Skip to contentSkip to conversation

ipmi: Fix I2C client removal in the SSIF driver

PATCH In review [email protected] 2 participants 4 messages Aug 30, 2018 – Aug 31, 2018 IPMI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipmi: Fix I2C client removal in the SSIF driver

From: Corey Minyard <[email protected]>

The SSIF driver was removing any client that came in through the platform interface, but it should only remove clients that it added. On a failure in the probe function, this could result in the following oops when the driver is removed and the client gets unregistered twice:

 CPU: 107 PID: 30266 Comm: rmmod Not tainted 4.18.0+ #80
 Hardware name: Cavium Inc. Saber/Saber, BIOS Cavium reference firmware version 7.0 08/04/2018
 pstate: 60400009 (nZCv daif +PAN -UAO)
 pc : kernfs_find_ns+0x28/0x120
 lr : kernfs_find_and_get_ns+0x40/0x60
 sp : ffff00002310fb50
 x29: ffff00002310fb50 x28: ffff800a8240f800
 x27: 0000000000000000 x26: 0000000000000000
 x25: 0000000056000000 x24: ffff000009073000
 x23: ffff000008998b38 x22: 0000000000000000
 x21: ffff800ed86de820 x20: 0000000000000000
 x19: ffff00000913a1d8 x18: 0000000000000000
 x17: 0000000000000000 x16: 0000000000000000
 x15: 0000000000000000 x14: 5300737265766972
 x13: 643d4d4554535953 x12: 0000000000000030
 x11: 0000000000000030 x10: 0101010101010101
 x9 : ffff800ea06cc3f9 x8 : 0000000000000000
 x7 : 0000000000000141 x6 : ffff000009073000
 x5 : ffff800adb706b00 x4 : 0000000000000000
 x3 : 00000000ffffffff x2 : 0000000000000000
 x1 : ffff000008998b38 x0 : ffff000008356760
 Process rmmod (pid: 30266, stack limit = 0x00000000e218418d)
 Call trace:
  kernfs_find_ns+0x28/0x120
  kernfs_find_and_get_ns+0x40/0x60
  sysfs_unmerge_group+0x2c/0x6c
  dpm_sysfs_remove+0x34/0x70
  device_del+0x58/0x30c
  device_unregister+0x30/0x7c
  i2c_unregister_device+0x84/0x90 [i2c_core]
  ssif_platform_remove+0x38/0x98 [ipmi_ssif]
  platform_drv_remove+0x2c/0x6c
  device_release_driver_internal+0x168/0x1f8
  driver_detach+0x50/0xbc
  bus_remove_driver+0x74/0xe8
  driver_unregister+0x34/0x5c
  platform_driver_unregister+0x20/0x2c
  cleanup_ipmi_ssif+0x50/0xd82c [ipmi_ssif]
  __arm64_sys_delete_module+0x1b4/0x220
  el0_svc_handler+0x104/0x160
  el0_svc+0x8/0xc
 Code: aa1e03e0 aa0203f6 aa0103f7 d503201f (7940e280)
 ---[ end trace 09f0e34cce8e2d8c ]---
 Kernel panic - not syncing: Fatal exception
 SMP: stopping secondary CPUs
 Kernel Offset: disabled
 CPU features: 0x23800c38

So track the clients that the SSIF driver adds and only remove those.

---
 drivers/char/ipmi/ipmi_ssif.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)

George, your fix was close, but not quite right. The following should fix the problem, obvious in hindsight :-). Thanks for working with me on this.

-corey

drivers/char/ipmi/ipmi_ssif.c +6 -11
index e7e8b86..ca4f7c4 100644
@@ -182,6 +182,8 @@ struct ssif_addr_info {
 	struct device *dev;
 	struct i2c_client *client;
 
+	struct i2c_client *added_client;
+
 	struct mutex clients_mutex;
 	struct list_head clients;
 
@@ -1639,15 +1641,7 @@ static int ssif_probe(struct i2c_client *client, const struct i2c_device_id *id)
 
  out:
 	if (rv) {
-		/*
-		 * Note that if addr_info->client is assigned, we
-		 * leave it.  The i2c client hangs around even if we
-		 * return a failure here, and the failure here is not
-		 * propagated back to the i2c code.  This seems to be
-		 * design intent, strange as it may be.  But if we
-		 * don't leave it, ssif_platform_remove will not remove
-		 * the client like it should.
-		 */
+		addr_info->client = NULL;
 		dev_err(&client->dev, "Unable to start IPMI SSIF: %d\n", rv);
 		kfree(ssif_info);
 	}
@@ -1667,7 +1661,8 @@ static int ssif_adapter_handler(struct device *adev, void *opaque)
 	if (adev->type != &i2c_adapter_type)
 		return 0;
 
-	i2c_new_device(to_i2c_adapter(adev), &addr_info->binfo);
+	addr_info->added_client = i2c_new_device(to_i2c_adapter(adev),
+						 &addr_info->binfo);
 
 	if (!addr_info->adapter_name)
 		return 1; /* Only try the first I2C adapter by default. */
@@ -1840,7 +1835,7 @@ static int ssif_platform_remove(struct platform_device *dev)
 		return 0;
 
 	mutex_lock(&ssif_infos_mutex);
-	i2c_unregister_device(addr_info->client);
+	i2c_unregister_device(addr_info->added_client);
 
 	list_del(&addr_info->link);
 	kfree(addr_info);
Signature · 1 line
2.7.4
#1 lore raw
Quoted 62 lines · On 08/30/2018 11:44 PM, [email protected] wrote:

From: Corey Minyard <[email protected]>

The SSIF driver was removing any client that came in through the platform interface, but it should only remove clients that it added. On a failure in the probe function, this could result in the following oops when the driver is removed and the client gets unregistered twice:

  CPU: 107 PID: 30266 Comm: rmmod Not tainted 4.18.0+ #80
  Hardware name: Cavium Inc. Saber/Saber, BIOS Cavium reference firmware version 7.0 08/04/2018
  pstate: 60400009 (nZCv daif +PAN -UAO)
  pc : kernfs_find_ns+0x28/0x120
  lr : kernfs_find_and_get_ns+0x40/0x60
  sp : ffff00002310fb50
  x29: ffff00002310fb50 x28: ffff800a8240f800
  x27: 0000000000000000 x26: 0000000000000000
  x25: 0000000056000000 x24: ffff000009073000
  x23: ffff000008998b38 x22: 0000000000000000
  x21: ffff800ed86de820 x20: 0000000000000000
  x19: ffff00000913a1d8 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 5300737265766972
  x13: 643d4d4554535953 x12: 0000000000000030
  x11: 0000000000000030 x10: 0101010101010101
  x9 : ffff800ea06cc3f9 x8 : 0000000000000000
  x7 : 0000000000000141 x6 : ffff000009073000
  x5 : ffff800adb706b00 x4 : 0000000000000000
  x3 : 00000000ffffffff x2 : 0000000000000000
  x1 : ffff000008998b38 x0 : ffff000008356760
  Process rmmod (pid: 30266, stack limit = 0x00000000e218418d)
  Call trace:
   kernfs_find_ns+0x28/0x120
   kernfs_find_and_get_ns+0x40/0x60
   sysfs_unmerge_group+0x2c/0x6c
   dpm_sysfs_remove+0x34/0x70
   device_del+0x58/0x30c
   device_unregister+0x30/0x7c
   i2c_unregister_device+0x84/0x90 [i2c_core]
   ssif_platform_remove+0x38/0x98 [ipmi_ssif]
   platform_drv_remove+0x2c/0x6c
   device_release_driver_internal+0x168/0x1f8
   driver_detach+0x50/0xbc
   bus_remove_driver+0x74/0xe8
   driver_unregister+0x34/0x5c
   platform_driver_unregister+0x20/0x2c
   cleanup_ipmi_ssif+0x50/0xd82c [ipmi_ssif]
   __arm64_sys_delete_module+0x1b4/0x220
   el0_svc_handler+0x104/0x160
   el0_svc+0x8/0xc
  Code: aa1e03e0 aa0203f6 aa0103f7 d503201f (7940e280)
  ---[ end trace 09f0e34cce8e2d8c ]---
  Kernel panic - not syncing: Fatal exception
  SMP: stopping secondary CPUs
  Kernel Offset: disabled
  CPU features: 0x23800c38

So track the clients that the SSIF driver adds and only remove those.

Quoted 8 lines
---
  drivers/char/ipmi/ipmi_ssif.c | 17 ++++++-----------
  1 file changed, 6 insertions(+), 11 deletions(-)

George, your fix was close, but not quite right. The following should fix the problem, obvious in hindsight :-). Thanks for working with me on this.

Thanks Corey!!!
This works for me now.

Quoted 54 lines
(quoted diff) +0 -1
-corey
drivers/char/ipmi/ipmi_ssif.c +6 -12
index e7e8b86..ca4f7c4 100644
@@ -182,6 +182,8 @@ struct ssif_addr_info {
         struct device *dev;
         struct i2c_client *client;
 
+       struct i2c_client *added_client;
+
         struct mutex clients_mutex;
         struct list_head clients;
 
@@ -1639,15 +1641,7 @@ static int ssif_probe(struct i2c_client *client, const struct i2c_device_id *id)
 
   out:
         if (rv) {
-               /*
-                * Note that if addr_info->client is assigned, we
-                * leave it.  The i2c client hangs around even if we
-                * return a failure here, and the failure here is not
-                * propagated back to the i2c code.  This seems to be
-                * design intent, strange as it may be.  But if we
-                * don't leave it, ssif_platform_remove will not remove
-                * the client like it should.
-                */
+               addr_info->client = NULL;
                 dev_err(&client->dev, "Unable to start IPMI SSIF: %d\n", rv);
                 kfree(ssif_info);
         }
@@ -1667,7 +1661,8 @@ static int ssif_adapter_handler(struct device *adev, void *opaque)
         if (adev->type != &i2c_adapter_type)
                 return 0;
 
-       i2c_new_device(to_i2c_adapter(adev), &addr_info->binfo);
+       addr_info->added_client = i2c_new_device(to_i2c_adapter(adev),
+                                                &addr_info->binfo);
 
         if (!addr_info->adapter_name)
                 return 1; /* Only try the first I2C adapter by default. */
@@ -1840,7 +1835,7 @@ static int ssif_platform_remove(struct platform_device *dev)
                 return 0;
 
         mutex_lock(&ssif_infos_mutex);
-       i2c_unregister_device(addr_info->client);
+       i2c_unregister_device(addr_info->added_client);
 
         list_del(&addr_info->link);
         kfree(addr_info);
--

2.7.4

#2 lore raw
Quoted 69 lines · George Cherian
Quoted 66 lines · On 08/30/2018 11:44 PM, [email protected] wrote:

From: Corey Minyard <[email protected]>

The SSIF driver was removing any client that came in through the platform interface, but it should only remove clients that it added.  On a failure in the probe function, this could result in the following oops when the driver is removed and the client gets unregistered twice:

  CPU: 107 PID: 30266 Comm: rmmod Not tainted 4.18.0+ #80
  Hardware name: Cavium Inc. Saber/Saber, BIOS Cavium reference firmware version 7.0 08/04/2018
  pstate: 60400009 (nZCv daif +PAN -UAO)
  pc : kernfs_find_ns+0x28/0x120
  lr : kernfs_find_and_get_ns+0x40/0x60

  sp : ffff00002310fb50
  x29: ffff00002310fb50 x28: ffff800a8240f800
  x27: 0000000000000000 x26: 0000000000000000
  x25: 0000000056000000 x24: ffff000009073000
  x23: ffff000008998b38 x22: 0000000000000000
  x21: ffff800ed86de820 x20: 0000000000000000
  x19: ffff00000913a1d8 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 5300737265766972
  x13: 643d4d4554535953 x12: 0000000000000030
  x11: 0000000000000030 x10: 0101010101010101
  x9 : ffff800ea06cc3f9 x8 : 0000000000000000
  x7 : 0000000000000141 x6 : ffff000009073000
  x5 : ffff800adb706b00 x4 : 0000000000000000
  x3 : 00000000ffffffff x2 : 0000000000000000
  x1 : ffff000008998b38 x0 : ffff000008356760 Process rmmod (pid: 30266, stack limit = 0x00000000e218418d)

  Call trace:
   kernfs_find_ns+0x28/0x120
   kernfs_find_and_get_ns+0x40/0x60
   sysfs_unmerge_group+0x2c/0x6c
   dpm_sysfs_remove+0x34/0x70
   device_del+0x58/0x30c
   device_unregister+0x30/0x7c
   i2c_unregister_device+0x84/0x90 [i2c_core]
   ssif_platform_remove+0x38/0x98 [ipmi_ssif]
   platform_drv_remove+0x2c/0x6c
   device_release_driver_internal+0x168/0x1f8
   driver_detach+0x50/0xbc
   bus_remove_driver+0x74/0xe8
   driver_unregister+0x34/0x5c
   platform_driver_unregister+0x20/0x2c
   cleanup_ipmi_ssif+0x50/0xd82c [ipmi_ssif]
   __arm64_sys_delete_module+0x1b4/0x220
   el0_svc_handler+0x104/0x160
   el0_svc+0x8/0xc
  Code: aa1e03e0 aa0203f6 aa0103f7 d503201f (7940e280)
  ---[ end trace 09f0e34cce8e2d8c ]---
  Kernel panic - not syncing: Fatal exception

  SMP: stopping secondary CPUs
  Kernel Offset: disabled
  CPU features: 0x23800c38

So track the clients that the SSIF driver adds and only remove those.

---
  drivers/char/ipmi/ipmi_ssif.c | 17 ++++++-----------

  1 file changed, 6 insertions(+), 11 deletions(-)
Quoted 7 lines
Quoted 4 lines

George, your fix was close, but not quite right.  The following should fix the problem, obvious in hindsight :-).  Thanks for working with me on this.

Thanks Corey!!!
This works for me now.

Thanks.  Can I add a Tested-by: from you?

-corey

Quoted 57 lines
Quoted 57 lines
(quoted diff) +0 -1
-corey
drivers/char/ipmi/ipmi_ssif.c +2 -0
index e7e8b86..ca4f7c4 100644
@@ -182,6 +182,8 @@ struct ssif_addr_info {
          struct device *dev;
          struct i2c_client *client;
 
+       struct i2c_client *added_client;
+
          struct mutex clients_mutex;
          struct list_head clients;
 
@@ -1639,15 +1641,7 @@ static int ssif_probe(struct i2c_client 
 *client, const struct i2c_device_id *id)
 

   out:
         if (rv) {
-               /*
-                * Note that if addr_info->client is assigned, we -                * leave it.  The i2c client hangs around even if we -                * return a failure here, and the failure here is not -                * propagated back to the i2c code.  This seems to be -                * design intent, strange as it may be.  But if we -                * don't leave it, ssif_platform_remove will not remove -                * the client like it should.
-                */
+               addr_info->client = NULL;
                 dev_err(&client->dev, "Unable to start IPMI SSIF: %d\n", rv); kfree(ssif_info);
         }

drivers/char/ipmi/ipmi_ssif.c +2 -1
@@ -1667,7 +1661,8 @@ static int ssif_adapter_handler(struct device 
 *adev, void *opaque)
          if (adev->type != &i2c_adapter_type)
                  return 0;
 
-       i2c_new_device(to_i2c_adapter(adev), &addr_info->binfo);
+       addr_info->added_client = i2c_new_device(to_i2c_adapter(adev),
+ &addr_info->binfo);
 
          if (!addr_info->adapter_name)

                 return 1; /* Only try the first I2C adapter by default. */

drivers/char/ipmi/ipmi_ssif.c +1 -1
@@ -1840,7 +1835,7 @@ static int ssif_platform_remove(struct 
 platform_device *dev)
                  return 0;
 
          mutex_lock(&ssif_infos_mutex);
-       i2c_unregister_device(addr_info->client);
+       i2c_unregister_device(addr_info->added_client);
 
          list_del(&addr_info->link);

         kfree(addr_info);

Signature · 1 line
2.7.4
Tested-by
#3 lore raw
Quoted 76 lines · Corey Minyard
Quoted 74 lines · George Cherian
Quoted 71 lines · On 08/30/2018 11:44 PM, [email protected] wrote:

From: Corey Minyard <[email protected]>

The SSIF driver was removing any client that came in through the platform interface, but it should only remove clients that it added.  On a failure in the probe function, this could result in the following oops when the driver is removed and the client gets unregistered twice:

  CPU: 107 PID: 30266 Comm: rmmod Not tainted 4.18.0+ #80
  Hardware name: Cavium Inc. Saber/Saber, BIOS Cavium reference
firmware version 7.0 08/04/2018
  pstate: 60400009 (nZCv daif +PAN -UAO)
  pc : kernfs_find_ns+0x28/0x120
  lr : kernfs_find_and_get_ns+0x40/0x60

  sp : ffff00002310fb50
  x29: ffff00002310fb50 x28: ffff800a8240f800
  x27: 0000000000000000 x26: 0000000000000000
  x25: 0000000056000000 x24: ffff000009073000
  x23: ffff000008998b38 x22: 0000000000000000
  x21: ffff800ed86de820 x20: 0000000000000000
  x19: ffff00000913a1d8 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 5300737265766972
  x13: 643d4d4554535953 x12: 0000000000000030
  x11: 0000000000000030 x10: 0101010101010101
  x9 : ffff800ea06cc3f9 x8 : 0000000000000000
  x7 : 0000000000000141 x6 : ffff000009073000
  x5 : ffff800adb706b00 x4 : 0000000000000000
  x3 : 00000000ffffffff x2 : 0000000000000000
  x1 : ffff000008998b38 x0 : ffff000008356760 Process rmmod (pid: 30266, stack limit = 0x00000000e218418d)

  Call trace:
   kernfs_find_ns+0x28/0x120
   kernfs_find_and_get_ns+0x40/0x60
   sysfs_unmerge_group+0x2c/0x6c
   dpm_sysfs_remove+0x34/0x70
   device_del+0x58/0x30c
   device_unregister+0x30/0x7c
   i2c_unregister_device+0x84/0x90 [i2c_core]
   ssif_platform_remove+0x38/0x98 [ipmi_ssif]
   platform_drv_remove+0x2c/0x6c
   device_release_driver_internal+0x168/0x1f8
   driver_detach+0x50/0xbc
   bus_remove_driver+0x74/0xe8
   driver_unregister+0x34/0x5c
   platform_driver_unregister+0x20/0x2c
   cleanup_ipmi_ssif+0x50/0xd82c [ipmi_ssif]
   __arm64_sys_delete_module+0x1b4/0x220
   el0_svc_handler+0x104/0x160
   el0_svc+0x8/0xc
  Code: aa1e03e0 aa0203f6 aa0103f7 d503201f (7940e280)
  ---[ end trace 09f0e34cce8e2d8c ]---
  Kernel panic - not syncing: Fatal exception

  SMP: stopping secondary CPUs
  Kernel Offset: disabled
  CPU features: 0x23800c38

So track the clients that the SSIF driver adds and only remove those.

---
  drivers/char/ipmi/ipmi_ssif.c | 17 ++++++-----------

  1 file changed, 6 insertions(+), 11 deletions(-)

George, your fix was close, but not quite right.  The following should fix the problem, obvious in hindsight :-).  Thanks for working with me on this.

Quoted 5 lines
Quoted 3 lines

Thanks Corey!!!
This works for me now.

Thanks.  Can I add a Tested-by: from you?

Yes.

Can you CC this to stable too.

-George

Quoted 64 lines

-corey

Quoted 60 lines
Quoted 60 lines
(quoted diff) +0 -1
-corey
drivers/char/ipmi/ipmi_ssif.c +2 -0
index e7e8b86..ca4f7c4 100644
@@ -182,6 +182,8 @@ struct ssif_addr_info {
          struct device *dev;
          struct i2c_client *client;
 
+       struct i2c_client *added_client;
+
          struct mutex clients_mutex;
          struct list_head clients;
 
@@ -1639,15 +1641,7 @@ static int ssif_probe(struct i2c_client
 *client, const struct i2c_device_id *id)
 

   out:
         if (rv) {
-               /*
-                * Note that if addr_info->client is assigned, we -                * leave it.  The i2c client hangs around even if we -                * return a failure here, and the failure here is not -                * propagated back to the i2c code.  This seems to be -                * design intent, strange as it may be.  But if we -                * don't leave it, ssif_platform_remove will not remove -                * the client like it should.
-                */
+               addr_info->client = NULL; dev_err(&client->dev, "Unable to start IPMI SSIF:
%d\n", rv);
                 kfree(ssif_info);
         }

drivers/char/ipmi/ipmi_ssif.c +2 -1
@@ -1667,7 +1661,8 @@ static int ssif_adapter_handler(struct device
 *adev, void *opaque)
          if (adev->type != &i2c_adapter_type)
                  return 0;
 
-       i2c_new_device(to_i2c_adapter(adev), &addr_info->binfo);
+       addr_info->added_client = i2c_new_device(to_i2c_adapter(adev),
+ &addr_info->binfo);
 
          if (!addr_info->adapter_name)

                 return 1; /* Only try the first I2C adapter by default. */

drivers/char/ipmi/ipmi_ssif.c +1 -1
@@ -1840,7 +1835,7 @@ static int ssif_platform_remove(struct
 platform_device *dev)
                  return 0;
 
          mutex_lock(&ssif_infos_mutex);
-       i2c_unregister_device(addr_info->client);
+       i2c_unregister_device(addr_info->added_client);
 
          list_del(&addr_info->link);

         kfree(addr_info);

Signature · 1 line
2.7.4