>On Tue, Jul 6, 2021 at 10:41 AM 王擎 <wangqing@xxxxxxxx> wrote: >> >> >> >On Mon, Jul 5, 2021 at 9:57 PM Wang Qing <wangqing@xxxxxxxx> wrote: >> >> >> >> Use GFP_ATOMIC when local_lock_irqsave in __alloc_pages_bulk >> >> >> >> Reported-by: syzbot+e45919db2eab5e837646@xxxxxxxxxxxxxxxxxxxxxxxxx >> >> Signed-off-by: Wang Qing <wangqing@xxxxxxxx> >> >> --- >> >> mm/page_alloc.c | 2 +- >> >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> >> >> diff --git a/mm/page_alloc.c b/mm/page_alloc.c >> >> index d6e94cc..3016ba5 >> >> --- a/mm/page_alloc.c >> >> +++ b/mm/page_alloc.c >> >> @@ -5309,7 +5309,7 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, >> >> } >> >> nr_account++; >> >> >> >> - prep_new_page(page, 0, gfp, 0); >> >> + prep_new_page(page, 0, gfp | GFP_ATOMIC, 0); >> > >> >Hi Wang Qing, >> > >> >I didn't get the point here. IIUC, prep_new_page() will not allocate >> >memory. So why do we need GFP_ATOMIC? What I missed here? >> > >> >Thanks. >> >> prep_new_page() will allocate memory in some scenarios. For details, >> you can check the bugs detected by syzkaller: >> https://syzkaller.appspot.com/bug?id=91c2030241ada0e5d21877f8f2f44c98cffc04bb >> >> Call Trace: >> __dump_stack lib/dump_stack.c:79 [inline] >> dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:96 >> ___might_sleep.cold+0x1f1/0x237 kernel/sched/core.c:9153 >> prepare_alloc_pages+0x3da/0x580 mm/page_alloc.c:5179 >> __alloc_pages+0x12f/0x500 mm/page_alloc.c:5375 >> alloc_pages+0x18c/0x2a0 mm/mempolicy.c:2272 >> stack_depot_save+0x39d/0x4e0 lib/stackdepot.c:303 >> save_stack+0x15e/0x1e0 mm/page_owner.c:120 >> __set_page_owner+0x50/0x290 mm/page_owner.c:181 >> prep_new_page mm/page_alloc.c:2445 [inline] >> __alloc_pages_bulk+0x8b9/0x1870 mm/page_alloc.c:5313 > >Got it. But I don't think the fix you mentioned above was >appropriate. What if GFP_KERNEL | GFP_ATOMIC? Yes agree, but I haven't figured out what will happen this way, the test has been passed in syzkaller. Or how about gfp | GFP_ATOMIC & ~GFP_KERNEL ? Thanks, Qing > >Thanks. > >> >> Thanks. >> >> Qing >> >> > >> >> if (page_list) >> >> list_add(&page->lru, page_list); >> >> else >> >> -- >> >> 2.7.4 >> >> >> >>
