Re: [PATCH v22 1/4] mm: add MAP_DROPPABLE for designating always lazily freeable mappings

From: David Hildenbrand
Date: Thu Jul 11 2024 - 18:31:54 EST


On 10.07.24 05:27, David Hildenbrand wrote:
On 09.07.24 15:05, Jason A. Donenfeld wrote:
The vDSO getrandom() implementation works with a buffer allocated with a
new system call that has certain requirements:

- It shouldn't be written to core dumps.
* Easy: VM_DONTDUMP.
- It should be zeroed on fork.
* Easy: VM_WIPEONFORK.

- It shouldn't be written to swap.
* Uh-oh: mlock is rlimited.
* Uh-oh: mlock isn't inherited by forks.

It turns out that the vDSO getrandom() function has three really nice
characteristics that we can exploit to solve this problem:

1) Due to being wiped during fork(), the vDSO code is already robust to
having the contents of the pages it reads zeroed out midway through
the function's execution.

2) In the absolute worst case of whatever contingency we're coding for,
we have the option to fallback to the getrandom() syscall, and
everything is fine.

3) The buffers the function uses are only ever useful for a maximum of
60 seconds -- a sort of cache, rather than a long term allocation.

These characteristics mean that we can introduce VM_DROPPABLE, which
has the following semantics:

a) It never is written out to swap.
b) Under memory pressure, mm can just drop the pages (so that they're
zero when read back again).
c) It is inherited by fork.
d) It doesn't count against the mlock budget, since nothing is locked.

This is fairly simple to implement, with the one snag that we have to
use 64-bit VM_* flags, but this shouldn't be a problem, since the only
consumers will probably be 64-bit anyway.

This way, allocations used by vDSO getrandom() can use:

VM_DROPPABLE | VM_DONTDUMP | VM_WIPEONFORK | VM_NORESERVE

And there will be no problem with using memory when not in use, not
wiping on fork(), coredumps, or writing out to swap.

In order to let vDSO getrandom() use this, expose these via mmap(2) as
MAP_DROPPABLE.

Finally, the provided self test ensures that this is working as desired.

Acked-by: David Hildenbrand <david@xxxxxxxxxx>


I'll try to think of some corner cases we might be missing.

Sorry that I keep coming up with corner cases :) But these should be easy to handle:

1) We should disallow KSM.

diff --git a/mm/ksm.c b/mm/ksm.c
index df6bae3a5a2c..d6744183ba41 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -713,7 +713,7 @@ static bool vma_ksm_compatible(struct vm_area_struct *vma)
{
if (vma->vm_flags & (VM_SHARED | VM_MAYSHARE | VM_PFNMAP |
VM_IO | VM_DONTEXPAND | VM_HUGETLB |
- VM_MIXEDMAP))
+ VM_MIXEDMAP | VM_DROPPABLE))
return false; /* just ignore the advice */
if (vma_is_dax(vma))


We don't want to suddenly get pages that are swapbacked.


2) We should disable userfaultfd

diff --git a/include/linux/userfaultfd_k.h b/include/linux/userfaultfd_k.h
index 05d59f74fc88..a12bcf042551 100644
--- a/include/linux/userfaultfd_k.h
+++ b/include/linux/userfaultfd_k.h
@@ -218,6 +218,9 @@ static inline bool vma_can_userfault(struct vm_area_struct *vma,
{
vm_flags &= __VM_UFFD_FLAGS;
+ if (vm_flags & VM_DROPPABLE)
+ return false;
+
if ((vm_flags & VM_UFFD_MINOR) &&
(!is_vm_hugetlb_page(vma) && !vma_is_shmem(vma)))
return false;


Otherwise someone could place swapbacked pages in there (using UFFDIO_MOVE)
I think. But conceptually, I don't think userfaultfd might not make sense at
all with uffd. And if there are good reasons for it in the future, we could
enable the parts that make sense.


I think other places like khugepaged should handle it correctly (not set
swapbacked) due to your changes to folio_add_new_anon_rmap().

--
Cheers,

David / dhildenb