Skip to contentSkip to conversation

cxl/memdev: Avoid mailbox functionality on device memory CXL devices

PATCH In review Ira Weiny 4 participants 5 messages Jul 28, 2023 – Aug 1, 2023 CXL

lore
Share & tools
lore

Ira Weiny

author
#0 lore raw

PATCH 1/1 cxl/memdev: Avoid mailbox functionality on device memory CXL devices

Using the proposed type-2 cxl-test device[1] the following splat was observed:

  BUG: kernel NULL pointer dereference, address: 0000000000000278
  [...]
  RIP: 0010:devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
  [...]
  Call Trace:
   <TASK>
   ? __die+0x1f/0x70
   ? page_fault_oops+0x149/0x420
   ? fixup_exception+0x22/0x310
   ? kernelmode_fixup_or_oops+0x84/0x110
   ? exc_page_fault+0x6d/0x150
   ? asm_exc_page_fault+0x22/0x30
   ? devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
   cxl_mock_mem_probe+0x632/0x870 [cxl_mock_mem]
   platform_probe+0x40/0x90
   really_probe+0x19e/0x3e0
   ? __pfx___driver_attach+0x10/0x10
   __driver_probe_device+0x78/0x160
   driver_probe_device+0x1f/0x90
   __driver_attach+0xce/0x1c0
   bus_for_each_dev+0x63/0xa0
   bus_add_driver+0x112/0x210
   driver_register+0x55/0x100
   ? __pfx_cxl_mock_mem_driver_init+0x10/0x10 [cxl_mock_mem]
   [...]

Commit f6b8ab32e3ec made the mailbox functionality optional. However, some mailbox functionality was merged after that patch. Therefore some mailbox functionality can be accessed on a device which did not set up the mailbox.

While no devices currently exist, commit f6b8ab32e3ec is incomplete. Complete the checks for memdev state to bring the code to a consistent state for when type-2 devices are introduced.

[1] https://lore.kernel.org/all/168592160379.1948938.12863272903570476312.stgit@dwillia2-xfh.jf.intel.com/

Fixes f6b8ab32e3ec ("cxl/memdev: Make mailbox functionality optional")
Cc Dan Williams <[email protected]>
---
 drivers/cxl/core/mbox.c   |  9 +++++++++
 drivers/cxl/core/memdev.c | 26 ++++++++++++++++++++++++++
 drivers/cxl/mem.c         | 18 ++++++++++--------
 drivers/cxl/pci.c         |  5 ++++-
 drivers/cxl/pmem.c        |  3 +++
 5 files changed, 52 insertions(+), 9 deletions(-)
drivers/cxl/core/mbox.c +9 -0
index d6d067fbee97..eb1758fb8cdf 100644
@@ -482,6 +482,9 @@ int cxl_query_cmd(struct cxl_memdev *cxlmd,
 
 	dev_dbg(dev, "Query IOCTL\n");
 
+	if (!mds)
+		return -EIO;
+
 	if (get_user(n_commands, &q->n_commands))
 		return -EFAULT;
 
@@ -586,6 +589,9 @@ int cxl_send_cmd(struct cxl_memdev *cxlmd, struct cxl_send_command __user *s)
 
 	dev_dbg(dev, "Send IOCTL\n");
 
+	if (!mds)
+		return -EIO;
+
 	if (copy_from_user(&send, s, sizeof(send)))
 		return -EFAULT;
 
@@ -1245,6 +1251,9 @@ int cxl_mem_get_poison(struct cxl_memdev *cxlmd, u64 offset, u64 len,
 	int nr_records = 0;
 	int rc;
 
+	if (!mds)
+		return -EIO;
+
 	rc = mutex_lock_interruptible(&mds->poison.lock);
 	if (rc)
 		return rc;
drivers/cxl/core/memdev.c +26 -0
index f99e7ec3cc40..629e479f751b 100644
@@ -201,6 +201,19 @@ static ssize_t security_erase_store(struct device *dev,
 static struct device_attribute dev_attr_security_erase =
 	__ATTR(erase, 0200, NULL, security_erase_store);
 
+static umode_t cxl_memdev_security_visible(struct kobject *kobj,
+					   struct attribute *a, int n)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
+	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
+
+	if (!mds)
+		return 0;
+
+	return a->mode;
+}
+
 static int cxl_get_poison_by_memdev(struct cxl_memdev *cxlmd)
 {
 	struct cxl_dev_state *cxlds = cxlmd->cxlds;
@@ -332,6 +345,9 @@ int cxl_inject_poison(struct cxl_memdev *cxlmd, u64 dpa)
 	struct cxl_region *cxlr;
 	int rc;
 
+	if (!mds)
+		return -EIO;
+
 	if (!IS_ENABLED(CONFIG_DEBUG_FS))
 		return 0;
 
@@ -380,6 +396,9 @@ int cxl_clear_poison(struct cxl_memdev *cxlmd, u64 dpa)
 	struct cxl_region *cxlr;
 	int rc;
 
+	if (!mds)
+		return -EIO;
+
 	if (!IS_ENABLED(CONFIG_DEBUG_FS))
 		return 0;
 
@@ -480,6 +499,7 @@ static struct attribute_group cxl_memdev_pmem_attribute_group = {
 static struct attribute_group cxl_memdev_security_attribute_group = {
 	.name = "security",
 	.attrs = cxl_memdev_security_attributes,
+	.is_visible = cxl_memdev_security_visible,
 };
 
 static const struct attribute_group *cxl_memdev_attribute_groups[] = {
@@ -542,6 +562,9 @@ static void cxl_memdev_security_shutdown(struct device *dev)
 	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
 	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
 
+	if (!mds)
+		return;
+
 	if (mds->security.poll)
 		cancel_delayed_work_sync(&mds->security.poll_dwork);
 }
@@ -997,6 +1020,9 @@ static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
 	struct device *dev = &cxlmd->dev;
 	struct kernfs_node *sec;
 
+	if (!mds)
+		return 0;
+
 	sec = sysfs_get_dirent(dev->kobj.sd, "security");
 	if (!sec) {
 		dev_err(dev, "sysfs_get_dirent 'security' failed\n");
drivers/cxl/mem.c +10 -8
index 317c7548e4e9..4755a890018d 100644
@@ -132,12 +132,14 @@ static int cxl_mem_probe(struct device *dev)
 	dentry = cxl_debugfs_create_dir(dev_name(dev));
 	debugfs_create_devm_seqfile(dev, "dpamem", dentry, cxl_mem_dpa_show);
 
-	if (test_bit(CXL_POISON_ENABLED_INJECT, mds->poison.enabled_cmds))
-		debugfs_create_file("inject_poison", 0200, dentry, cxlmd,
-				    &cxl_poison_inject_fops);
-	if (test_bit(CXL_POISON_ENABLED_CLEAR, mds->poison.enabled_cmds))
-		debugfs_create_file("clear_poison", 0200, dentry, cxlmd,
-				    &cxl_poison_clear_fops);
+	if (mds) {
+		if (test_bit(CXL_POISON_ENABLED_INJECT, mds->poison.enabled_cmds))
+			debugfs_create_file("inject_poison", 0200, dentry, cxlmd,
+					    &cxl_poison_inject_fops);
+		if (test_bit(CXL_POISON_ENABLED_CLEAR, mds->poison.enabled_cmds))
+			debugfs_create_file("clear_poison", 0200, dentry, cxlmd,
+					    &cxl_poison_clear_fops);
+	}
 
 	rc = devm_add_action_or_reset(dev, remove_debugfs, dentry);
 	if (rc)
@@ -222,8 +224,8 @@ static umode_t cxl_mem_visible(struct kobject *kobj, struct attribute *a, int n)
 		struct cxl_memdev_state *mds =
 			to_cxl_memdev_state(cxlmd->cxlds);
 
-		if (!test_bit(CXL_POISON_ENABLED_LIST,
-			      mds->poison.enabled_cmds))
+		if (!mds || !test_bit(CXL_POISON_ENABLED_LIST,
+				      mds->poison.enabled_cmds))
 			return 0;
 	}
 	return a->mode;
drivers/cxl/pci.c +4 -1
index 1cb1494c28fe..93f6140432cd 100644
@@ -122,7 +122,7 @@ static irqreturn_t cxl_pci_mbox_irq(int irq, void *id)
 	struct cxl_dev_state *cxlds = dev_id->cxlds;
 	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
 
-	if (!cxl_mbox_background_complete(cxlds))
+	if (!mds || !cxl_mbox_background_complete(cxlds))
 		return IRQ_NONE;
 
 	reg = readq(cxlds->regs.mbox + CXLDEV_MBOX_BG_CMD_STATUS_OFFSET);
@@ -624,6 +624,9 @@ static irqreturn_t cxl_event_thread(int irq, void *id)
 	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
 	u32 status;
 
+	if (!mds)
+		return IRQ_HANDLED;
+
 	do {
 		/*
 		 * CXL 3.0 8.2.8.3.1: The lower 32 bits are the status;
drivers/cxl/pmem.c +3 -0
index 7cb8994f8809..f1adfdd1a2b3 100644
@@ -70,6 +70,9 @@ static int cxl_nvdimm_probe(struct device *dev)
 	struct nvdimm *nvdimm;
 	int rc;
 
+	if (WARN_ON_ONCE(!mds))
+		return -EIO;
+
 	set_exclusive_cxl_commands(mds, exclusive_cmds);
 	rc = devm_add_action_or_reset(dev, clear_exclusive, mds);
 	if (rc)

---
base-commit: 20ea1e7d13c1b544fe67c4a8dc3943bb1ab33e6f
change-id: 20230728-cxl-fix-devmemdev-5003ce927f68

Best regards,

Signature · 1 line
Ira Weiny <[email protected]>
maintainer
#1 lore raw
Quoted 38 lines · Ira Weiny

Using the proposed type-2 cxl-test device[1] the following splat was observed:

   BUG: kernel NULL pointer dereference, address: 0000000000000278
   [...]
   RIP: 0010:devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
   [...]
   Call Trace:
    <TASK>
    ? __die+0x1f/0x70
    ? page_fault_oops+0x149/0x420
    ? fixup_exception+0x22/0x310
    ? kernelmode_fixup_or_oops+0x84/0x110
    ? exc_page_fault+0x6d/0x150
    ? asm_exc_page_fault+0x22/0x30
    ? devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
    cxl_mock_mem_probe+0x632/0x870 [cxl_mock_mem]
    platform_probe+0x40/0x90
    really_probe+0x19e/0x3e0
    ? __pfx___driver_attach+0x10/0x10
    __driver_probe_device+0x78/0x160
    driver_probe_device+0x1f/0x90
    __driver_attach+0xce/0x1c0
    bus_for_each_dev+0x63/0xa0
    bus_add_driver+0x112/0x210
    driver_register+0x55/0x100
    ? __pfx_cxl_mock_mem_driver_init+0x10/0x10 [cxl_mock_mem]
    [...]

Commit f6b8ab32e3ec made the mailbox functionality optional. However, some mailbox functionality was merged after that patch. Therefore some mailbox functionality can be accessed on a device which did not set up the mailbox.

While no devices currently exist, commit f6b8ab32e3ec is incomplete. Complete the checks for memdev state to bring the code to a consistent state for when type-2 devices are introduced.

Quoted 5 lines

[1] https://lore.kernel.org/all/168592160379.1948938.12863272903570476312.stgit@dwillia2-xfh.jf.intel.com/

Fixes f6b8ab32e3ec ("cxl/memdev: Make mailbox functionality optional")
Cc Dan Williams <[email protected]>

I think you'll need to coordinate with this patch from Davidlohr?
https://lore.kernel.org/linux-cxl/[email protected]/

Quoted 193 lines
---
  drivers/cxl/core/mbox.c   |  9 +++++++++
  drivers/cxl/core/memdev.c | 26 ++++++++++++++++++++++++++
  drivers/cxl/mem.c         | 18 ++++++++++--------
  drivers/cxl/pci.c         |  5 ++++-
  drivers/cxl/pmem.c        |  3 +++
  5 files changed, 52 insertions(+), 9 deletions(-)
drivers/cxl/core/mbox.c +9 -0
index d6d067fbee97..eb1758fb8cdf 100644
@@ -482,6 +482,9 @@ int cxl_query_cmd(struct cxl_memdev *cxlmd,
 
  	dev_dbg(dev, "Query IOCTL\n");
 
+	if (!mds)
+		return -EIO;
+
  	if (get_user(n_commands, &q->n_commands))
  		return -EFAULT;
 
@@ -586,6 +589,9 @@ int cxl_send_cmd(struct cxl_memdev *cxlmd, struct cxl_send_command __user *s)
 
  	dev_dbg(dev, "Send IOCTL\n");
 
+	if (!mds)
+		return -EIO;
+
  	if (copy_from_user(&send, s, sizeof(send)))
  		return -EFAULT;
 
@@ -1245,6 +1251,9 @@ int cxl_mem_get_poison(struct cxl_memdev *cxlmd, u64 offset, u64 len,
  	int nr_records = 0;
  	int rc;
 
+	if (!mds)
+		return -EIO;
+
  	rc = mutex_lock_interruptible(&mds->poison.lock);
  	if (rc)
  		return rc;
drivers/cxl/core/memdev.c +26 -0
index f99e7ec3cc40..629e479f751b 100644
@@ -201,6 +201,19 @@ static ssize_t security_erase_store(struct device *dev,
  static struct device_attribute dev_attr_security_erase =
  	__ATTR(erase, 0200, NULL, security_erase_store);
 
+static umode_t cxl_memdev_security_visible(struct kobject *kobj,
+					   struct attribute *a, int n)
+{
+	struct device *dev = kobj_to_dev(kobj);
+	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
+	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
+
+	if (!mds)
+		return 0;
+
+	return a->mode;
+}
+
  static int cxl_get_poison_by_memdev(struct cxl_memdev *cxlmd)
  {
  	struct cxl_dev_state *cxlds = cxlmd->cxlds;
@@ -332,6 +345,9 @@ int cxl_inject_poison(struct cxl_memdev *cxlmd, u64 dpa)
  	struct cxl_region *cxlr;
  	int rc;
 
+	if (!mds)
+		return -EIO;
+
  	if (!IS_ENABLED(CONFIG_DEBUG_FS))
  		return 0;
 
@@ -380,6 +396,9 @@ int cxl_clear_poison(struct cxl_memdev *cxlmd, u64 dpa)
  	struct cxl_region *cxlr;
  	int rc;
 
+	if (!mds)
+		return -EIO;
+
  	if (!IS_ENABLED(CONFIG_DEBUG_FS))
  		return 0;
 
@@ -480,6 +499,7 @@ static struct attribute_group cxl_memdev_pmem_attribute_group = {
  static struct attribute_group cxl_memdev_security_attribute_group = {
  	.name = "security",
  	.attrs = cxl_memdev_security_attributes,
+	.is_visible = cxl_memdev_security_visible,
  };
 
  static const struct attribute_group *cxl_memdev_attribute_groups[] = {
@@ -542,6 +562,9 @@ static void cxl_memdev_security_shutdown(struct device *dev)
  	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
  	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
 
+	if (!mds)
+		return;
+
  	if (mds->security.poll)
  		cancel_delayed_work_sync(&mds->security.poll_dwork);
  }
@@ -997,6 +1020,9 @@ static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
  	struct device *dev = &cxlmd->dev;
  	struct kernfs_node *sec;
 
+	if (!mds)
+		return 0;
+
  	sec = sysfs_get_dirent(dev->kobj.sd, "security");
  	if (!sec) {
  		dev_err(dev, "sysfs_get_dirent 'security' failed\n");
drivers/cxl/mem.c +10 -8
index 317c7548e4e9..4755a890018d 100644
@@ -132,12 +132,14 @@ static int cxl_mem_probe(struct device *dev)
  	dentry = cxl_debugfs_create_dir(dev_name(dev));
  	debugfs_create_devm_seqfile(dev, "dpamem", dentry, cxl_mem_dpa_show);
 
-	if (test_bit(CXL_POISON_ENABLED_INJECT, mds->poison.enabled_cmds))
-		debugfs_create_file("inject_poison", 0200, dentry, cxlmd,
-				    &cxl_poison_inject_fops);
-	if (test_bit(CXL_POISON_ENABLED_CLEAR, mds->poison.enabled_cmds))
-		debugfs_create_file("clear_poison", 0200, dentry, cxlmd,
-				    &cxl_poison_clear_fops);
+	if (mds) {
+		if (test_bit(CXL_POISON_ENABLED_INJECT, mds->poison.enabled_cmds))
+			debugfs_create_file("inject_poison", 0200, dentry, cxlmd,
+					    &cxl_poison_inject_fops);
+		if (test_bit(CXL_POISON_ENABLED_CLEAR, mds->poison.enabled_cmds))
+			debugfs_create_file("clear_poison", 0200, dentry, cxlmd,
+					    &cxl_poison_clear_fops);
+	}
 
  	rc = devm_add_action_or_reset(dev, remove_debugfs, dentry);
  	if (rc)
@@ -222,8 +224,8 @@ static umode_t cxl_mem_visible(struct kobject *kobj, struct attribute *a, int n)
  		struct cxl_memdev_state *mds =
  			to_cxl_memdev_state(cxlmd->cxlds);
 
-		if (!test_bit(CXL_POISON_ENABLED_LIST,
-			      mds->poison.enabled_cmds))
+		if (!mds || !test_bit(CXL_POISON_ENABLED_LIST,
+				      mds->poison.enabled_cmds))
  			return 0;
  	}
  	return a->mode;
drivers/cxl/pci.c +4 -1
index 1cb1494c28fe..93f6140432cd 100644
@@ -122,7 +122,7 @@ static irqreturn_t cxl_pci_mbox_irq(int irq, void *id)
  	struct cxl_dev_state *cxlds = dev_id->cxlds;
  	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
 
-	if (!cxl_mbox_background_complete(cxlds))
+	if (!mds || !cxl_mbox_background_complete(cxlds))
  		return IRQ_NONE;
 
  	reg = readq(cxlds->regs.mbox + CXLDEV_MBOX_BG_CMD_STATUS_OFFSET);
@@ -624,6 +624,9 @@ static irqreturn_t cxl_event_thread(int irq, void *id)
  	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
  	u32 status;
 
+	if (!mds)
+		return IRQ_HANDLED;
+
  	do {
  		/*
  		 * CXL 3.0 8.2.8.3.1: The lower 32 bits are the status;
drivers/cxl/pmem.c +3 -1
index 7cb8994f8809..f1adfdd1a2b3 100644
@@ -70,6 +70,9 @@ static int cxl_nvdimm_probe(struct device *dev)
  	struct nvdimm *nvdimm;
  	int rc;
 
+	if (WARN_ON_ONCE(!mds))
+		return -EIO;
+
  	set_exclusive_cxl_commands(mds, exclusive_cmds);
  	rc = devm_add_action_or_reset(dev, clear_exclusive, mds);
  	if (rc)
 
---

base-commit: 20ea1e7d13c1b544fe67c4a8dc3943bb1ab33e6f
change-id: 20230728-cxl-fix-devmemdev-5003ce927f68

Best regards,

#2 lore raw
In reply to Ira Weiny
Quoted 6 lines · Ira Weiny

Using the proposed type-2 cxl-test device[1] the following splat was observed:

  BUG: kernel NULL pointer dereference, address: 0000000000000278
  [...]
  RIP: 0010:devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]

It would be useful to decode this to a line number, the rest of this call trace is not adding much.

Quoted 23 lines

[...]

  Call Trace:
   <TASK>
   ? __die+0x1f/0x70
   ? page_fault_oops+0x149/0x420
   ? fixup_exception+0x22/0x310
   ? kernelmode_fixup_or_oops+0x84/0x110
   ? exc_page_fault+0x6d/0x150
   ? asm_exc_page_fault+0x22/0x30
   ? devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
   cxl_mock_mem_probe+0x632/0x870 [cxl_mock_mem]
   platform_probe+0x40/0x90
   really_probe+0x19e/0x3e0
   ? __pfx___driver_attach+0x10/0x10
   __driver_probe_device+0x78/0x160
   driver_probe_device+0x1f/0x90
   __driver_attach+0xce/0x1c0
   bus_for_each_dev+0x63/0xa0
   bus_add_driver+0x112/0x210
   driver_register+0x55/0x100
   ? __pfx_cxl_mock_mem_driver_init+0x10/0x10 [cxl_mock_mem]
   [...]
Quoted 4 lines

Commit f6b8ab32e3ec made the mailbox functionality optional. However, some mailbox functionality was merged after that patch. Therefore some mailbox functionality can be accessed on a device which did not set up the mailbox.

cxl_memdev_security_init() definitely needs to move out of devm_cxl_add_memdev() and after that I do not think @mds NULL checks need to be sprinkled everywhere. In other words something is wrong at a higher level if we get into some of these helper functions without the memory device state.

So definitely this uncovered a problem where cxl_memdev_security_init() needs to move, but the rest of the mds NULL checks need clear reproduction scenarios and expect most of them are precluded higher in the call stack.

maintainer
#3 lore raw
Quoted 39 lines · Dan Williams
Quoted 6 lines · Ira Weiny

Using the proposed type-2 cxl-test device[1] the following splat was observed:

  BUG: kernel NULL pointer dereference, address: 0000000000000278
  [...]
  RIP: 0010:devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]

It would be useful to decode this to a line number, the rest of this call trace is not adding much.

Quoted 23 lines

[...]

  Call Trace:
   <TASK>
   ? __die+0x1f/0x70
   ? page_fault_oops+0x149/0x420
   ? fixup_exception+0x22/0x310
   ? kernelmode_fixup_or_oops+0x84/0x110
   ? exc_page_fault+0x6d/0x150
   ? asm_exc_page_fault+0x22/0x30
   ? devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
   cxl_mock_mem_probe+0x632/0x870 [cxl_mock_mem]
   platform_probe+0x40/0x90
   really_probe+0x19e/0x3e0
   ? __pfx___driver_attach+0x10/0x10
   __driver_probe_device+0x78/0x160
   driver_probe_device+0x1f/0x90
   __driver_attach+0xce/0x1c0
   bus_for_each_dev+0x63/0xa0
   bus_add_driver+0x112/0x210
   driver_register+0x55/0x100
   ? __pfx_cxl_mock_mem_driver_init+0x10/0x10 [cxl_mock_mem]
   [...]
Quoted 4 lines

Commit f6b8ab32e3ec made the mailbox functionality optional. However, some mailbox functionality was merged after that patch. Therefore some mailbox functionality can be accessed on a device which did not set up the mailbox.

Quoted 5 lines

cxl_memdev_security_init() definitely needs to move out of devm_cxl_add_memdev() and after that I do not think @mds NULL checks need to be sprinkled everywhere. In other words something is wrong at a higher level if we get into some of these helper functions without the memory device state.

Right, so we can move it directly into cxl_pci_probe() - just as with other mbox based functionality. This leaves me wondering, however, what to do about the cxl_memdev_security_shutdown() counterpart. As with the below diff, leaving it as is and just adding a mds nil check might still be considering a layering violation in that it would be asymmetrical wrt to the init; but this is tightly coupled with cxl_memdev_unregister().

Ira does the below fix the crash?

Thanks,
Davidlohr

----8<-------

drivers/cxl/core/memdev.c +7 -10
index 14b547c07f54..4d1bf80c0e54 100644
@@ -561,7 +561,7 @@ static void cxl_memdev_security_shutdown(struct device *dev)
 	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
 	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
 
-	if (mds->security.poll)
+	if (mds && mds->security.poll)
 		cancel_delayed_work_sync(&mds->security.poll_dwork);
 }
 
@@ -1009,11 +1009,11 @@ static void put_sanitize(void *data)
 	sysfs_put(mds->security.sanitize_node);
 }
 
-static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
+int cxl_memdev_security_state_init(struct cxl_memdev_state *mds)
 {
-	struct cxl_dev_state *cxlds = cxlmd->cxlds;
-	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
-	struct device *dev = &cxlmd->dev;
+
+	struct cxl_dev_state *cxlds = &mds->cxlds;
+	struct device *dev = &cxlds->cxlmd->dev;
 	struct kernfs_node *sec;
 
 	sec = sysfs_get_dirent(dev->kobj.sd, "security");
@@ -1029,7 +1029,8 @@ static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
 	}
 
 	return devm_add_action_or_reset(cxlds->dev, put_sanitize, mds);
- }
+}
+EXPORT_SYMBOL_NS_GPL(cxl_memdev_security_state_init, CXL);
 
 struct cxl_memdev *devm_cxl_add_memdev(struct cxl_dev_state *cxlds)
 {
@@ -1059,10 +1060,6 @@ struct cxl_memdev *devm_cxl_add_memdev(struct cxl_dev_state *cxlds)
 	if (rc)
 		goto err;
 
-	rc = cxl_memdev_security_init(cxlmd);
-	if (rc)
-		goto err;
-
 	rc = devm_add_action_or_reset(cxlds->dev, cxl_memdev_unregister, cxlmd);
 	if (rc)
 		return ERR_PTR(rc);
drivers/cxl/cxlmem.h +1 -0
index f86afef90c91..441270770519 100644
@@ -884,6 +884,7 @@ static inline void cxl_mem_active_dec(void)
 #endif
 
 int cxl_mem_sanitize(struct cxl_memdev_state *mds, u16 cmd);
+int cxl_memdev_security_state_init(struct cxl_memdev_state *mds);
 
 struct cxl_hdm {
 	struct cxl_component_regs regs;
drivers/cxl/pci.c +4 -0
index 1cb1494c28fe..5242dbf0044d 100644
@@ -887,6 +887,10 @@ static int cxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (IS_ERR(cxlmd))
 		return PTR_ERR(cxlmd);
 
+	rc = cxl_memdev_security_state_init(mds);
+	if (rc)
+		return rc;
+
 	rc = cxl_memdev_setup_fw_upload(mds);
 	if (rc)
 		return rc;
author
#4 lore raw
Quoted 47 lines · Davidlohr Bueso
Quoted 39 lines · Dan Williams
Quoted 6 lines · Ira Weiny

Using the proposed type-2 cxl-test device[1] the following splat was observed:

  BUG: kernel NULL pointer dereference, address: 0000000000000278
  [...]
  RIP: 0010:devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]

It would be useful to decode this to a line number, the rest of this call trace is not adding much.

Quoted 23 lines

[...]

  Call Trace:
   <TASK>
   ? __die+0x1f/0x70
   ? page_fault_oops+0x149/0x420
   ? fixup_exception+0x22/0x310
   ? kernelmode_fixup_or_oops+0x84/0x110
   ? exc_page_fault+0x6d/0x150
   ? asm_exc_page_fault+0x22/0x30
   ? devm_cxl_add_memdev+0x1de/0x2c0 [cxl_core]
   cxl_mock_mem_probe+0x632/0x870 [cxl_mock_mem]
   platform_probe+0x40/0x90
   really_probe+0x19e/0x3e0
   ? __pfx___driver_attach+0x10/0x10
   __driver_probe_device+0x78/0x160
   driver_probe_device+0x1f/0x90
   __driver_attach+0xce/0x1c0
   bus_for_each_dev+0x63/0xa0
   bus_add_driver+0x112/0x210
   driver_register+0x55/0x100
   ? __pfx_cxl_mock_mem_driver_init+0x10/0x10 [cxl_mock_mem]
   [...]
Quoted 4 lines

Commit f6b8ab32e3ec made the mailbox functionality optional. However, some mailbox functionality was merged after that patch. Therefore some mailbox functionality can be accessed on a device which did not set up the mailbox.

Quoted 5 lines

cxl_memdev_security_init() definitely needs to move out of devm_cxl_add_memdev() and after that I do not think @mds NULL checks need to be sprinkled everywhere. In other words something is wrong at a higher level if we get into some of these helper functions without the memory device state.

Quoted 8 lines

Right, so we can move it directly into cxl_pci_probe() - just as with other mbox based functionality. This leaves me wondering, however, what to do about the cxl_memdev_security_shutdown() counterpart. As with the below diff, leaving it as is and just adding a mds nil check might still be considering a layering violation in that it would be asymmetrical wrt to the init; but this is tightly coupled with cxl_memdev_unregister().

Ira does the below fix the crash?

I had to apply it by hand but yea it fixes the immediate crash.

Did you want to submit that as part of other work?

Ira

Quoted 82 lines

Thanks,
Davidlohr

----8<-------

drivers/cxl/core/memdev.c +7 -10
index 14b547c07f54..4d1bf80c0e54 100644
@@ -561,7 +561,7 @@ static void cxl_memdev_security_shutdown(struct device *dev)
  	struct cxl_memdev *cxlmd = to_cxl_memdev(dev);
  	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlmd->cxlds);
 
-	if (mds->security.poll)
+	if (mds && mds->security.poll)
  		cancel_delayed_work_sync(&mds->security.poll_dwork);
  }
 
@@ -1009,11 +1009,11 @@ static void put_sanitize(void *data)
  	sysfs_put(mds->security.sanitize_node);
  }
 
-static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
+int cxl_memdev_security_state_init(struct cxl_memdev_state *mds)
  {
-	struct cxl_dev_state *cxlds = cxlmd->cxlds;
-	struct cxl_memdev_state *mds = to_cxl_memdev_state(cxlds);
-	struct device *dev = &cxlmd->dev;
+
+	struct cxl_dev_state *cxlds = &mds->cxlds;
+	struct device *dev = &cxlds->cxlmd->dev;
  	struct kernfs_node *sec;
 
  	sec = sysfs_get_dirent(dev->kobj.sd, "security");
@@ -1029,7 +1029,8 @@ static int cxl_memdev_security_init(struct cxl_memdev *cxlmd)
  	}
 
  	return devm_add_action_or_reset(cxlds->dev, put_sanitize, mds);
- }
+}
+EXPORT_SYMBOL_NS_GPL(cxl_memdev_security_state_init, CXL);
 
  struct cxl_memdev *devm_cxl_add_memdev(struct cxl_dev_state *cxlds)
  {
@@ -1059,10 +1060,6 @@ struct cxl_memdev *devm_cxl_add_memdev(struct cxl_dev_state *cxlds)
  	if (rc)
  		goto err;
 
-	rc = cxl_memdev_security_init(cxlmd);
-	if (rc)
-		goto err;
-
  	rc = devm_add_action_or_reset(cxlds->dev, cxl_memdev_unregister, cxlmd);
  	if (rc)
  		return ERR_PTR(rc);
drivers/cxl/cxlmem.h +1 -0
index f86afef90c91..441270770519 100644
@@ -884,6 +884,7 @@ static inline void cxl_mem_active_dec(void)
  #endif
 
  int cxl_mem_sanitize(struct cxl_memdev_state *mds, u16 cmd);
+int cxl_memdev_security_state_init(struct cxl_memdev_state *mds);
 
  struct cxl_hdm {
  	struct cxl_component_regs regs;
drivers/cxl/pci.c +4 -0
index 1cb1494c28fe..5242dbf0044d 100644
@@ -887,6 +887,10 @@ static int cxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
  	if (IS_ERR(cxlmd))
  		return PTR_ERR(cxlmd);
 
+	rc = cxl_memdev_security_state_init(mds);
+	if (rc)
+		return rc;
+
  	rc = cxl_memdev_setup_fw_upload(mds);
  	if (rc)
  		return rc;