Re: [PATCH] mm, kasan: don't poison boot memory

From: Mike Rapoport
Date: Thu Feb 25 2021 - 11:10:30 EST


On Thu, Feb 25, 2021 at 10:22:44AM -0500, George Kennedy wrote:
>
> > > > > On 2/24/2021 5:37 AM, Mike Rapoport wrote:
>
> Applied just your latest patch, but same failure.
>
> I thought there was an earlier comment (which I can't find now) that stated
> that memblock_reserve() wouldn't reserve the page, which is what's needed
> here.

Actually, I think that memblock_reserve() should be just fine, but it seems
I'm missing something in address calculation each time.

What would happen if you stuck

memblock_reserve(0xbe453000, PAGE_SIZE);

say, at the beginning of find_ibft_region()?

> [   30.308229] iBFT detected..
> [   30.308796]
> ==================================================================
> [   30.308890] BUG: KASAN: use-after-free in ibft_init+0x134/0xc33
> [   30.308890] Read of size 4 at addr ffff8880be453004 by task swapper/0/1
> [   30.308890]
> [   30.308890] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 5.11.0-f9593a0 #12
> [   30.308890] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
> 0.0.0 02/06/2015
> [   30.308890] Call Trace:
> [   30.308890]  dump_stack+0xdb/0x120
> [   30.308890]  ? ibft_init+0x134/0xc33
> [   30.308890]  print_address_description.constprop.7+0x41/0x60
> [   30.308890]  ? ibft_init+0x134/0xc33
> [   30.308890]  ? ibft_init+0x134/0xc33
> [   30.308890]  kasan_report.cold.10+0x78/0xd1
> [   30.308890]  ? ibft_init+0x134/0xc33
> [   30.308890]  __asan_report_load_n_noabort+0xf/0x20
> [   30.308890]  ibft_init+0x134/0xc33
> [   30.308890]  ? write_comp_data+0x2f/0x90
> [   30.308890]  ? ibft_check_initiator_for+0x159/0x159
> [   30.308890]  ? write_comp_data+0x2f/0x90
> [   30.308890]  ? ibft_check_initiator_for+0x159/0x159
> [   30.308890]  do_one_initcall+0xc4/0x3e0
> [   30.308890]  ? perf_trace_initcall_level+0x3e0/0x3e0
> [   30.308890]  ? unpoison_range+0x14/0x40
> [   30.308890]  ? ____kasan_kmalloc.constprop.5+0x8f/0xc0
> [   30.308890]  ? kernel_init_freeable+0x420/0x652
> [   30.308890]  ? __kasan_kmalloc+0x9/0x10
> [   30.308890]  ? __sanitizer_cov_trace_pc+0x21/0x50
> [   30.308890]  kernel_init_freeable+0x596/0x652
> [   30.308890]  ? console_on_rootfs+0x7d/0x7d
> [   30.308890]  ? __sanitizer_cov_trace_pc+0x21/0x50
> [   30.308890]  ? rest_init+0xf0/0xf0
> [   30.308890]  kernel_init+0x16/0x1d0
> [   30.308890]  ? rest_init+0xf0/0xf0
> [   30.308890]  ret_from_fork+0x22/0x30
> [   30.308890]
> [   30.308890] The buggy address belongs to the page:
> [   30.308890] page:0000000001b7b17c refcount:0 mapcount:0
> mapping:0000000000000000 index:0x1 pfn:0xbe453
> [   30.308890] flags: 0xfffffc0000000()
> [   30.308890] raw: 000fffffc0000000 ffffea0002ef9788 ffffea0002f91488
> 0000000000000000
> [   30.308890] raw: 0000000000000001 0000000000000000 00000000ffffffff
> 0000000000000000
> [   30.308890] page dumped because: kasan: bad access detected
> [   30.308890] page_owner tracks the page as freed
> [   30.308890] page last allocated via order 0, migratetype Movable,
> gfp_mask 0x100dca(GFP_HIGHUSER_MOVABLE|__GFP_ZERO), pid 204, ts 28121288605
> [   30.308890]  prep_new_page+0xfb/0x140
> [   30.308890]  get_page_from_freelist+0x3503/0x5730
> [   30.308890]  __alloc_pages_nodemask+0x2d8/0x650
> [   30.308890]  alloc_pages_vma+0xe2/0x560
> [   30.308890]  __handle_mm_fault+0x930/0x26c0
> [   30.308890]  handle_mm_fault+0x1f9/0x810
> [   30.308890]  do_user_addr_fault+0x6f7/0xca0
> [   30.308890]  exc_page_fault+0xaf/0x1a0
> [   30.308890]  asm_exc_page_fault+0x1e/0x30
> [   30.308890] page last free stack trace:
> [   30.308890]  free_pcp_prepare+0x122/0x290
> [   30.308890]  free_unref_page_list+0xe6/0x490
> [   30.308890]  release_pages+0x2ed/0x1270
> [   30.308890]  free_pages_and_swap_cache+0x245/0x2e0
> [   30.308890]  tlb_flush_mmu+0x11e/0x680
> [   30.308890]  tlb_finish_mmu+0xa6/0x3e0
> [   30.308890]  exit_mmap+0x2b3/0x540
> [   30.308890]  mmput+0x11d/0x450
> [   30.308890]  do_exit+0xaa6/0x2d40
> [   30.308890]  do_group_exit+0x128/0x340
> [   30.308890]  __x64_sys_exit_group+0x43/0x50
> [   30.308890]  do_syscall_64+0x37/0x50
> [   30.308890]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [   30.308890]
> [   30.308890] Memory state around the buggy address:
> [   30.308890]  ffff8880be452f00: ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ff ff
> [   30.308890]  ffff8880be452f80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ff ff
> [   30.308890] >ffff8880be453000: ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ff ff
> [   30.308890]                    ^
> [   30.308890]  ffff8880be453080: ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ff ff
> [   30.308890]  ffff8880be453100: ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ff ff
> [   30.308890]
> ==================================================================
>
> George
>

--
Sincerely yours,
Mike.