Messages in this thread Patch in this message | | | Date | Wed, 31 Jul 2002 00:35:00 -0400 | From | Brian Gerst <> | Subject | Re: [PATCH] fix x86 page table init |
| |
Dan Aloni wrote: > On Tue, Jul 30, 2002 at 09:45:17PM -0400, Brian Gerst wrote: > >>The recent changes to the x86 page table init reintroduced a bug with >>the 4k pagetables. The page table must be filled with entries before it >>is inserted into the pmd or else you risk a tlb miss on a kernel code >>page causing an oops. This patch takes a different approach than before >>- it allows for reuse of the boot pagetable pages instead of allocating >>new ones. > > > In the patch below, isn't there a bootmem page leak in case !pmd_none(*pmd)?
Whoops. That's what happens when one needs to recreate a patch from memory. Revised patch attached.
-- Brian Gerst diff -urN linux-bk/arch/i386/mm/init.c linux/arch/i386/mm/init.c --- linux-bk/arch/i386/mm/init.c Tue Jul 30 20:59:27 2002 +++ linux/arch/i386/mm/init.c Wed Jul 31 00:32:37 2002 @@ -70,10 +70,14 @@ */ static pte_t * __init one_page_table_init(pmd_t *pmd) { - pte_t *page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE); - set_pmd(pmd, __pmd(__pa(page_table) | _KERNPG_TABLE)); - if (page_table != pte_offset_kernel(pmd, 0)) - BUG(); + pte_t *page_table; + if (pmd_none(*pmd)) { + page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE); + set_pmd(pmd, __pmd(__pa(page_table) | _KERNPG_TABLE)); + if (page_table != pte_offset_kernel(pmd, 0)) + BUG(); + } else + page_table = pte_offset_kernel(pmd, 0); return page_table; } @@ -107,9 +111,7 @@ pmd = pmd_offset(pgd, vaddr); for (; (pmd_ofs < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_ofs++) { - if (pmd_none(*pmd)) - one_page_table_init(pmd); - + one_page_table_init(pmd); vaddr += PMD_SIZE; } pmd_ofs = 0; | |