Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1644212

[PATCH 4.4 28/56] vfio/type1: Remove locked page accounting workqueue

From Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Newsgroups linux.kernel
Subject [PATCH 4.4 28/56] vfio/type1: Remove locked page accounting workqueue
Date 2017-05-18 13:10 +0200
Message-ID <tIrtH-1zL-99@gated-at.bofh.it> (permalink)
References <tIrtD-1zL-3@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw

4.4-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Alex Williamson <alex.williamson@redhat.com>

commit 0cfef2b7410b64d7a430947e0b533314c4f97153 upstream.

If the mmap_sem is contented then the vfio type1 IOMMU backend will
defer locked page accounting updates to a workqueue task.  This has a
few problems and depending on which side the user tries to play, they
might be over-penalized for unmaps that haven't yet been accounted or
race the workqueue to enter more mappings than they're allowed.  The
original intent of this workqueue mechanism seems to be focused on
reducing latency through the ioctl, but we cannot do so at the cost
of correctness.  Remove this workqueue mechanism and update the
callers to allow for failure.  We can also now recheck the limit under
write lock to make sure we don't exceed it.

vfio_pin_pages_remote() also now necessarily includes an unwind path
which we can jump to directly if the consecutive page pinning finds
that we're exceeding the user's memory limits.  This avoids the
current lazy approach which does accounting and mapping up to the
fault, only to return an error on the next iteration to unwind the
entire vfio_dma.

Cc: stable@vger.kernel.org
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Kirti Wankhede <kwankhede@nvidia.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>


---
 drivers/vfio/vfio_iommu_type1.c |  100 ++++++++++++++++------------------------
 1 file changed, 42 insertions(+), 58 deletions(-)

--- a/drivers/vfio/vfio_iommu_type1.c
+++ b/drivers/vfio/vfio_iommu_type1.c
@@ -130,57 +130,34 @@ static void vfio_unlink_dma(struct vfio_
 	rb_erase(&old->node, &iommu->dma_list);
 }
 
-struct vwork {
-	struct mm_struct	*mm;
-	long			npage;
-	struct work_struct	work;
-};
-
-/* delayed decrement/increment for locked_vm */
-static void vfio_lock_acct_bg(struct work_struct *work)
+static int vfio_lock_acct(long npage, bool *lock_cap)
 {
-	struct vwork *vwork = container_of(work, struct vwork, work);
-	struct mm_struct *mm;
+	int ret = 0;
 
-	mm = vwork->mm;
-	down_write(&mm->mmap_sem);
-	mm->locked_vm += vwork->npage;
-	up_write(&mm->mmap_sem);
-	mmput(mm);
-	kfree(vwork);
-}
+	if (!npage)
+		return 0;
 
-static void vfio_lock_acct(long npage)
-{
-	struct vwork *vwork;
-	struct mm_struct *mm;
+	if (!current->mm)
+		return -ESRCH; /* process exited */
 
-	if (!current->mm || !npage)
-		return; /* process exited or nothing to do */
+	down_write(&current->mm->mmap_sem);
+	if (npage > 0) {
+		if (lock_cap ? !*lock_cap : !capable(CAP_IPC_LOCK)) {
+			unsigned long limit;
 
-	if (down_write_trylock(&current->mm->mmap_sem)) {
-		current->mm->locked_vm += npage;
-		up_write(&current->mm->mmap_sem);
-		return;
-	}
+			limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
 
-	/*
-	 * Couldn't get mmap_sem lock, so must setup to update
-	 * mm->locked_vm later. If locked_vm were atomic, we
-	 * wouldn't need this silliness
-	 */
-	vwork = kmalloc(sizeof(struct vwork), GFP_KERNEL);
-	if (!vwork)
-		return;
-	mm = get_task_mm(current);
-	if (!mm) {
-		kfree(vwork);
-		return;
+			if (current->mm->locked_vm + npage > limit)
+				ret = -ENOMEM;
+		}
 	}
-	INIT_WORK(&vwork->work, vfio_lock_acct_bg);
-	vwork->mm = mm;
-	vwork->npage = npage;
-	schedule_work(&vwork->work);
+
+	if (!ret)
+		current->mm->locked_vm += npage;
+
+	up_write(&current->mm->mmap_sem);
+
+	return ret;
 }
 
 /*
@@ -262,9 +239,9 @@ static int vaddr_get_pfn(unsigned long v
 static long vfio_pin_pages(unsigned long vaddr, long npage,
 			   int prot, unsigned long *pfn_base)
 {
-	unsigned long limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
+	unsigned long pfn = 0, limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
 	bool lock_cap = capable(CAP_IPC_LOCK);
-	long ret, i;
+	long ret, i = 1;
 	bool rsvd;
 
 	if (!current->mm)
@@ -283,16 +260,11 @@ static long vfio_pin_pages(unsigned long
 		return -ENOMEM;
 	}
 
-	if (unlikely(disable_hugepages)) {
-		if (!rsvd)
-			vfio_lock_acct(1);
-		return 1;
-	}
+	if (unlikely(disable_hugepages))
+		goto out;
 
 	/* Lock all the consecutive pages from pfn_base */
-	for (i = 1, vaddr += PAGE_SIZE; i < npage; i++, vaddr += PAGE_SIZE) {
-		unsigned long pfn = 0;
-
+	for (vaddr += PAGE_SIZE; i < npage; i++, vaddr += PAGE_SIZE) {
 		ret = vaddr_get_pfn(vaddr, prot, &pfn);
 		if (ret)
 			break;
@@ -308,12 +280,24 @@ static long vfio_pin_pages(unsigned long
 			put_pfn(pfn, prot);
 			pr_warn("%s: RLIMIT_MEMLOCK (%ld) exceeded\n",
 				__func__, limit << PAGE_SHIFT);
-			break;
+			ret = -ENOMEM;
+			goto unpin_out;
 		}
 	}
 
+out:
 	if (!rsvd)
-		vfio_lock_acct(i);
+		ret = vfio_lock_acct(i, &lock_cap);
+
+unpin_out:
+	if (ret) {
+		if (!rsvd) {
+			for (pfn = *pfn_base ; i ; pfn++, i--)
+				put_pfn(pfn, prot);
+		}
+
+		return ret;
+	}
 
 	return i;
 }
@@ -328,7 +312,7 @@ static long vfio_unpin_pages(unsigned lo
 		unlocked += put_pfn(pfn++, prot);
 
 	if (do_accounting)
-		vfio_lock_acct(-unlocked);
+		vfio_lock_acct(-unlocked, NULL);
 
 	return unlocked;
 }
@@ -390,7 +374,7 @@ static void vfio_unmap_unpin(struct vfio
 		cond_resched();
 	}
 
-	vfio_lock_acct(-unlocked);
+	vfio_lock_acct(-unlocked, NULL);
 }
 
 static void vfio_remove_dma(struct vfio_iommu *iommu, struct vfio_dma *dma)

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 4.4 00/56] 4.4.69-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 54/56] wlcore: Pass win_size taken from ieee80211_sta to FW Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 34/56] fs/xattr.c: zero out memory copied to userspace in getxattr Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 52/56] mac80211: pass block ack session timeout to to driver Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 42/56] padata: free correct variable Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 56/56] ipmi: Fix kernel panic at ipmi_ssif_thread() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 17/56] usb: hub: Do not attempt to autosuspend disconnected devices Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 24/56] KVM: arm/arm64: fix races in kvm_psci_vcpu_on Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 09/56] staging: vt6656: use off stack for in buffer USB transfers. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 33/56] ext4: evict inline data when writing to memory map Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 51/56] mac80211: pass RX aggregation window size to driver Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 38/56] SMB3: Work around mount failure when using SMB3 dialect to Macs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 55/56] wlcore: Add RX_BA_WIN_SIZE_CHANGE_EVENT event Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 45/56] serial: omap: fix runtime-pm handling on unbind Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 03/56] target/fileio: Fix zero-length READ and WRITE handling Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 50/56] Bluetooth: hci_intel: add missing tty-device sanity check Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 46/56] serial: omap: suspend device on probe errors Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 28/56] vfio/type1: Remove locked page accounting workqueue Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 44/56] serial: samsung: Use right device for DMA-mapping calls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 36/56] fs/block_dev: always invalidate cleancache in invalidate_bdev() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 08/56] USB: Proper handling of Race Condition when two USB class drivers try to call init_usb_class simultaneously Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 05/56] iscsi-target: Set session_fall_back_to_erl0 when forcing reinstatement Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 12/56] staging: comedi: jr3_pci: fix possible null pointer dereference Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:10 +0200
  [PATCH 4.4 20/56] selftests/x86/ldt_gdt_32: Work around a glibc sigaction() bug Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 04/56] target: Convert ACL change queue_depth se_session reference usage Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 27/56] dm era: save spacemap metadata root after the pre-commit Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 22/56] um: Fix PTRACE_POKEUSER on x86_64 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 10/56] staging: vt6656: use off stack for out buffer USB transfers. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 18/56] usb: misc: legousbtower: Fix buffers on stack Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 14/56] usb: misc: add missing continue in switch Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 21/56] x86, pmem: Fix cache flushing for iovec write < 8 bytes Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 16/56] usb: hub: Fix error loop seen after hub communication errors Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 15/56] usb: Make sure usb/phy/of gets built-in Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 11/56] staging: gdm724x: gdm_mux: fix use-after-free on module unload Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 29/56] IB/core: Fix sysfs registration error flow Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 19/56] x86/boot: Fix BSS corruption/overwrite bug in early x86 kernel startup Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  [PATCH 4.4 25/56] block: fix blk_integrity_register to use templates interval_exp if not 0 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-05-18 13:20 +0200
  Re: [PATCH 4.4 00/56] 4.4.69-stable review Shuah Khan <shuahkh@osg.samsung.com> - 2017-05-18 19:40 +0200
  Re: [PATCH 4.4 00/56] 4.4.69-stable review Guenter Roeck <linux@roeck-us.net> - 2017-05-19 03:20 +0200

csiph-web