Re: [lkp-robot] [x86/cpu_entry_area] 10043e02db: kernel_BUG_at_arch/x86/mm/physaddr.c

[Thomas Gleixner <tglx@linutronix.de>]

On Tue, 26 Dec 2017, kernel test robot wrote:

> 
> FYI, we noticed the following commit (built with gcc-6):
> 
> commit: 10043e02db7f8a4161f76434931051e7d797a5f6 ("x86/cpu_entry_area: Add debugstore entries to cpu_entry_area")
> https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git WIP.x86/pti

...

> [    0.000000] kernel BUG at arch/x86/mm/physaddr.c:27!
> PANIC: early exception 0x06 IP 10:ffffffff8115586f error 0 cr2 0xffff88000e468000
> [    0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 4.14.0-00160-g10043e02 #1
> [    0.000000] task: ffffffff8a4683c0 task.stack: ffffffff8a400000
> [    0.000000] RIP: 0010:__phys_addr+0x268/0x276
> [    0.000000] RSP: 0000:ffffffff8a407bd8 EFLAGS: 00010002 ORIG_RAX: 0000000000000000
> [    0.000000] RAX: 0000000000000000 RBX: 0000780000000000 RCX: 1ffffffff17a9a01
> [    0.000000] RDX: dffffc0000000000 RSI: 0000000000000003 RDI: ffffffff8bd4d340
> [    0.000000] RBP: ffffffff8a407bf8 R08: 0000000000000001 R09: ffffffff8a407a48
> [    0.000000] R10: ffff880000010000 R11: ffff880000010fff R12: 0000000000000001
> [    0.000000] R13: 0000000000000001 R14: 0000000000000000 R15: fffffbd00c401000
> [    0.000000] FS:  0000000000000000(0000) GS:ffffffff8cb4d000(0000) knlGS:0000000000000000
> [    0.000000] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [    0.000000] CR2: ffff88000e468000 CR3: 000000000cde8000 CR4: 00000000000406b0
> [    0.000000] Call Trace:
> [    0.000000]  kasan_populate_shadow+0x3f2/0x497

So this dies simply because kasan_populate_shadow() runs out of memory and
has no sanity check whatsoever.

static __init void *early_alloc(size_t size, int nid)
{
        return memblock_virt_alloc_try_nid_nopanic(size, size,
                __pa(MAX_DMA_ADDRESS), BOOTMEM_ALLOC_ACCESSIBLE, nid);
}

kasan_populate_pmd()
{
	.....

                p = early_alloc(PAGE_SIZE, nid);
                entry = pfn_pte(PFN_DOWN(__pa(p)), PAGE_KERNEL);

I've instrumented the whole thing and early_alloc() returns NULL at some
point and then __pa(NULL) dies in the VIRTUAL_DEBUG code. Well, it would
die with VIRTUAL_DEBUG=n as well at some other place.

Not really a problem caused by the patch above, it's merily exposing a code
path which relies blindly on "enough memory available" assumptions.

Throwing more memory at the VM makes the problem go away...

Thanks,

	tglx