Re: [PATCH v3] usb: gadget: configfs: Fix KASAN use-after-free

Re: [PATCH v3] usb: gadget: configfs: Fix KASAN use-after-free

[Thadeu Lima de Souza Cascardo]

On Tue, Jan 17, 2017 at 12:29:09PM +0200, Felipe Balbi wrote:
> 
> Hi,
> 
> Jim Lin <jilin@nvidia.com> writes:
> > When gadget is disconnected, running sequence is like this.
> > . composite_disconnect
> > . Call trace:
> >   usb_string_copy+0xd0/0x128
> >   gadget_config_name_configuration_store+0x4
> >   gadget_config_name_attr_store+0x40/0x50
> >   configfs_write_file+0x198/0x1f4
> >   vfs_write+0x100/0x220
> >   SyS_write+0x58/0xa8
> > . configfs_composite_unbind
> > . configfs_composite_bind
> >
> > In configfs_composite_bind, it has
> > "cn->strings.s = cn->configuration;"
> >
> > When usb_string_copy is invoked. it would
> > allocate memory, copy input string, release previous pointed memory space,
> > and use new allocated memory.
> >
> > When gadget is connected, host sends down request to get information.
> > Call trace:
> >   usb_gadget_get_string+0xec/0x168
> >   lookup_string+0x64/0x98
> >   composite_setup+0xa34/0x1ee8
> >
> > If gadget is disconnected and connected quickly, in the failed case,
> > cn->configuration memory has been released by usb_string_copy kfree but
> > configfs_composite_bind hasn't been run in time to assign new allocated
> > "cn->configuration" pointer to "cn->strings.s".
> >
> > When "strlen(s->s) of usb_gadget_get_string is being executed, the dangling
> > memory is accessed, "BUG: KASAN: use-after-free" error occurs.
> >
> > Signed-off-by: Jim Lin <jilin@nvidia.com>
> > ---
> > Changes in v2:
> > Changes in v3:
> >  Change commit description
> 
> well, I need to be sure you tested this with Linus' tree. The reason I'm
> asking is because this could be a bug caused by Android changes. From
> your previous patch, the problem started with android_setup().
> 
> Please test with v4.10-rc4 and any configfs-based gadget.
> 
> -- 
> balbi

I tested this with dummy_hcd on top of a 5.8 kernel and I got lsusb to respond
with an error instead of the right manufacturer string, after overwriting such
a string after binding.

With the patch applied, after the string is overwritten, lsusb will show the
updated string.

Because of commit 81c7462883b0cc0a4eeef0687f80ad5b5baee5f6 ("USB: replace
hardcode maximum usb string length by definition"), the patch will need a
fixup. Should I send a v2 with my sign-off?

Thanks.
Cascardo.

[PATCH v4] usb: gadget: configfs: Fix KASAN use-after-free

[Macpaul Lin]

From: Jim Lin <jilin@nvidia.com>

When gadget is disconnected, running sequence is like this.
. composite_disconnect
. Call trace:
  usb_string_copy+0xd0/0x128
  gadget_config_name_configuration_store+0x4
  gadget_config_name_attr_store+0x40/0x50
  configfs_write_file+0x198/0x1f4
  vfs_write+0x100/0x220
  SyS_write+0x58/0xa8
. configfs_composite_unbind
. configfs_composite_bind

In configfs_composite_bind, it has
"cn->strings.s = cn->configuration;"

When usb_string_copy is invoked. it would
allocate memory, copy input string, release previous pointed memory space,
and use new allocated memory.

When gadget is connected, host sends down request to get information.
Call trace:
  usb_gadget_get_string+0xec/0x168
  lookup_string+0x64/0x98
  composite_setup+0xa34/0x1ee8

If gadget is disconnected and connected quickly, in the failed case,
cn->configuration memory has been released by usb_string_copy kfree but
configfs_composite_bind hasn't been run in time to assign new allocated
"cn->configuration" pointer to "cn->strings.s".

When "strlen(s->s) of usb_gadget_get_string is being executed, the dangling
memory is accessed, "BUG: KASAN: use-after-free" error occurs.

Signed-off-by: Jim Lin <jilin@nvidia.com>
Signed-off-by: Macpaul Lin <macpaul.lin@mediatek.com>
Cc: stable@vger.kernel.org
---
Changes in v2:
Changes in v3:
 - Change commit description
Changes in v4:
 - Fix build error and adapt patch to kernel-5.12-rc1.
   Replace definition "MAX_USB_STRING_WITH_NULL_LEN" with
   "USB_MAX_STRING_WITH_NULL_LEN".
 - Note: The patch v2 and v3 has been verified by
   Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
   http://spinics.net/lists/kernel/msg3840792.html
   and
   Macpaul Lin <macpaul.lin@mediatek.com> on Android kernels.
   http://lkml.org/lkml/2020/6/11/8
 - The patch is suggested to be applied to LTS versions.

 drivers/usb/gadget/configfs.c |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/drivers/usb/gadget/configfs.c b/drivers/usb/gadget/configfs.c
index 0d56f33..15a607c 100644
--- a/drivers/usb/gadget/configfs.c
+++ b/drivers/usb/gadget/configfs.c
@@ -97,6 +97,8 @@ struct gadget_config_name {
 	struct list_head list;
 };
 
+#define USB_MAX_STRING_WITH_NULL_LEN	(USB_MAX_STRING_LEN+1)
+
 static int usb_string_copy(const char *s, char **s_copy)
 {
 	int ret;
@@ -106,12 +108,16 @@ static int usb_string_copy(const char *s, char **s_copy)
 	if (ret > USB_MAX_STRING_LEN)
 		return -EOVERFLOW;
 
-	str = kstrdup(s, GFP_KERNEL);
-	if (!str)
-		return -ENOMEM;
+	if (copy) {
+		str = copy;
+	} else {
+		str = kmalloc(USB_MAX_STRING_WITH_NULL_LEN, GFP_KERNEL);
+		if (!str)
+			return -ENOMEM;
+	}
+	strcpy(str, s);
 	if (str[ret - 1] == '\n')
 		str[ret - 1] = '\0';
-	kfree(copy);
 	*s_copy = str;
 	return 0;
 }
-- 
1.7.9.5

Re: [PATCH v4] usb: gadget: configfs: Fix KASAN use-after-free

[Macpaul Lin]

On Thu, 2021-03-11 at 14:42 +0800, Macpaul Lin wrote:
> From: Jim Lin <jilin@nvidia.com>
> 
> When gadget is disconnected, running sequence is like this.
> . composite_disconnect
> . Call trace:
>   usb_string_copy+0xd0/0x128
>   gadget_config_name_configuration_store+0x4
>   gadget_config_name_attr_store+0x40/0x50
>   configfs_write_file+0x198/0x1f4
>   vfs_write+0x100/0x220
>   SyS_write+0x58/0xa8
> . configfs_composite_unbind
> . configfs_composite_bind
> 
> In configfs_composite_bind, it has
> "cn->strings.s = cn->configuration;"
> 
> When usb_string_copy is invoked. it would
> allocate memory, copy input string, release previous pointed memory space,
> and use new allocated memory.
> 
> When gadget is connected, host sends down request to get information.
> Call trace:
>   usb_gadget_get_string+0xec/0x168
>   lookup_string+0x64/0x98
>   composite_setup+0xa34/0x1ee8
> 
> If gadget is disconnected and connected quickly, in the failed case,
> cn->configuration memory has been released by usb_string_copy kfree but
> configfs_composite_bind hasn't been run in time to assign new allocated
> "cn->configuration" pointer to "cn->strings.s".
> 
> When "strlen(s->s) of usb_gadget_get_string is being executed, the dangling
> memory is accessed, "BUG: KASAN: use-after-free" error occurs.
> 
> Signed-off-by: Jim Lin <jilin@nvidia.com>
> Signed-off-by: Macpaul Lin <macpaul.lin@mediatek.com>
> Cc: stable@vger.kernel.org
> ---
> Changes in v2:
> Changes in v3:
>  - Change commit description
> Changes in v4:
>  - Fix build error and adapt patch to kernel-5.12-rc1.
>    Replace definition "MAX_USB_STRING_WITH_NULL_LEN" with
>    "USB_MAX_STRING_WITH_NULL_LEN".
>  - Note: The patch v2 and v3 has been verified by
>    Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
>    http://spinics.net/lists/kernel/msg3840792.html

Dear Cascardo,

Would you please help to confirm if you've tested it on Linux PC,
Chrome OS, or an Android OS?

Thanks!
Macpaul Lin

>    and
>    Macpaul Lin <macpaul.lin@mediatek.com> on Android kernels.
>    http://lkml.org/lkml/2020/6/11/8
>  - The patch is suggested to be applied to LTS versions.
> 
>  drivers/usb/gadget/configfs.c |   14 ++++++++++----
>  1 file changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/usb/gadget/configfs.c b/drivers/usb/gadget/configfs.c
> index 0d56f33..15a607c 100644
> --- a/drivers/usb/gadget/configfs.c
> +++ b/drivers/usb/gadget/configfs.c
> @@ -97,6 +97,8 @@ struct gadget_config_name {
>  	struct list_head list;
>  };
>  
> +#define USB_MAX_STRING_WITH_NULL_LEN	(USB_MAX_STRING_LEN+1)
> +
>  static int usb_string_copy(const char *s, char **s_copy)
>  {
>  	int ret;
> @@ -106,12 +108,16 @@ static int usb_string_copy(const char *s, char **s_copy)
>  	if (ret > USB_MAX_STRING_LEN)
>  		return -EOVERFLOW;
>  
> -	str = kstrdup(s, GFP_KERNEL);
> -	if (!str)
> -		return -ENOMEM;
> +	if (copy) {
> +		str = copy;
> +	} else {
> +		str = kmalloc(USB_MAX_STRING_WITH_NULL_LEN, GFP_KERNEL);
> +		if (!str)
> +			return -ENOMEM;
> +	}
> +	strcpy(str, s);
>  	if (str[ret - 1] == '\n')
>  		str[ret - 1] = '\0';
> -	kfree(copy);
>  	*s_copy = str;
>  	return 0;
>  }

Re: [PATCH v4] usb: gadget: configfs: Fix KASAN use-after-free

[Thadeu Lima de Souza Cascardo]

On Thu, Mar 11, 2021 at 02:53:52PM +0800, Macpaul Lin wrote:
> On Thu, 2021-03-11 at 14:42 +0800, Macpaul Lin wrote:
> > From: Jim Lin <jilin@nvidia.com>
> > 
> > When gadget is disconnected, running sequence is like this.
> > . composite_disconnect
> > . Call trace:
> >   usb_string_copy+0xd0/0x128
> >   gadget_config_name_configuration_store+0x4
> >   gadget_config_name_attr_store+0x40/0x50
> >   configfs_write_file+0x198/0x1f4
> >   vfs_write+0x100/0x220
> >   SyS_write+0x58/0xa8
> > . configfs_composite_unbind
> > . configfs_composite_bind
> > 
> > In configfs_composite_bind, it has
> > "cn->strings.s = cn->configuration;"
> > 
> > When usb_string_copy is invoked. it would
> > allocate memory, copy input string, release previous pointed memory space,
> > and use new allocated memory.
> > 
> > When gadget is connected, host sends down request to get information.
> > Call trace:
> >   usb_gadget_get_string+0xec/0x168
> >   lookup_string+0x64/0x98
> >   composite_setup+0xa34/0x1ee8
> > 
> > If gadget is disconnected and connected quickly, in the failed case,
> > cn->configuration memory has been released by usb_string_copy kfree but
> > configfs_composite_bind hasn't been run in time to assign new allocated
> > "cn->configuration" pointer to "cn->strings.s".
> > 
> > When "strlen(s->s) of usb_gadget_get_string is being executed, the dangling
> > memory is accessed, "BUG: KASAN: use-after-free" error occurs.
> > 
> > Signed-off-by: Jim Lin <jilin@nvidia.com>
> > Signed-off-by: Macpaul Lin <macpaul.lin@mediatek.com>
> > Cc: stable@vger.kernel.org
> > ---
> > Changes in v2:
> > Changes in v3:
> >  - Change commit description
> > Changes in v4:
> >  - Fix build error and adapt patch to kernel-5.12-rc1.
> >    Replace definition "MAX_USB_STRING_WITH_NULL_LEN" with
> >    "USB_MAX_STRING_WITH_NULL_LEN".
> >  - Note: The patch v2 and v3 has been verified by
> >    Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
> >    http://spinics.net/lists/kernel/msg3840792.html
> 
> Dear Cascardo,
> 
> Would you please help to confirm if you've tested it on Linux PC,
> Chrome OS, or an Android OS?

I tested v3 on Ubuntu GNU/Linux. I will test v4.

Cascardo.

> 
> Thanks!
> Macpaul Lin
> 
> >    and
> >    Macpaul Lin <macpaul.lin@mediatek.com> on Android kernels.
> >    http://lkml.org/lkml/2020/6/11/8
> >  - The patch is suggested to be applied to LTS versions.
> > 
> >  drivers/usb/gadget/configfs.c |   14 ++++++++++----
> >  1 file changed, 10 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/usb/gadget/configfs.c b/drivers/usb/gadget/configfs.c
> > index 0d56f33..15a607c 100644
> > --- a/drivers/usb/gadget/configfs.c
> > +++ b/drivers/usb/gadget/configfs.c
> > @@ -97,6 +97,8 @@ struct gadget_config_name {
> >  	struct list_head list;
> >  };
> >  
> > +#define USB_MAX_STRING_WITH_NULL_LEN	(USB_MAX_STRING_LEN+1)
> > +
> >  static int usb_string_copy(const char *s, char **s_copy)
> >  {
> >  	int ret;
> > @@ -106,12 +108,16 @@ static int usb_string_copy(const char *s, char **s_copy)
> >  	if (ret > USB_MAX_STRING_LEN)
> >  		return -EOVERFLOW;
> >  
> > -	str = kstrdup(s, GFP_KERNEL);
> > -	if (!str)
> > -		return -ENOMEM;
> > +	if (copy) {
> > +		str = copy;
> > +	} else {
> > +		str = kmalloc(USB_MAX_STRING_WITH_NULL_LEN, GFP_KERNEL);
> > +		if (!str)
> > +			return -ENOMEM;
> > +	}
> > +	strcpy(str, s);
> >  	if (str[ret - 1] == '\n')
> >  		str[ret - 1] = '\0';
> > -	kfree(copy);
> >  	*s_copy = str;
> >  	return 0;
> >  }
>