[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5b8c6b49-e17a-2c0b-4440-ccf3c5493cb2@huawei.com>
Date: Fri, 29 Jul 2022 10:55:15 +0800
From: Miaohe Lin <linmiaohe@...wei.com>
To: Mike Kravetz <mike.kravetz@...cle.com>, <linux-mm@...ck.org>,
<linux-kernel@...r.kernel.org>
CC: Muchun Song <songmuchun@...edance.com>,
Michal Hocko <mhocko@...e.com>, Peter Xu <peterx@...hat.com>,
Naoya Horiguchi <naoya.horiguchi@...ux.dev>,
David Hildenbrand <david@...hat.com>,
"Aneesh Kumar K . V" <aneesh.kumar@...ux.vnet.ibm.com>,
Andrea Arcangeli <aarcange@...hat.com>,
"Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
Davidlohr Bueso <dave@...olabs.net>,
Prakash Sangappa <prakash.sangappa@...cle.com>,
James Houghton <jthoughton@...gle.com>,
Mina Almasry <almasrymina@...gle.com>,
Pasha Tatashin <pasha.tatashin@...een.com>,
Axel Rasmussen <axelrasmussen@...gle.com>,
Ray Fucillo <Ray.Fucillo@...ersystems.com>,
Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [RFC PATCH v4 6/8] hugetlb: add vma based lock for pmd sharing
synchronization
On 2022/7/7 4:23, Mike Kravetz wrote:
> Allocate a rw semaphore and hang off vm_private_data for
> synchronization use by vmas that could be involved in pmd sharing. Only
> add infrastructure for the new lock here. Actual use will be added in
> subsequent patch.
>
> Signed-off-by: Mike Kravetz <mike.kravetz@...cle.com>
> ---
> include/linux/hugetlb.h | 36 +++++++++-
> kernel/fork.c | 6 +-
> mm/hugetlb.c | 150 ++++++++++++++++++++++++++++++++++++----
> mm/rmap.c | 8 ++-
> 4 files changed, 178 insertions(+), 22 deletions(-)
>
<snip>
>
> /* Forward declaration */
> static int hugetlb_acct_memory(struct hstate *h, long delta);
> +static bool vma_pmd_shareable(struct vm_area_struct *vma);
>
> static inline bool subpool_is_free(struct hugepage_subpool *spool)
> {
> @@ -904,6 +905,89 @@ resv_map_set_hugetlb_cgroup_uncharge_info(struct resv_map *resv_map,
> #endif
> }
>
> +static bool __vma_shareable_flags_pmd(struct vm_area_struct *vma)
> +{
> + return vma->vm_flags & (VM_MAYSHARE | VM_SHARED) &&
Should me make __vma_aligned_range_pmd_shareable check (VM_MAYSHARE | VM_SHARED) like above
instead of VM_MAYSHARE to make code more consistent?
> + vma->vm_private_data;
> +}
> +
> +void hugetlb_vma_lock_read(struct vm_area_struct *vma)
> +{
> + if (__vma_shareable_flags_pmd(vma))
> + down_read((struct rw_semaphore *)vma->vm_private_data);
> +}
> +
> +void hugetlb_vma_unlock_read(struct vm_area_struct *vma)
> +{
> + if (__vma_shareable_flags_pmd(vma))
> + up_read((struct rw_semaphore *)vma->vm_private_data);
> +}
> +
> +void hugetlb_vma_lock_write(struct vm_area_struct *vma)
> +{
> + if (__vma_shareable_flags_pmd(vma))
> + down_write((struct rw_semaphore *)vma->vm_private_data);
> +}
> +
> +void hugetlb_vma_unlock_write(struct vm_area_struct *vma)
> +{
> + if (__vma_shareable_flags_pmd(vma))
> + up_write((struct rw_semaphore *)vma->vm_private_data);
> +}
> +
> +int hugetlb_vma_trylock_write(struct vm_area_struct *vma)
> +{
> + if (!__vma_shareable_flags_pmd(vma))
> + return 1;
> +
> + return down_write_trylock((struct rw_semaphore *)vma->vm_private_data);
> +}
> +
> +void hugetlb_vma_assert_locked(struct vm_area_struct *vma)
> +{
> + if (__vma_shareable_flags_pmd(vma))
> + lockdep_assert_held((struct rw_semaphore *)
> + vma->vm_private_data);
> +}
> +
> +static void hugetlb_free_vma_lock(struct vm_area_struct *vma)
> +{
> + /* Only present in sharable vmas */
> + if (!vma || !(vma->vm_flags & (VM_MAYSHARE | VM_SHARED)))
> + return;
> +
> + if (vma->vm_private_data) {
> + kfree(vma->vm_private_data);
> + vma->vm_private_data = NULL;
> + }
> +}
> +
> +static void hugetlb_alloc_vma_lock(struct vm_area_struct *vma)
> +{
> + struct rw_semaphore *vma_sema;
> +
> + /* Only establish in (flags) sharable vmas */
> + if (!vma || !(vma->vm_flags & (VM_MAYSHARE | VM_SHARED)))
> + return;
> +> + if (!vma_pmd_shareable(vma)) {
> + vma->vm_private_data = NULL;
> + return;
> + }
> +
> + vma_sema = kmalloc(sizeof(*vma_sema), GFP_KERNEL);
> + if (!vma_sema) {
> + /*
> + * If we can not allocate semaphore, then vma can not
> + * participate in pmd sharing.
> + */
> + vma->vm_private_data = NULL;
> + } else {
> + init_rwsem(vma_sema);
> + vma->vm_private_data = vma_sema;
> + }
This code is really subtle. If it's called from hugetlb_vm_op_open during fork after
hugetlb_dup_vma_private is done, there should already be a kmalloc-ed vma_sema for this
vma (because hugetlb_alloc_vma_lock is also called by hugetlb_dup_vma_private). So we
can't simply change the value of vm_private_data here or vma_sema will be leaked ? But
when hugetlb_alloc_vma_lock is called from hugetlb_reserve_pages, it should work fine.
Or am I miss something?
Thanks.
Powered by blists - more mailing lists