On Thu, Dec 21, 2023 at 1:53 AM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> On Wed, 20 Dec 2023 13:28:39 +0800 "jiajun.xie" <jiajun.xie.sh@xxxxxxxxx> wrote:
>
> > From: Jiajun Xie <jiajun.xie.sh@xxxxxxxxx>
> >
> > The bug happens when highest bit of holebegin is 1, suppose
> > holebign is 0x8000000111111000, after shift, hba would be
> > 0xfff8000000111111, then vma_interval_tree_foreach would look
> > it up fail or leads to the wrong result.
> >
> > error call seq e.g.:
> > - mmap(..., offset=0x8000000111111000)
> > |- syscall(mmap, ... unsigned long, off):
> > |- ksys_mmap_pgoff( ... , off >> PAGE_SHIFT);
> >
> > here pgoff is correctly shifted to 0x8000000111111,
> > but pass 0x8000000111111000 as holebegin to unmap
> > would then cause terrible result, as shown below:
> >
> > - unmap_mapping_range(..., loff_t const holebegin)
> > |- pgoff_t hba = holebegin >> PAGE_SHIFT;
> > /* hba = 0xfff8000000111111 unexpectedly */
> >
> > turn holebegin to be unsigned first would fix the bug.
> >
>
> Thanks. Are you able to describe the runtime effects of this
> (obviously bad, but it's good to spell it out) and under what
> circumstances it occurs?
Thanks for the quick reply.
The issue happens in Heterogeneous computing, where the
device(e.g. gpu) and host share the same virtual address space.
A simple workflow pattern which hit the issue is:
/* host */
1. userspace first mmap a file backed VA range with specified offset.
e.g. (offset=0x800..., mmap return: va_a)
2. write some data to the corresponding sys page
e.g. (va_a = 0xAABB)
/* device */
3. gpu workload touches VA, triggers gpu fault and notify the host.
/* host */
4. reviced gpu fault notification, then it will:
4.1 unmap host pages and also takes care of cpu tlb
(use unmap_mapping_range with offset=0x800...)
4.2 migrate sys page to device
4.3 setup device page table and resolve device fault.
/* device */
5. gpu workload continued, it accessed va_a and got 0xAABB.
6. gpu workload continued, it wrote 0xBBCC to va_a.
/* host */
7. userspace access va_a, as expected, it will:
7.1 trigger cpu vm fault.
7.2 driver handling fault to migrate gpu local page to host.
8. userspace then could correctly get 0xBBCC from va_a
9. done
But in step 4.1, if we hitted the bug this patch mentioned, then user space
would never trigger cpu fault, and still get the old value: 0xAABB.
