] x86/mm: Remove kernel_unmap_pages_in_pgd() and efi_cleanup_page_tables()

2016-06-21 Thread Matt Fleming
On Mon, 20 Jun, at 04:43:33PM, Andy Lutomirski wrote: > kernel_unmap_pages_in_pgd() is dangerous: if a pgd entry in > init_mm.pgd were to be cleared, callers would need to ensure that > the pgd entry hadn't been propagated to any other pgd. > > Its only caller was efi_cleanup_page_tables(), and

Re: [PATCH v3 03/13] x86/mm: Remove kernel_unmap_pages_in_pgd() and efi_cleanup_page_tables()

2016-06-21 Thread Matt Fleming
On Mon, 20 Jun, at 04:43:33PM, Andy Lutomirski wrote: > kernel_unmap_pages_in_pgd() is dangerous: if a pgd entry in > init_mm.pgd were to be cleared, callers would need to ensure that > the pgd entry hadn't been propagated to any other pgd. > > Its only caller was efi_cleanup_page_tables(), and

[PATCH v3 03/13] x86/mm: Remove kernel_unmap_pages_in_pgd() and efi_cleanup_page_tables()

2016-06-20 Thread Andy Lutomirski
kernel_unmap_pages_in_pgd() is dangerous: if a pgd entry in init_mm.pgd were to be cleared, callers would need to ensure that the pgd entry hadn't been propagated to any other pgd. Its only caller was efi_cleanup_page_tables(), and that, in turn, was unused, so just delete both functions. This

[PATCH v3 03/13] x86/mm: Remove kernel_unmap_pages_in_pgd() and efi_cleanup_page_tables()

2016-06-20 Thread Andy Lutomirski
kernel_unmap_pages_in_pgd() is dangerous: if a pgd entry in init_mm.pgd were to be cleared, callers would need to ensure that the pgd entry hadn't been propagated to any other pgd. Its only caller was efi_cleanup_page_tables(), and that, in turn, was unused, so just delete both functions. This