thr3ads.net - Linux Virtualization - [PATCH v2 0/9] proc/vmcore: sanitize access to virtio-mem memory [Oct 2021]

If this information is useful, please help other people find it:
Share via:

David Hildenbrand

2021-Oct-05 12:14 UTC

[PATCH v2 0/9] proc/vmcore: sanitize access to virtio-mem memory

As so often with virtio-mem changes that mess with common MM
infrastructure, this might be a good candiate to go via Andrew's tree.

--

After removing /dev/kmem, sanitizing /proc/kcore and handling /dev/mem,
this series tackles the last sane way how a VM could accidentially access
logically unplugged memory managed by a virtio-mem device: /proc/vmcore

When dumping memory via "makedumpfile", PG_offline pages, used by
virtio-mem to flag logically unplugged memory, are already properly
excluded; however, especially when accessing/copying /proc/vmcore "the
usual way", we can still end up reading logically unplugged memory part of
a virtio-mem device.

Patch #1-#3 are cleanups. Patch #4 extends the existing oldmem_pfn_is_ram
mechanism. Patch #5-#7 are virtio-mem refactorings for patch #8, which
implements the virtio-mem logic to query the state of device blocks.

Patch #8:

"
Although virtio-mem currently supports reading unplugged memory in the
hypervisor, this will change in the future, indicated to the device via
a new feature flag. We similarly sanitized /proc/kcore access recently.
[...]
Distributions that support virtio-mem+kdump have to make sure that the
virtio_mem module will be part of the kdump kernel or the kdump initrd;
dracut was recently [2] extended to include virtio-mem in the generated
initrd. As long as no special kdump kernels are used, this will
automatically make sure that virtio-mem will be around in the kdump initrd
and sanitize /proc/vmcore access -- with dracut.
"

This is the last remaining bit to support
VIRTIO_MEM_F_UNPLUGGED_INACCESSIBLE [3] in the Linux implementation of
virtio-mem.

Note: this is best-effort. We'll never be able to control what runs inside
the second kernel, really, but we also don't have to care: we only care
about sane setups where we don't want our VM getting zapped once we
touch the wrong memory location while dumping. While we usually expect sane
setups to use "makedumfile", nothing really speaks against just
copying
/proc/vmcore, especially in environments where HWpoisioning isn't typically
expected. Also, we really don't want to put all our trust completely on the
memmap, so sanitizing also makes sense when just using "makedumpfile".

[1] https://lkml.kernel.org/r/20210526093041.8800-1-david at redhat.com
[2] https://github.com/dracutdevs/dracut/pull/1157
[3] https://lists.oasis-open.org/archives/virtio-comment/202109/msg00021.html

v1 -> v2:
- "x86/xen: simplify xen_oldmem_pfn_is_ram()"
-- Simplify even more
- "x86/xen: print a warning when HVMOP_get_mem_type fails"
-- Added
- "virtio-mem: kdump mode to sanitize /proc/vmcore access"
-- Fix wrong range check

Cc: Andrew Morton <akpm at linux-foundation.org>
Cc: Thomas Gleixner <tglx at linutronix.de>
Cc: Ingo Molnar <mingo at redhat.com>
Cc: Borislav Petkov <bp at alien8.de>
Cc: "H. Peter Anvin" <hpa at zytor.com>
Cc: Boris Ostrovsky <boris.ostrovsky at oracle.com>
Cc: Juergen Gross <jgross at suse.com>
Cc: Stefano Stabellini <sstabellini at kernel.org>
Cc: "Michael S. Tsirkin" <mst at redhat.com>
Cc: Jason Wang <jasowang at redhat.com>
Cc: Dave Young <dyoung at redhat.com>
Cc: Baoquan He <bhe at redhat.com>
Cc: Vivek Goyal <vgoyal at redhat.com>
Cc: Michal Hocko <mhocko at suse.com>
Cc: Oscar Salvador <osalvador at suse.de>
Cc: Mike Rapoport <rppt at kernel.org>
Cc: "Rafael J. Wysocki" <rafael.j.wysocki at intel.com>
Cc: x86 at kernel.org
Cc: xen-devel at lists.xenproject.org
Cc: virtualization at lists.linux-foundation.org
Cc: kexec at lists.infradead.org
Cc: linux-fsdevel at vger.kernel.org
Cc: linux-mm at kvack.org

David Hildenbrand (9):
  x86/xen: update xen_oldmem_pfn_is_ram() documentation
  x86/xen: simplify xen_oldmem_pfn_is_ram()
  x86/xen: print a warning when HVMOP_get_mem_type fails
  proc/vmcore: let pfn_is_ram() return a bool
  proc/vmcore: convert oldmem_pfn_is_ram callback to more generic vmcore
    callbacks
  virtio-mem: factor out hotplug specifics from virtio_mem_init() into
    virtio_mem_init_hotplug()
  virtio-mem: factor out hotplug specifics from virtio_mem_probe() into
    virtio_mem_init_hotplug()
  virtio-mem: factor out hotplug specifics from virtio_mem_remove() into
    virtio_mem_deinit_hotplug()
  virtio-mem: kdump mode to sanitize /proc/vmcore access

 arch/x86/kernel/aperture_64.c |  13 +-
 arch/x86/xen/mmu_hvm.c        |  37 ++---
 drivers/virtio/virtio_mem.c   | 297 ++++++++++++++++++++++++----------
 fs/proc/vmcore.c              | 105 ++++++++----
 include/linux/crash_dump.h    |  26 ++-
 5 files changed, 333 insertions(+), 145 deletions(-)


base-commit: 9e1ff307c779ce1f0f810c7ecce3d95bbae40896
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 1/9] x86/xen: update xen_oldmem_pfn_is_ram() documentation

The callback is only used for the vmcore nowadays.

Reviewed-by: Boris Ostrovsky <boris.ostrvsky at oracle.com>
Signed-off-by: David Hildenbrand <david at redhat.com>
---
 arch/x86/xen/mmu_hvm.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/arch/x86/xen/mmu_hvm.c b/arch/x86/xen/mmu_hvm.c
index 57409373750f..b242d1f4b426 100644
--- a/arch/x86/xen/mmu_hvm.c
+++ b/arch/x86/xen/mmu_hvm.c
@@ -9,12 +9,9 @@
 
 #ifdef CONFIG_PROC_VMCORE
 /*
- * This function is used in two contexts:
- * - the kdump kernel has to check whether a pfn of the crashed kernel
- *   was a ballooned page. vmcore is using this function to decide
- *   whether to access a pfn of the crashed kernel.
- * - the kexec kernel has to check whether a pfn was ballooned by the
- *   previous kernel. If the pfn is ballooned, handle it properly.
+ * The kdump kernel has to check whether a pfn of the crashed kernel
+ * was a ballooned page. vmcore is using this function to decide
+ * whether to access a pfn of the crashed kernel.
  * Returns 0 if the pfn is not backed by a RAM page, the caller may
  * handle the pfn special in this case.
  */
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 2/9] x86/xen: simplify xen_oldmem_pfn_is_ram()

Let's simplify return handling.

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 arch/x86/xen/mmu_hvm.c | 15 +--------------
 1 file changed, 1 insertion(+), 14 deletions(-)

diff --git a/arch/x86/xen/mmu_hvm.c b/arch/x86/xen/mmu_hvm.c
index b242d1f4b426..d1b38c77352b 100644
--- a/arch/x86/xen/mmu_hvm.c
+++ b/arch/x86/xen/mmu_hvm.c
@@ -21,23 +21,10 @@ static int xen_oldmem_pfn_is_ram(unsigned long pfn)
 		.domid = DOMID_SELF,
 		.pfn = pfn,
 	};
-	int ram;
 
 	if (HYPERVISOR_hvm_op(HVMOP_get_mem_type, &a))
 		return -ENXIO;
-
-	switch (a.mem_type) {
-	case HVMMEM_mmio_dm:
-		ram = 0;
-		break;
-	case HVMMEM_ram_rw:
-	case HVMMEM_ram_ro:
-	default:
-		ram = 1;
-		break;
-	}
-
-	return ram;
+	return a.mem_type != HVMMEM_mmio_dm;
 }
 #endif
 
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 3/9] x86/xen: print a warning when HVMOP_get_mem_type fails

HVMOP_get_mem_type is not expected to fail, "This call failing is
indication of something going quite wrong and it would be good to know
about this." [1]

Let's add a pr_warn_once().

[1] https://lkml.kernel.org/r/3b935aa0-6d85-0bcd-100e-15098add3c4c at oracle.com

Suggested-by: Boris Ostrovsky <boris.ostrovsky at oracle.com>
Signed-off-by: David Hildenbrand <david at redhat.com>
---
 arch/x86/xen/mmu_hvm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/x86/xen/mmu_hvm.c b/arch/x86/xen/mmu_hvm.c
index d1b38c77352b..6ba8826dcdcc 100644
--- a/arch/x86/xen/mmu_hvm.c
+++ b/arch/x86/xen/mmu_hvm.c
@@ -22,8 +22,10 @@ static int xen_oldmem_pfn_is_ram(unsigned long pfn)
 		.pfn = pfn,
 	};
 
-	if (HYPERVISOR_hvm_op(HVMOP_get_mem_type, &a))
+	if (HYPERVISOR_hvm_op(HVMOP_get_mem_type, &a)) {
+		pr_warn_once("Unexpected HVMOP_get_mem_type failure\n");
 		return -ENXIO;
+	}
 	return a.mem_type != HVMMEM_mmio_dm;
 }
 #endif
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 4/9] proc/vmcore: let pfn_is_ram() return a bool

The callback should deal with errors internally, it doesn't make sense to
expose these via pfn_is_ram(). We'll rework the callbacks next. Right now
we consider errors as if "it's RAM"; no functional change.

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 fs/proc/vmcore.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 9a15334da208..a9bd80ab670e 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -84,11 +84,11 @@ void unregister_oldmem_pfn_is_ram(void)
 }
 EXPORT_SYMBOL_GPL(unregister_oldmem_pfn_is_ram);
 
-static int pfn_is_ram(unsigned long pfn)
+static bool pfn_is_ram(unsigned long pfn)
 {
 	int (*fn)(unsigned long pfn);
 	/* pfn is ram unless fn() checks pagetype */
-	int ret = 1;
+	bool ret = true;
 
 	/*
 	 * Ask hypervisor if the pfn is really ram.
@@ -97,7 +97,7 @@ static int pfn_is_ram(unsigned long pfn)
 	 */
 	fn = oldmem_pfn_is_ram;
 	if (fn)
-		ret = fn(pfn);
+		ret = !!fn(pfn);
 
 	return ret;
 }
@@ -124,7 +124,7 @@ ssize_t read_from_oldmem(char *buf, size_t count,
 			nr_bytes = count;
 
 		/* If pfn is not ram, return zeros for sparse dump files */
-		if (pfn_is_ram(pfn) == 0)
+		if (!pfn_is_ram(pfn))
 			memset(buf, 0, nr_bytes);
 		else {
 			if (encrypted)
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 5/9] proc/vmcore: convert oldmem_pfn_is_ram callback to more generic vmcore callbacks

Let's support multiple registered callbacks, making sure that
registering vmcore callbacks cannot fail. Make the callback return a
bool instead of an int, handling how to deal with errors internally.
Drop unused HAVE_OLDMEM_PFN_IS_RAM.

We soon want to make use of this infrastructure from other drivers:
virtio-mem, registering one callback for each virtio-mem device, to
prevent reading unplugged virtio-mem memory.

Handle it via a generic vmcore_cb structure, prepared for future
extensions: for example, once we support virtio-mem on s390x where the
vmcore is completely constructed in the second kernel, we want to detect
and add plugged virtio-mem memory ranges to the vmcore in order for them
to get dumped properly.

Handle corner cases that are unexpected and shouldn't happen in sane
setups: registering a callback after the vmcore has already been opened
(warn only) and unregistering a callback after the vmcore has already been
opened (warn and essentially read only zeroes from that point on).

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 arch/x86/kernel/aperture_64.c | 13 ++++-
 arch/x86/xen/mmu_hvm.c        | 11 ++--
 fs/proc/vmcore.c              | 99 ++++++++++++++++++++++++-----------
 include/linux/crash_dump.h    | 26 +++++++--
 4 files changed, 111 insertions(+), 38 deletions(-)

diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
index 10562885f5fc..af3ba08b684b 100644
--- a/arch/x86/kernel/aperture_64.c
+++ b/arch/x86/kernel/aperture_64.c
@@ -73,12 +73,23 @@ static int gart_mem_pfn_is_ram(unsigned long pfn)
 		      (pfn >= aperture_pfn_start + aperture_page_count));
 }
 
+#ifdef CONFIG_PROC_VMCORE
+static bool gart_oldmem_pfn_is_ram(struct vmcore_cb *cb, unsigned long pfn)
+{
+	return !!gart_mem_pfn_is_ram(pfn);
+}
+
+static struct vmcore_cb gart_vmcore_cb = {
+	.pfn_is_ram = gart_oldmem_pfn_is_ram,
+};
+#endif
+
 static void __init exclude_from_core(u64 aper_base, u32 aper_order)
 {
 	aperture_pfn_start = aper_base >> PAGE_SHIFT;
 	aperture_page_count = (32 * 1024 * 1024) << aper_order >>
PAGE_SHIFT;
 #ifdef CONFIG_PROC_VMCORE
-	WARN_ON(register_oldmem_pfn_is_ram(&gart_mem_pfn_is_ram));
+	register_vmcore_cb(&gart_vmcore_cb);
 #endif
 #ifdef CONFIG_PROC_KCORE
 	WARN_ON(register_mem_pfn_is_ram(&gart_mem_pfn_is_ram));
diff --git a/arch/x86/xen/mmu_hvm.c b/arch/x86/xen/mmu_hvm.c
index 6ba8826dcdcc..509bdee3ab90 100644
--- a/arch/x86/xen/mmu_hvm.c
+++ b/arch/x86/xen/mmu_hvm.c
@@ -12,10 +12,10 @@
  * The kdump kernel has to check whether a pfn of the crashed kernel
  * was a ballooned page. vmcore is using this function to decide
  * whether to access a pfn of the crashed kernel.
- * Returns 0 if the pfn is not backed by a RAM page, the caller may
+ * Returns "false" if the pfn is not backed by a RAM page, the caller
may
  * handle the pfn special in this case.
  */
-static int xen_oldmem_pfn_is_ram(unsigned long pfn)
+static bool xen_vmcore_pfn_is_ram(struct vmcore_cb *cb, unsigned long pfn)
 {
 	struct xen_hvm_get_mem_type a = {
 		.domid = DOMID_SELF,
@@ -24,10 +24,13 @@ static int xen_oldmem_pfn_is_ram(unsigned long pfn)
 
 	if (HYPERVISOR_hvm_op(HVMOP_get_mem_type, &a)) {
 		pr_warn_once("Unexpected HVMOP_get_mem_type failure\n");
-		return -ENXIO;
+		return true;
 	}
 	return a.mem_type != HVMMEM_mmio_dm;
 }
+static struct vmcore_cb xen_vmcore_cb = {
+	.pfn_is_ram = xen_vmcore_pfn_is_ram,
+};
 #endif
 
 static void xen_hvm_exit_mmap(struct mm_struct *mm)
@@ -61,6 +64,6 @@ void __init xen_hvm_init_mmu_ops(void)
 	if (is_pagetable_dying_supported())
 		pv_ops.mmu.exit_mmap = xen_hvm_exit_mmap;
 #ifdef CONFIG_PROC_VMCORE
-	WARN_ON(register_oldmem_pfn_is_ram(&xen_oldmem_pfn_is_ram));
+	register_vmcore_cb(&xen_vmcore_cb);
 #endif
 }
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index a9bd80ab670e..7a04b2eca287 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -62,46 +62,75 @@ core_param(novmcoredd, vmcoredd_disabled, bool, 0);
 /* Device Dump Size */
 static size_t vmcoredd_orig_sz;
 
-/*
- * Returns > 0 for RAM pages, 0 for non-RAM pages, < 0 on error
- * The called function has to take care of module refcounting.
- */
-static int (*oldmem_pfn_is_ram)(unsigned long pfn);
-
-int register_oldmem_pfn_is_ram(int (*fn)(unsigned long pfn))
+static DECLARE_RWSEM(vmcore_cb_rwsem);
+/* List of registered vmcore callbacks. */
+static LIST_HEAD(vmcore_cb_list);
+/* Whether we had a surprise unregistration of a callback. */
+static bool vmcore_cb_unstable;
+/* Whether the vmcore has been opened once. */
+static bool vmcore_opened;
+
+void register_vmcore_cb(struct vmcore_cb *cb)
 {
-	if (oldmem_pfn_is_ram)
-		return -EBUSY;
-	oldmem_pfn_is_ram = fn;
-	return 0;
+	down_write(&vmcore_cb_rwsem);
+	INIT_LIST_HEAD(&cb->next);
+	list_add_tail(&cb->next, &vmcore_cb_list);
+	/*
+	 * Registering a vmcore callback after the vmcore was opened is
+	 * very unusual (e.g., manual driver loading).
+	 */
+	if (vmcore_opened)
+		pr_warn_once("Unexpected vmcore callback registration\n");
+	up_write(&vmcore_cb_rwsem);
 }
-EXPORT_SYMBOL_GPL(register_oldmem_pfn_is_ram);
+EXPORT_SYMBOL_GPL(register_vmcore_cb);
 
-void unregister_oldmem_pfn_is_ram(void)
+void unregister_vmcore_cb(struct vmcore_cb *cb)
 {
-	oldmem_pfn_is_ram = NULL;
-	wmb();
+	down_write(&vmcore_cb_rwsem);
+	list_del(&cb->next);
+	/*
+	 * Unregistering a vmcore callback after the vmcore was opened is
+	 * very unusual (e.g., forced driver removal), but we cannot stop
+	 * unregistering.
+	 */
+	if (vmcore_opened) {
+		pr_warn_once("Unexpected vmcore callback unregistration\n");
+		vmcore_cb_unstable = true;
+	}
+	up_write(&vmcore_cb_rwsem);
 }
-EXPORT_SYMBOL_GPL(unregister_oldmem_pfn_is_ram);
+EXPORT_SYMBOL_GPL(unregister_vmcore_cb);
 
 static bool pfn_is_ram(unsigned long pfn)
 {
-	int (*fn)(unsigned long pfn);
-	/* pfn is ram unless fn() checks pagetype */
+	struct vmcore_cb *cb;
 	bool ret = true;
 
-	/*
-	 * Ask hypervisor if the pfn is really ram.
-	 * A ballooned page contains no data and reading from such a page
-	 * will cause high load in the hypervisor.
-	 */
-	fn = oldmem_pfn_is_ram;
-	if (fn)
-		ret = !!fn(pfn);
+	lockdep_assert_held_read(&vmcore_cb_rwsem);
+	if (unlikely(vmcore_cb_unstable))
+		return false;
+
+	list_for_each_entry(cb, &vmcore_cb_list, next) {
+		if (unlikely(!cb->pfn_is_ram))
+			continue;
+		ret = cb->pfn_is_ram(cb, pfn);
+		if (!ret)
+			break;
+	}
 
 	return ret;
 }
 
+static int open_vmcore(struct inode *inode, struct file *file)
+{
+	down_read(&vmcore_cb_rwsem);
+	vmcore_opened = true;
+	up_read(&vmcore_cb_rwsem);
+
+	return 0;
+}
+
 /* Reads a page from the oldmem device from given offset. */
 ssize_t read_from_oldmem(char *buf, size_t count,
 			 u64 *ppos, int userbuf,
@@ -117,6 +146,7 @@ ssize_t read_from_oldmem(char *buf, size_t count,
 	offset = (unsigned long)(*ppos % PAGE_SIZE);
 	pfn = (unsigned long)(*ppos / PAGE_SIZE);
 
+	down_read(&vmcore_cb_rwsem);
 	do {
 		if (count > (PAGE_SIZE - offset))
 			nr_bytes = PAGE_SIZE - offset;
@@ -136,8 +166,10 @@ ssize_t read_from_oldmem(char *buf, size_t count,
 				tmp = copy_oldmem_page(pfn, buf, nr_bytes,
 						       offset, userbuf);
 
-			if (tmp < 0)
+			if (tmp < 0) {
+				up_read(&vmcore_cb_rwsem);
 				return tmp;
+			}
 		}
 		*ppos += nr_bytes;
 		count -= nr_bytes;
@@ -147,6 +179,7 @@ ssize_t read_from_oldmem(char *buf, size_t count,
 		offset = 0;
 	} while (count);
 
+	up_read(&vmcore_cb_rwsem);
 	return read;
 }
 
@@ -537,14 +570,19 @@ static int vmcore_remap_oldmem_pfn(struct vm_area_struct
*vma,
 			    unsigned long from, unsigned long pfn,
 			    unsigned long size, pgprot_t prot)
 {
+	int ret;
+
 	/*
 	 * Check if oldmem_pfn_is_ram was registered to avoid
 	 * looping over all pages without a reason.
 	 */
-	if (oldmem_pfn_is_ram)
-		return remap_oldmem_pfn_checked(vma, from, pfn, size, prot);
+	down_read(&vmcore_cb_rwsem);
+	if (!list_empty(&vmcore_cb_list) || vmcore_cb_unstable)
+		ret = remap_oldmem_pfn_checked(vma, from, pfn, size, prot);
 	else
-		return remap_oldmem_pfn_range(vma, from, pfn, size, prot);
+		ret = remap_oldmem_pfn_range(vma, from, pfn, size, prot);
+	up_read(&vmcore_cb_rwsem);
+	return ret;
 }
 
 static int mmap_vmcore(struct file *file, struct vm_area_struct *vma)
@@ -668,6 +706,7 @@ static int mmap_vmcore(struct file *file, struct
vm_area_struct *vma)
 #endif
 
 static const struct proc_ops vmcore_proc_ops = {
+	.proc_open	= open_vmcore,
 	.proc_read	= read_vmcore,
 	.proc_lseek	= default_llseek,
 	.proc_mmap	= mmap_vmcore,
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index 2618577a4d6d..0c547d866f1e 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -91,9 +91,29 @@ static inline void vmcore_unusable(void)
 		elfcorehdr_addr = ELFCORE_ADDR_ERR;
 }
 
-#define HAVE_OLDMEM_PFN_IS_RAM 1
-extern int register_oldmem_pfn_is_ram(int (*fn)(unsigned long pfn));
-extern void unregister_oldmem_pfn_is_ram(void);
+/**
+ * struct vmcore_cb - driver callbacks for /proc/vmcore handling
+ * @pfn_is_ram: check whether a PFN really is RAM and should be accessed when
+ *              reading the vmcore. Will return "true" if it is RAM
or if the
+ *              callback cannot tell. If any callback returns
"false", it's not
+ *              RAM and the page must not be accessed; zeroes should be
+ *              indicated in the vmcore instead. For example, a ballooned page
+ *              contains no data and reading from such a page will cause high
+ *              load in the hypervisor.
+ * @next: List head to manage registered callbacks internally; initialized by
+ *        register_vmcore_cb().
+ *
+ * vmcore callbacks allow drivers managing physical memory ranges to
+ * coordinate with vmcore handling code, for example, to prevent accessing
+ * physical memory ranges that should not be accessed when reading the vmcore,
+ * although included in the vmcore header as memory ranges to dump.
+ */
+struct vmcore_cb {
+	bool (*pfn_is_ram)(struct vmcore_cb *cb, unsigned long pfn);
+	struct list_head next;
+};
+extern void register_vmcore_cb(struct vmcore_cb *cb);
+extern void unregister_vmcore_cb(struct vmcore_cb *cb);
 
 #else /* !CONFIG_CRASH_DUMP */
 static inline bool is_kdump_kernel(void) { return 0; }
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 6/9] virtio-mem: factor out hotplug specifics from virtio_mem_init() into virtio_mem_init_hotplug()

Let's prepare for a new virtio-mem kdump mode in which we don't actually
hot(un)plug any memory but only observe the state of device blocks.

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 drivers/virtio/virtio_mem.c | 81 ++++++++++++++++++++-----------------
 1 file changed, 44 insertions(+), 37 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index bef8ad6bf466..2ba7e8d6ba8d 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -2392,41 +2392,10 @@ static int virtio_mem_init_vq(struct virtio_mem *vm)
 	return 0;
 }
 
-static int virtio_mem_init(struct virtio_mem *vm)
+static int virtio_mem_init_hotplug(struct virtio_mem *vm)
 {
 	const struct range pluggable_range = mhp_get_pluggable_range(true);
 	uint64_t sb_size, addr;
-	uint16_t node_id;
-
-	if (!vm->vdev->config->get) {
-		dev_err(&vm->vdev->dev, "config access disabled\n");
-		return -EINVAL;
-	}
-
-	/*
-	 * We don't want to (un)plug or reuse any memory when in kdump. The
-	 * memory is still accessible (but not mapped).
-	 */
-	if (is_kdump_kernel()) {
-		dev_warn(&vm->vdev->dev, "disabled in kdump kernel\n");
-		return -EBUSY;
-	}
-
-	/* Fetch all properties that can't change. */
-	virtio_cread_le(vm->vdev, struct virtio_mem_config, plugged_size,
-			&vm->plugged_size);
-	virtio_cread_le(vm->vdev, struct virtio_mem_config, block_size,
-			&vm->device_block_size);
-	virtio_cread_le(vm->vdev, struct virtio_mem_config, node_id,
-			&node_id);
-	vm->nid = virtio_mem_translate_node_id(vm, node_id);
-	virtio_cread_le(vm->vdev, struct virtio_mem_config, addr,
&vm->addr);
-	virtio_cread_le(vm->vdev, struct virtio_mem_config, region_size,
-			&vm->region_size);
-
-	/* Determine the nid for the device based on the lowest address. */
-	if (vm->nid == NUMA_NO_NODE)
-		vm->nid = memory_add_physaddr_to_nid(vm->addr);
 
 	/* bad device setup - warn only */
 	if (!IS_ALIGNED(vm->addr, memory_block_size_bytes()))
@@ -2496,10 +2465,6 @@ static int virtio_mem_init(struct virtio_mem *vm)
 					      vm->offline_threshold);
 	}
 
-	dev_info(&vm->vdev->dev, "start address: 0x%llx",
vm->addr);
-	dev_info(&vm->vdev->dev, "region size: 0x%llx",
vm->region_size);
-	dev_info(&vm->vdev->dev, "device block size: 0x%llx",
-		 (unsigned long long)vm->device_block_size);
 	dev_info(&vm->vdev->dev, "memory block size: 0x%lx",
 		 memory_block_size_bytes());
 	if (vm->in_sbm)
@@ -2508,10 +2473,52 @@ static int virtio_mem_init(struct virtio_mem *vm)
 	else
 		dev_info(&vm->vdev->dev, "big block size: 0x%llx",
 			 (unsigned long long)vm->bbm.bb_size);
+
+	return 0;
+}
+
+static int virtio_mem_init(struct virtio_mem *vm)
+{
+	uint16_t node_id;
+
+	if (!vm->vdev->config->get) {
+		dev_err(&vm->vdev->dev, "config access disabled\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * We don't want to (un)plug or reuse any memory when in kdump. The
+	 * memory is still accessible (but not mapped).
+	 */
+	if (is_kdump_kernel()) {
+		dev_warn(&vm->vdev->dev, "disabled in kdump kernel\n");
+		return -EBUSY;
+	}
+
+	/* Fetch all properties that can't change. */
+	virtio_cread_le(vm->vdev, struct virtio_mem_config, plugged_size,
+			&vm->plugged_size);
+	virtio_cread_le(vm->vdev, struct virtio_mem_config, block_size,
+			&vm->device_block_size);
+	virtio_cread_le(vm->vdev, struct virtio_mem_config, node_id,
+			&node_id);
+	vm->nid = virtio_mem_translate_node_id(vm, node_id);
+	virtio_cread_le(vm->vdev, struct virtio_mem_config, addr,
&vm->addr);
+	virtio_cread_le(vm->vdev, struct virtio_mem_config, region_size,
+			&vm->region_size);
+
+	/* Determine the nid for the device based on the lowest address. */
+	if (vm->nid == NUMA_NO_NODE)
+		vm->nid = memory_add_physaddr_to_nid(vm->addr);
+
+	dev_info(&vm->vdev->dev, "start address: 0x%llx",
vm->addr);
+	dev_info(&vm->vdev->dev, "region size: 0x%llx",
vm->region_size);
+	dev_info(&vm->vdev->dev, "device block size: 0x%llx",
+		 (unsigned long long)vm->device_block_size);
 	if (vm->nid != NUMA_NO_NODE && IS_ENABLED(CONFIG_NUMA))
 		dev_info(&vm->vdev->dev, "nid: %d", vm->nid);
 
-	return 0;
+	return virtio_mem_init_hotplug(vm);
 }
 
 static int virtio_mem_create_resource(struct virtio_mem *vm)
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 7/9] virtio-mem: factor out hotplug specifics from virtio_mem_probe() into virtio_mem_init_hotplug()

Let's prepare for a new virtio-mem kdump mode in which we don't actually
hot(un)plug any memory but only observe the state of device blocks.

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 drivers/virtio/virtio_mem.c | 87 +++++++++++++++++++------------------
 1 file changed, 45 insertions(+), 42 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 2ba7e8d6ba8d..1be3ee7f684d 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -260,6 +260,8 @@ static void virtio_mem_fake_offline_going_offline(unsigned
long pfn,
 static void virtio_mem_fake_offline_cancel_offline(unsigned long pfn,
 						   unsigned long nr_pages);
 static void virtio_mem_retry(struct virtio_mem *vm);
+static int virtio_mem_create_resource(struct virtio_mem *vm);
+static void virtio_mem_delete_resource(struct virtio_mem *vm);
 
 /*
  * Register a virtio-mem device so it will be considered for the online_page
@@ -2395,7 +2397,8 @@ static int virtio_mem_init_vq(struct virtio_mem *vm)
 static int virtio_mem_init_hotplug(struct virtio_mem *vm)
 {
 	const struct range pluggable_range = mhp_get_pluggable_range(true);
-	uint64_t sb_size, addr;
+	uint64_t unit_pages, sb_size, addr;
+	int rc;
 
 	/* bad device setup - warn only */
 	if (!IS_ALIGNED(vm->addr, memory_block_size_bytes()))
@@ -2474,7 +2477,48 @@ static int virtio_mem_init_hotplug(struct virtio_mem *vm)
 		dev_info(&vm->vdev->dev, "big block size: 0x%llx",
 			 (unsigned long long)vm->bbm.bb_size);
 
+	/* create the parent resource for all memory */
+	rc = virtio_mem_create_resource(vm);
+	if (rc)
+		return rc;
+
+	/* use a single dynamic memory group to cover the whole memory device */
+	if (vm->in_sbm)
+		unit_pages = PHYS_PFN(memory_block_size_bytes());
+	else
+		unit_pages = PHYS_PFN(vm->bbm.bb_size);
+	rc = memory_group_register_dynamic(vm->nid, unit_pages);
+	if (rc < 0)
+		goto out_del_resource;
+	vm->mgid = rc;
+
+	/*
+	 * If we still have memory plugged, we have to unplug all memory first.
+	 * Registering our parent resource makes sure that this memory isn't
+	 * actually in use (e.g., trying to reload the driver).
+	 */
+	if (vm->plugged_size) {
+		vm->unplug_all_required = true;
+		dev_info(&vm->vdev->dev, "unplugging all memory is
required\n");
+	}
+
+	/* register callbacks */
+	vm->memory_notifier.notifier_call = virtio_mem_memory_notifier_cb;
+	rc = register_memory_notifier(&vm->memory_notifier);
+	if (rc)
+		goto out_unreg_group;
+	rc = register_virtio_mem_device(vm);
+	if (rc)
+		goto out_unreg_mem;
+
 	return 0;
+out_unreg_mem:
+	unregister_memory_notifier(&vm->memory_notifier);
+out_unreg_group:
+	memory_group_unregister(vm->mgid);
+out_del_resource:
+	virtio_mem_delete_resource(vm);
+	return rc;
 }
 
 static int virtio_mem_init(struct virtio_mem *vm)
@@ -2578,7 +2622,6 @@ static bool virtio_mem_has_memory_added(struct virtio_mem
*vm)
 static int virtio_mem_probe(struct virtio_device *vdev)
 {
 	struct virtio_mem *vm;
-	uint64_t unit_pages;
 	int rc;
 
 	BUILD_BUG_ON(sizeof(struct virtio_mem_req) != 24);
@@ -2608,40 +2651,6 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 	if (rc)
 		goto out_del_vq;
 
-	/* create the parent resource for all memory */
-	rc = virtio_mem_create_resource(vm);
-	if (rc)
-		goto out_del_vq;
-
-	/* use a single dynamic memory group to cover the whole memory device */
-	if (vm->in_sbm)
-		unit_pages = PHYS_PFN(memory_block_size_bytes());
-	else
-		unit_pages = PHYS_PFN(vm->bbm.bb_size);
-	rc = memory_group_register_dynamic(vm->nid, unit_pages);
-	if (rc < 0)
-		goto out_del_resource;
-	vm->mgid = rc;
-
-	/*
-	 * If we still have memory plugged, we have to unplug all memory first.
-	 * Registering our parent resource makes sure that this memory isn't
-	 * actually in use (e.g., trying to reload the driver).
-	 */
-	if (vm->plugged_size) {
-		vm->unplug_all_required = true;
-		dev_info(&vm->vdev->dev, "unplugging all memory is
required\n");
-	}
-
-	/* register callbacks */
-	vm->memory_notifier.notifier_call = virtio_mem_memory_notifier_cb;
-	rc = register_memory_notifier(&vm->memory_notifier);
-	if (rc)
-		goto out_unreg_group;
-	rc = register_virtio_mem_device(vm);
-	if (rc)
-		goto out_unreg_mem;
-
 	virtio_device_ready(vdev);
 
 	/* trigger a config update to start processing the requested_size */
@@ -2649,12 +2658,6 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 	queue_work(system_freezable_wq, &vm->wq);
 
 	return 0;
-out_unreg_mem:
-	unregister_memory_notifier(&vm->memory_notifier);
-out_unreg_group:
-	memory_group_unregister(vm->mgid);
-out_del_resource:
-	virtio_mem_delete_resource(vm);
 out_del_vq:
 	vdev->config->del_vqs(vdev);
 out_free_vm:
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 8/9] virtio-mem: factor out hotplug specifics from virtio_mem_remove() into virtio_mem_deinit_hotplug()

Let's prepare for a new virtio-mem kdump mode in which we don't actually
hot(un)plug any memory but only observe the state of device blocks.

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 drivers/virtio/virtio_mem.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 1be3ee7f684d..76d8aef3cfd2 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -2667,9 +2667,8 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 	return rc;
 }
 
-static void virtio_mem_remove(struct virtio_device *vdev)
+static void virtio_mem_deinit_hotplug(struct virtio_mem *vm)
 {
-	struct virtio_mem *vm = vdev->priv;
 	unsigned long mb_id;
 	int rc;
 
@@ -2716,7 +2715,8 @@ static void virtio_mem_remove(struct virtio_device *vdev)
 	 * away. Warn at least.
 	 */
 	if (virtio_mem_has_memory_added(vm)) {
-		dev_warn(&vdev->dev, "device still has system memory
added\n");
+		dev_warn(&vm->vdev->dev,
+			 "device still has system memory added\n");
 	} else {
 		virtio_mem_delete_resource(vm);
 		kfree_const(vm->resource_name);
@@ -2730,6 +2730,13 @@ static void virtio_mem_remove(struct virtio_device *vdev)
 	} else {
 		vfree(vm->bbm.bb_states);
 	}
+}
+
+static void virtio_mem_remove(struct virtio_device *vdev)
+{
+	struct virtio_mem *vm = vdev->priv;
+
+	virtio_mem_deinit_hotplug(vm);
 
 	/* reset the device and cleanup the queues */
 	vdev->config->reset(vdev);
-- 
2.31.1

David Hildenbrand

2021-Oct-05 12:14 UTC

head link

[PATCH v2 9/9] virtio-mem: kdump mode to sanitize /proc/vmcore access

Although virtio-mem currently supports reading unplugged memory in the
hypervisor, this will change in the future, indicated to the device via
a new feature flag. We similarly sanitized /proc/kcore access recently. [1]

Let's register a vmcore callback, to allow vmcore code to check if a PFN
belonging to a virtio-mem device is either currently plugged and should
be dumped or is currently unplugged and should not be accessed, instead
mapping the shared zeropage or returning zeroes when reading.

This is important when not capturing /proc/vmcore via tools like
"makedumpfile" that can identify logically unplugged virtio-mem memory
via
PG_offline in the memmap, but simply by e.g., copying the file.

Distributions that support virtio-mem+kdump have to make sure that the
virtio_mem module will be part of the kdump kernel or the kdump initrd;
dracut was recently [2] extended to include virtio-mem in the generated
initrd. As long as no special kdump kernels are used, this will
automatically make sure that virtio-mem will be around in the kdump initrd
and sanitize /proc/vmcore access -- with dracut.

With this series, we'll send one virtio-mem state request for every
~2 MiB chunk of virtio-mem memory indicated in the vmcore that we intend
to read/map.

In the future, we might want to allow building virtio-mem for kdump
mode only, even without CONFIG_MEMORY_HOTPLUG and friends: this way,
we could support special stripped-down kdump kernels that have many
other config options disabled; we'll tackle that once required. Further,
we might want to try sensing bigger blocks (e.g., memory sections)
first before falling back to device blocks on demand.

Tested with Fedora rawhide, which contains a recent kexec-tools version
(considering "System RAM (virtio_mem)" when creating the vmcore
header) and
a recent dracut version (including the virtio_mem module in the kdump
initrd).

[1] https://lkml.kernel.org/r/20210526093041.8800-1-david at redhat.com
[2] https://github.com/dracutdevs/dracut/pull/1157

Signed-off-by: David Hildenbrand <david at redhat.com>
---
 drivers/virtio/virtio_mem.c | 136 ++++++++++++++++++++++++++++++++----
 1 file changed, 124 insertions(+), 12 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 76d8aef3cfd2..3573b78f6b05 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -223,6 +223,9 @@ struct virtio_mem {
 	 * When this lock is held the pointers can't change, ONLINE and
 	 * OFFLINE blocks can't change the state and no subblocks will get
 	 * plugged/unplugged.
+	 *
+	 * In kdump mode, used to serialize requests, last_block_addr and
+	 * last_block_plugged.
 	 */
 	struct mutex hotplug_mutex;
 	bool hotplug_active;
@@ -230,6 +233,9 @@ struct virtio_mem {
 	/* An error occurred we cannot handle - stop processing requests. */
 	bool broken;
 
+	/* Cached valued of is_kdump_kernel() when the device was probed. */
+	bool in_kdump;
+
 	/* The driver is being removed. */
 	spinlock_t removal_lock;
 	bool removing;
@@ -243,6 +249,13 @@ struct virtio_mem {
 	/* Memory notifier (online/offline events). */
 	struct notifier_block memory_notifier;
 
+#ifdef CONFIG_PROC_VMCORE
+	/* vmcore callback for /proc/vmcore handling in kdump mode */
+	struct vmcore_cb vmcore_cb;
+	uint64_t last_block_addr;
+	bool last_block_plugged;
+#endif /* CONFIG_PROC_VMCORE */
+
 	/* Next device in the list of virtio-mem devices. */
 	struct list_head next;
 };
@@ -2293,6 +2306,12 @@ static void virtio_mem_run_wq(struct work_struct *work)
 	uint64_t diff;
 	int rc;
 
+	if (unlikely(vm->in_kdump)) {
+		dev_warn_once(&vm->vdev->dev,
+			     "unexpected workqueue run in kdump kernel\n");
+		return;
+	}
+
 	hrtimer_cancel(&vm->retry_timer);
 
 	if (vm->broken)
@@ -2521,6 +2540,86 @@ static int virtio_mem_init_hotplug(struct virtio_mem *vm)
 	return rc;
 }
 
+#ifdef CONFIG_PROC_VMCORE
+static int virtio_mem_send_state_request(struct virtio_mem *vm, uint64_t addr,
+					 uint64_t size)
+{
+	const uint64_t nb_vm_blocks = size / vm->device_block_size;
+	const struct virtio_mem_req req = {
+		.type = cpu_to_virtio16(vm->vdev, VIRTIO_MEM_REQ_STATE),
+		.u.state.addr = cpu_to_virtio64(vm->vdev, addr),
+		.u.state.nb_blocks = cpu_to_virtio16(vm->vdev, nb_vm_blocks),
+	};
+	int rc = -ENOMEM;
+
+	dev_dbg(&vm->vdev->dev, "requesting state: 0x%llx -
0x%llx\n", addr,
+		addr + size - 1);
+
+	switch (virtio_mem_send_request(vm, &req)) {
+	case VIRTIO_MEM_RESP_ACK:
+		return virtio16_to_cpu(vm->vdev, vm->resp.u.state.state);
+	case VIRTIO_MEM_RESP_ERROR:
+		rc = -EINVAL;
+		break;
+	default:
+		break;
+	}
+
+	dev_dbg(&vm->vdev->dev, "requesting state failed: %d\n",
rc);
+	return rc;
+}
+
+static bool virtio_mem_vmcore_pfn_is_ram(struct vmcore_cb *cb,
+					 unsigned long pfn)
+{
+	struct virtio_mem *vm = container_of(cb, struct virtio_mem,
+					     vmcore_cb);
+	uint64_t addr = PFN_PHYS(pfn);
+	bool is_ram;
+	int rc;
+
+	if (!virtio_mem_contains_range(vm, addr, PAGE_SIZE))
+		return true;
+	if (!vm->plugged_size)
+		return false;
+
+	/*
+	 * We have to serialize device requests and access to the information
+	 * about the block queried last.
+	 */
+	mutex_lock(&vm->hotplug_mutex);
+
+	addr = ALIGN_DOWN(addr, vm->device_block_size);
+	if (addr != vm->last_block_addr) {
+		rc = virtio_mem_send_state_request(vm, addr,
+						   vm->device_block_size);
+		/* On any kind of error, we're going to signal !ram. */
+		if (rc == VIRTIO_MEM_STATE_PLUGGED)
+			vm->last_block_plugged = true;
+		else
+			vm->last_block_plugged = false;
+		vm->last_block_addr = addr;
+	}
+
+	is_ram = vm->last_block_plugged;
+	mutex_unlock(&vm->hotplug_mutex);
+	return is_ram;
+}
+#endif /* CONFIG_PROC_VMCORE */
+
+static int virtio_mem_init_kdump(struct virtio_mem *vm)
+{
+#ifdef CONFIG_PROC_VMCORE
+	dev_info(&vm->vdev->dev, "memory hot(un)plug disabled in kdump
kernel\n");
+	vm->vmcore_cb.pfn_is_ram = virtio_mem_vmcore_pfn_is_ram;
+	register_vmcore_cb(&vm->vmcore_cb);
+	return 0;
+#else /* CONFIG_PROC_VMCORE */
+	dev_warn(&vm->vdev->dev, "disabled in kdump kernel without
vmcore\n");
+	return -EBUSY;
+#endif /* CONFIG_PROC_VMCORE */
+}
+
 static int virtio_mem_init(struct virtio_mem *vm)
 {
 	uint16_t node_id;
@@ -2530,15 +2629,6 @@ static int virtio_mem_init(struct virtio_mem *vm)
 		return -EINVAL;
 	}
 
-	/*
-	 * We don't want to (un)plug or reuse any memory when in kdump. The
-	 * memory is still accessible (but not mapped).
-	 */
-	if (is_kdump_kernel()) {
-		dev_warn(&vm->vdev->dev, "disabled in kdump kernel\n");
-		return -EBUSY;
-	}
-
 	/* Fetch all properties that can't change. */
 	virtio_cread_le(vm->vdev, struct virtio_mem_config, plugged_size,
 			&vm->plugged_size);
@@ -2562,6 +2652,12 @@ static int virtio_mem_init(struct virtio_mem *vm)
 	if (vm->nid != NUMA_NO_NODE && IS_ENABLED(CONFIG_NUMA))
 		dev_info(&vm->vdev->dev, "nid: %d", vm->nid);
 
+	/*
+	 * We don't want to (un)plug or reuse any memory when in kdump. The
+	 * memory is still accessible (but not exposed to Linux).
+	 */
+	if (vm->in_kdump)
+		return virtio_mem_init_kdump(vm);
 	return virtio_mem_init_hotplug(vm);
 }
 
@@ -2640,6 +2736,7 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 	hrtimer_init(&vm->retry_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
 	vm->retry_timer.function = virtio_mem_timer_expired;
 	vm->retry_timer_ms = VIRTIO_MEM_RETRY_TIMER_MIN_MS;
+	vm->in_kdump = is_kdump_kernel();
 
 	/* register the virtqueue */
 	rc = virtio_mem_init_vq(vm);
@@ -2654,8 +2751,10 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 	virtio_device_ready(vdev);
 
 	/* trigger a config update to start processing the requested_size */
-	atomic_set(&vm->config_changed, 1);
-	queue_work(system_freezable_wq, &vm->wq);
+	if (!vm->in_kdump) {
+		atomic_set(&vm->config_changed, 1);
+		queue_work(system_freezable_wq, &vm->wq);
+	}
 
 	return 0;
 out_del_vq:
@@ -2732,11 +2831,21 @@ static void virtio_mem_deinit_hotplug(struct virtio_mem
*vm)
 	}
 }
 
+static void virtio_mem_deinit_kdump(struct virtio_mem *vm)
+{
+#ifdef CONFIG_PROC_VMCORE
+	unregister_vmcore_cb(&vm->vmcore_cb);
+#endif /* CONFIG_PROC_VMCORE */
+}
+
 static void virtio_mem_remove(struct virtio_device *vdev)
 {
 	struct virtio_mem *vm = vdev->priv;
 
-	virtio_mem_deinit_hotplug(vm);
+	if (vm->in_kdump)
+		virtio_mem_deinit_kdump(vm);
+	else
+		virtio_mem_deinit_hotplug(vm);
 
 	/* reset the device and cleanup the queues */
 	vdev->config->reset(vdev);
@@ -2750,6 +2859,9 @@ static void virtio_mem_config_changed(struct virtio_device
*vdev)
 {
 	struct virtio_mem *vm = vdev->priv;
 
+	if (unlikely(vm->in_kdump))
+		return;
+
 	atomic_set(&vm->config_changed, 1);
 	virtio_mem_retry(vm);
 }
-- 
2.31.1

Linux Virtualization - Oct 2021 - [PATCH v2 0/9] proc/vmcore: sanitize access to virtio-mem memory

[PATCH v2 0/9] proc/vmcore: sanitize access to virtio-mem memory

[PATCH v2 1/9] x86/xen: update xen_oldmem_pfn_is_ram() documentation

[PATCH v2 2/9] x86/xen: simplify xen_oldmem_pfn_is_ram()

[PATCH v2 3/9] x86/xen: print a warning when HVMOP_get_mem_type fails

[PATCH v2 4/9] proc/vmcore: let pfn_is_ram() return a bool

[PATCH v2 5/9] proc/vmcore: convert oldmem_pfn_is_ram callback to more generic vmcore callbacks

[PATCH v2 6/9] virtio-mem: factor out hotplug specifics from virtio_mem_init() into virtio_mem_init_hotplug()

[PATCH v2 7/9] virtio-mem: factor out hotplug specifics from virtio_mem_probe() into virtio_mem_init_hotplug()

[PATCH v2 8/9] virtio-mem: factor out hotplug specifics from virtio_mem_remove() into virtio_mem_deinit_hotplug()

[PATCH v2 9/9] virtio-mem: kdump mode to sanitize /proc/vmcore access