Commit 89b15332 authored by Johannes Weiner's avatar Johannes Weiner Committed by Linus Torvalds
Browse files

mm: drop mmap_sem before calling balance_dirty_pages() in write fault 

One of our services is observing hanging ps/top/etc under heavy write
IO, and the task states show this is an mmap_sem priority inversion:

A write fault is holding the mmap_sem in read-mode and waiting for
(heavily cgroup-limited) IO in balance_dirty_pages():

    balance_dirty_pages+0x724/0x905
    balance_dirty_pages_ratelimited+0x254/0x390
    fault_dirty_shared_page.isra.96+0x4a/0x90
    do_wp_page+0x33e/0x400
    __handle_mm_fault+0x6f0/0xfa0
    handle_mm_fault+0xe4/0x200
    __do_page_fault+0x22b/0x4a0
    page_fault+0x45/0x50

Somebody tries to change the address space, contending for the mmap_sem in
write-mode:

    call_rwsem_down_write_failed_killable+0x13/0x20
    do_mprotect_pkey+0xa8/0x330
    SyS_mprotect+0xf/0x20
    do_syscall_64+0x5b/0x100
    entry_SYSCALL_64_after_hwframe+0x3d/0xa2

The waiting writer locks out all subsequent readers to avoid lock
starvation, and several threads can be seen hanging like this:

    call_rwsem_down_read_failed+0x14/0x30
    proc_pid_cmdline_read+0xa0/0x480
    __vfs_read+0x23/0x140
    vfs_read+0x87/0x130
    SyS_read+0x42/0x90
    do_syscall_64+0x5b/0x100
    entry_SYSCALL_64_after_hwframe+0x3d/0xa2

To fix this, do what we do for cache read faults already: drop the
mmap_sem before calling into anything IO bound, in this case the
balance_dirty_pages() function, and return VM_FAULT_RETRY.

Link: http://lkml.kernel.org/r/20190924194238.GA29030@cmpxchg.org


Signed-off-by: default avatarJohannes Weiner <hannes@cmpxchg.org>
Reviewed-by: default avatarMatthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: default avatarKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Josef Bacik <josef@toxicpanda.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent 1603c8d1

mm/filemap.c

+0 −21
Original line number Diff line number Diff line
@@ -2329,27 +2329,6 @@ EXPORT_SYMBOL(generic_file_read_iter); 


#ifdef CONFIG_MMU

#define MMAP_LOTSAMISS  (100)

static struct file *maybe_unlock_mmap_for_io(struct vm_fault *vmf,

					     struct file *fpin)

{

	int flags = vmf->flags;



	if (fpin)

		return fpin;



	/*

	 * FAULT_FLAG_RETRY_NOWAIT means we don't want to wait on page locks or

	 * anything, so we only pin the file and drop the mmap_sem if only

	 * FAULT_FLAG_ALLOW_RETRY is set.

	 */

	if ((flags & (FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_RETRY_NOWAIT)) ==

	    FAULT_FLAG_ALLOW_RETRY) {

		fpin = get_file(vmf->vma->vm_file);

		up_read(&vmf->vma->vm_mm->mmap_sem);

	}

	return fpin;

}



/*

 * lock_page_maybe_drop_mmap - lock the page, possibly dropping the mmap_sem

 * @vmf - the vm_fault for this fault.

mm/internal.h

+21 −0
Original line number Diff line number Diff line
@@ -362,6 +362,27 @@ vma_address(struct page *page, struct vm_area_struct *vma) 
	return max(start, vma->vm_start);

}



static inline struct file *maybe_unlock_mmap_for_io(struct vm_fault *vmf,

						    struct file *fpin)

{

	int flags = vmf->flags;



	if (fpin)

		return fpin;



	/*

	 * FAULT_FLAG_RETRY_NOWAIT means we don't want to wait on page locks or

	 * anything, so we only pin the file and drop the mmap_sem if only

	 * FAULT_FLAG_ALLOW_RETRY is set.

	 */

	if ((flags & (FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_RETRY_NOWAIT)) ==

	    FAULT_FLAG_ALLOW_RETRY) {

		fpin = get_file(vmf->vma->vm_file);

		up_read(&vmf->vma->vm_mm->mmap_sem);

	}

	return fpin;

}



#else /* !CONFIG_MMU */

static inline void clear_page_mlock(struct page *page) { }

static inline void mlock_vma_page(struct page *page) { }

mm/memory.c

+27 −11
Original line number Diff line number Diff line
@@ -2289,10 +2289,11 @@ static vm_fault_t do_page_mkwrite(struct vm_fault *vmf) 
 *

 * The function expects the page to be locked and unlocks it.

 */

static void fault_dirty_shared_page(struct vm_area_struct *vma,

				    struct page *page)

static vm_fault_t fault_dirty_shared_page(struct vm_fault *vmf)

{

	struct vm_area_struct *vma = vmf->vma;

	struct address_space *mapping;

	struct page *page = vmf->page;

	bool dirtied;

	bool page_mkwrite = vma->vm_ops && vma->vm_ops->page_mkwrite;
@@ -2307,16 +2308,30 @@ static void fault_dirty_shared_page(struct vm_area_struct *vma, 
	mapping = page_rmapping(page);

	unlock_page(page);



	if ((dirtied || page_mkwrite) && mapping) {

	if (!page_mkwrite)

		file_update_time(vma->vm_file);



	/*

		 * Some device drivers do not set page.mapping

		 * but still dirty their pages

	 * Throttle page dirtying rate down to writeback speed.

	 *

	 * mapping may be NULL here because some device drivers do not

	 * set page.mapping but still dirty their pages

	 *

	 * Drop the mmap_sem before waiting on IO, if we can. The file

	 * is pinning the mapping, as per above.

	 */

	if ((dirtied || page_mkwrite) && mapping) {

		struct file *fpin;



		fpin = maybe_unlock_mmap_for_io(vmf, NULL);

		balance_dirty_pages_ratelimited(mapping);

		if (fpin) {

			fput(fpin);

			return VM_FAULT_RETRY;

		}

	}



	if (!page_mkwrite)

		file_update_time(vma->vm_file);

	return 0;

}



/*
@@ -2571,6 +2586,7 @@ static vm_fault_t wp_page_shared(struct vm_fault *vmf) 
	__releases(vmf->ptl)

{

	struct vm_area_struct *vma = vmf->vma;

	vm_fault_t ret = VM_FAULT_WRITE;



	get_page(vmf->page);
@@ -2594,10 +2610,10 @@ static vm_fault_t wp_page_shared(struct vm_fault *vmf) 
		wp_page_reuse(vmf);

		lock_page(vmf->page);

	}

	fault_dirty_shared_page(vma, vmf->page);

	ret |= fault_dirty_shared_page(vmf);

	put_page(vmf->page);



	return VM_FAULT_WRITE;

	return ret;

}



/*
@@ -3641,7 +3657,7 @@ static vm_fault_t do_shared_fault(struct vm_fault *vmf) 
		return ret;

	}



	fault_dirty_shared_page(vma, vmf->page);

	ret |= fault_dirty_shared_page(vmf);

	return ret;

}

CRA Git | Maintained and supported by SUSTech CRA and CCSE