net: usb: avoid using usb_control_msg() directly
Subsystems (6)
cover net: usb: avoid using usb_control_msg() directly
A recent bug-fix shaded light on possible incorrect use of usb_control_msg() without proper error check. This resulted in introducing new usb core api wrapper functions usb_control_msg_send() and usb_control_msg_recv() by Greg KH. This patch series continue the clean-up using the new functions.
Himadri Pandya (4):
net: usbnet: use usb_control_msg_recv() and usb_control_msg_send()
net: sierra_net: use usb_control_msg_recv()
net: usb: rtl8150: use usb_control_msg_recv() and
usb_control_msg_send()
net: rndis_host: use usb_control_msg_recv() and usb_control_msg_send()
drivers/net/usb/rndis_host.c | 44 +++++++++++++--------------------- drivers/net/usb/rtl8150.c | 32 +++++-------------------- drivers/net/usb/sierra_net.c | 17 ++++++------- drivers/net/usb/usbnet.c | 46 ++++-------------------------------- 4 files changed, 34 insertions(+), 105 deletions(-)
Signature · 1 line
2.17.1
PATCH 1/4 net: usbnet: use usb_control_msg_recv() and usb_control_msg_send()
Potential incorrect use of usb_control_msg() has resulted in new wrapper functions to enforce its correct usage with proper error check. Hence use these new wrapper functions instead of calling usb_control_msg() directly.
--- drivers/net/usb/usbnet.c | 46 ++++------------------------------------ 1 file changed, 4 insertions(+), 42 deletions(-)
drivers/net/usb/usbnet.c +4 -42
@@ -1982,64 +1982,26 @@ EXPORT_SYMBOL(usbnet_link_change);static int __usbnet_read_cmd (struct usbnet *dev,u8 cmd,u8 reqtype,u16 value,u16 index,void *data,u16 size) { -void *buf =NULL ; -int err = -ENOMEM ; -netdev_dbg (dev->net,"usbnet_read_cmd cmd=0x%02x reqtype=%02x" " value=0x%04x index=0x%04x size=%d\n" , cmd, reqtype, value, index, size); -if (size) { - buf =kmalloc (size,GFP_KERNEL ); -if (!buf) - goto out; - } - - err =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), - cmd, reqtype, value, index, buf, size, +return usb_control_msg_recv (dev->udev,0 , + cmd, reqtype, value, index, data, size,USB_CTRL_GET_TIMEOUT ); -if (err >0 && err <= size) { -if (data) -memcpy (data, buf, err); -else -netdev_dbg (dev->net, -"Huh? Data requested but thrown away.\n" ); - } -kfree (buf); -out: -return err; }static int __usbnet_write_cmd (struct usbnet *dev,u8 cmd,u8 reqtype,u16 value,u16 index,const void *data,u16 size) { -void *buf =NULL ; -int err = -ENOMEM ; -netdev_dbg (dev->net,"usbnet_write_cmd cmd=0x%02x reqtype=%02x" " value=0x%04x index=0x%04x size=%d\n" , cmd, reqtype, value, index, size); -if (data) { - buf =kmemdup (data, size,GFP_KERNEL ); -if (!buf) - goto out; - }else { -if (size) { -WARN_ON_ONCE (1 ); - err = -EINVAL ; - goto out; - } - } - - err =usb_control_msg (dev->udev,usb_sndctrlpipe (dev->udev,0 ), - cmd, reqtype, value, index, buf, size, +return usb_control_msg_send (dev->udev,0 , + cmd, reqtype, value, index, data, size,USB_CTRL_SET_TIMEOUT ); -kfree (buf); - -out: -return err; }/*
Signature · 1 line
2.17.1
Greg KH
maintainerRe: [PATCH 1/4] net: usbnet: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 37 lines · Himadri Pandya
Potential incorrect use of
usb_control_msg()has resulted in new wrapper functions to enforce its correct usage with proper error check. Hence use these new wrapper functions instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>--- drivers/net/usb/usbnet.c | 46 ++++------------------------------------ 1 file changed, 4 insertions(+), 42 deletions(-)drivers/net/usb/usbnet.c +2
-12@@ -1982,64 +1982,26 @@ EXPORT_SYMBOL(usbnet_link_change);static int __usbnet_read_cmd (struct usbnet *dev,u8 cmd,u8 reqtype,u16 value,u16 index,void *data,u16 size) { -void *buf =NULL ; -int err = -ENOMEM ; -netdev_dbg (dev->net,"usbnet_read_cmd cmd=0x%02x reqtype=%02x" " value=0x%04x index=0x%04x size=%d\n" , cmd, reqtype, value, index, size); -if (size) { - buf =kmalloc (size,GFP_KERNEL ); -if (!buf) - goto out; - } - - err =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), - cmd, reqtype, value, index, buf, size, +return usb_control_msg_recv (dev->udev,0 , + cmd, reqtype, value, index, data, size,USB_CTRL_GET_TIMEOUT ); -if (err >0 && err <= size) {
drivers/net/usb/usbnet.c EXPORT_SYMBOL(usbnet_link_change); · line 1992
Quoted 10 lines
drivers/net/usb/usbnet.c +0
-9-if (data) -memcpy (data, buf, err); -else -netdev_dbg (dev->net, -"Huh? Data requested but thrown away.\n" ); - } -kfree (buf); -out: -return err; }
Now there is no real need for these wrapper functions at all, except for the debugging which I doubt anyone needs anymore.
So how about just deleting these and calling the real function instead?
thanks,
greg k-h
Re: [PATCH 1/4] net: usbnet: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 53 lines · Greg KH
Quoted 37 lines · Himadri Pandya
Potential incorrect use of
usb_control_msg()has resulted in new wrapper functions to enforce its correct usage with proper error check. Hence use these new wrapper functions instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>--- drivers/net/usb/usbnet.c | 46 ++++------------------------------------ 1 file changed, 4 insertions(+), 42 deletions(-)drivers/net/usb/usbnet.c +2
-12@@ -1982,64 +1982,26 @@ EXPORT_SYMBOL(usbnet_link_change);static int __usbnet_read_cmd (struct usbnet *dev,u8 cmd,u8 reqtype,u16 value,u16 index,void *data,u16 size) { -void *buf =NULL ; -int err = -ENOMEM ; -netdev_dbg (dev->net,"usbnet_read_cmd cmd=0x%02x reqtype=%02x" " value=0x%04x index=0x%04x size=%d\n" , cmd, reqtype, value, index, size); -if (size) { - buf =kmalloc (size,GFP_KERNEL ); -if (!buf) - goto out; - } - - err =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), - cmd, reqtype, value, index, buf, size, +return usb_control_msg_recv (dev->udev,0 , + cmd, reqtype, value, index, data, size,USB_CTRL_GET_TIMEOUT ); -if (err >0 && err <= size) {drivers/net/usb/usbnet.c EXPORT_SYMBOL(usbnet_link_change); · line 1992
Quoted 10 lines
drivers/net/usb/usbnet.c +0
-9-if (data) -memcpy (data, buf, err); -else -netdev_dbg (dev->net, -"Huh? Data requested but thrown away.\n" ); - } -kfree (buf); -out: -return err; }Now there is no real need for these wrapper functions at all, except for the debugging which I doubt anyone needs anymore.
Quoted 2 lines
So how about just deleting these and calling the real function instead?
Yes, that would be a better thing to do.
Thanks,
Himadri
Quoted 3 lines
thanks,
greg k-h
PATCH 2/4 net: sierra_net: use usb_control_msg_recv()
In reply to Himadri PandyaThe new usb api function usb_control_msg_recv() nicely wrapps usb_control_msg() with proper error check. Hence use it instead of directly calling usb_control_msg().
--- drivers/net/usb/sierra_net.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-)
drivers/net/usb/sierra_net.c +7 -10
@@ -478,16 +478,13 @@ static void sierra_net_kevent(struct work_struct *work)return ; ifnum = priv->ifnum; - len =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), -USB_CDC_GET_ENCAPSULATED_RESPONSE , -USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , ifnum, buf,SIERRA_NET_USBCTL_BUF_LEN , -USB_CTRL_SET_TIMEOUT ); - -if (len <0 ) { -netdev_err (dev->net, -"usb_control_msg failed, status %d\n" , len); - }else { + len =usb_control_msg_recv (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), +USB_CDC_GET_ENCAPSULATED_RESPONSE , +USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , ifnum, buf,SIERRA_NET_USBCTL_BUF_LEN , +USB_CTRL_SET_TIMEOUT ); + +if (len) {struct hip_hdr hh; dev_dbg(&dev->udev->dev, "%s: Received status message,"
Signature · 1 line
2.17.1
PATCH 3/4 net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
In reply to Himadri PandyaMany usage of usb_control_msg() do not have proper error check on return value leaving scope for bugs on short reads. New usb_control_msg_recv() and usb_control_msg_send() nicely wraps usb_control_msg() with proper error check. Hence use the wrappers instead of calling usb_control_msg() directly.
--- drivers/net/usb/rtl8150.c | 32 ++++++-------------------------- 1 file changed, 6 insertions(+), 26 deletions(-)
drivers/net/usb/rtl8150.c +6 -26
@@ -152,36 +152,16 @@ static const char driver_name [] = "rtl8150";*/ static int get_registers (rtl8150_t * dev,u16 indx,u16 size,void *data) { -void *buf; -int ret; - - buf =kmalloc (size,GFP_NOIO ); -if (!buf) -return -ENOMEM ; - - ret =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), -RTL8150_REQ_GET_REGS ,RTL8150_REQT_READ , - indx,0 , buf, size,500 ); -if (ret >0 && ret <= size) -memcpy (data, buf, ret); -kfree (buf); -return ret; +return usb_control_msg_recv (dev->udev,0 ,RTL8150_REQ_GET_REGS , +RTL8150_REQT_READ , indx,0 , data, + size,500 ); }static int set_registers (rtl8150_t * dev,u16 indx,u16 size,const void *data) { -void *buf; -int ret; - - buf =kmemdup (data, size,GFP_NOIO ); -if (!buf) -return -ENOMEM ; - - ret =usb_control_msg (dev->udev,usb_sndctrlpipe (dev->udev,0 ), -RTL8150_REQ_SET_REGS ,RTL8150_REQT_WRITE , - indx,0 , buf, size,500 ); -kfree (buf); -return ret; +return usb_control_msg_send (dev->udev,0 ,RTL8150_REQ_SET_REGS , +RTL8150_REQT_WRITE , indx,0 , data, + size,500 ); }static void async_set_reg_cb (struct urb *urb)
Signature · 1 line
2.17.1
Re: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
Hi,
Quoted 7 lines
Many usage of
usb_control_msg()do not have proper error check on return value leaving scope for bugs on short reads. Newusb_control_msg_recv()andusb_control_msg_send()nicely wrapsusb_control_msg()with proper error check. Hence use the wrappers instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>
Quoted 9 lines
--- drivers/net/usb/rtl8150.c | 32 ++++++-------------------------- 1 file changed, 6 insertions(+), 26 deletions(-)drivers/net/usb/rtl8150.c +0
-0
drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 158
Quoted 7 lines
drivers/net/usb/rtl8150.c +0
-4@@ -152,36 +152,16 @@ static const char driver_name [] = "rtl8150";*/ static int get_registers (rtl8150_t * dev,u16 indx,u16 size,void *data) { -void *buf; -int ret; - - buf =kmalloc (size,GFP_NOIO );
GFP_NOIO is used here for a reason. You need to use this helper while in contexts of error recovery and runtime PM.
drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 157
Quoted 13 lines
drivers/net/usb/rtl8150.c +3
-10-if (!buf) -return -ENOMEM ; - - ret =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), -RTL8150_REQ_GET_REGS ,RTL8150_REQT_READ , - indx,0 , buf, size,500 ); -if (ret >0 && ret <= size) -memcpy (data, buf, ret); -kfree (buf); -return ret; +return usb_control_msg_recv (dev->udev,0 ,RTL8150_REQ_GET_REGS , +RTL8150_REQT_READ , indx,0 , data, + size,500 );
This internally uses kmemdup() with GFP_KERNEL.
You cannot make this change. The API does not support it. I am afraid we will have to change the API first, before more such changes are done.
I would suggest dropping the whole series for now.
Regards
Oliver
Re: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 18 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
Hi,
Quoted 7 lines
Many usage of
usb_control_msg()do not have proper error check on return value leaving scope for bugs on short reads. Newusb_control_msg_recv()andusb_control_msg_send()nicely wrapsusb_control_msg()with proper error check. Hence use the wrappers instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>Nacked-by Oliver Neukum <[email protected]>Quoted 4 lines
--- drivers/net/usb/rtl8150.c | 32 ++++++-------------------------- 1 file changed, 6 insertions(+), 26 deletions(-)
Quoted 16 lines
drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 158
Quoted 12 lines
drivers/net/usb/rtl8150.c +0
-4@@ -152,36 +152,16 @@ static const char driver_name [] = "rtl8150";*/ static int get_registers (rtl8150_t * dev,u16 indx,u16 size,void *data) { -void *buf; -int ret; - - buf =kmalloc (size,GFP_NOIO );
GFP_NOIOis used here for a reason. You need to use this helper while in contexts of error recovery and runtime PM.
Understood. Apologies for proposing such a stupid change.
Quoted 14 lines
drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 157
Quoted 13 lines
drivers/net/usb/rtl8150.c +3
-10-if (!buf) -return -ENOMEM ; - - ret =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), -RTL8150_REQ_GET_REGS ,RTL8150_REQT_READ , - indx,0 , buf, size,500 ); -if (ret >0 && ret <= size) -memcpy (data, buf, ret); -kfree (buf); -return ret; +return usb_control_msg_recv (dev->udev,0 ,RTL8150_REQ_GET_REGS , +RTL8150_REQT_READ , indx,0 , data, + size,500 );
Quoted 6 lines
This internally uses
kmemdup()withGFP_KERNEL.
You cannot make this change. The API does not support it. I am afraid we will have to change the API first, before more such changes are done.I would suggest dropping the whole series for now.
Okay. Thanks for the review.
Regards,
Himadri
Quoted 4 lines
Regards
Oliver
Oliver Neukum
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Am Mittwoch, den 23.09.2020, 19:36 +0530 schrieb Himadri Pandya:
Quoted 9 lines · Oliver Neukum
Quoted 6 lines
Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
GFP_NOIOis used here for a reason. You need to use this helper while in contexts of error recovery and runtime PM.Understood. Apologies for proposing such a stupid change.
Hi,
sorry if you concluded that the patch was stupid. That was not my intent. It was the best the API allowed for. If an API makes it easy to make a mistake, the problem is with the API, not the developer.
Regards
Oliver
Re: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 12 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 19:36 +0530 schrieb Himadri Pandya:
Quoted 9 lines · Oliver Neukum
Quoted 6 lines
Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
GFP_NOIOis used here for a reason. You need to use this helper while in contexts of error recovery and runtime PM.Understood. Apologies for proposing such a stupid change.
Quoted 9 lines
Hi,
sorry if you concluded that the patch was stupid. That was not my intent. It was the best the API allowed for. If an API makes it easy to make a mistake, the problem is with the API, not the developer.
Regards
Oliver
I meant that it was stupid to change it without properly understanding the significance of GFP_NOIO in this context.
So now, do we re-write the wrapper functions with flag passed as a parameter?
Regards,
Himadri
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees
Oliver Neukum
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Am Mittwoch, den 23.09.2020, 20:02 +0530 schrieb Himadri Pandya:
Quoted 4 lines
I meant that it was stupid to change it without properly understanding the significance of
GFP_NOIOin this context.So now, do we re-write the wrapper functions with flag passed as a parameter?
Hi,
I hope I set you in CC for a patch set doing exactly that.
Do not let me or other maintainers discourage you from writing patches. Look at it this way. Had you not written this patch, I would not have looked into the matter. Patches are supposed to be reviewed. If you want additional information, just ask. We do not want people discouraged from writing substantial patches.
Regards
Oliver
Re: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 12 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 20:02 +0530 schrieb Himadri Pandya:
Quoted 4 lines
I meant that it was stupid to change it without properly understanding the significance of
GFP_NOIOin this context.So now, do we re-write the wrapper functions with flag passed as a parameter?
Hi,
I hope I set you in CC for a patch set doing exactly that.
Yes.
Quoted 6 lines
Do not let me or other maintainers discourage you from writing patches. Look at it this way. Had you not written this patch, I would not have looked into the matter. Patches are supposed to be reviewed. If you want additional information, just ask. We do not want people discouraged from writing substantial patches.
Understood :).
I'll send v2 after the update in API is merged.
Thanks,
Himadri
Quoted 4 lines
Regards
Oliver
Petko Manolov
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
In reply to Oliver NeukumQuoted 43 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
Hi,
Quoted 7 lines
Many usage of
usb_control_msg()do not have proper error check on return value leaving scope for bugs on short reads. Newusb_control_msg_recv()andusb_control_msg_send()nicely wrapsusb_control_msg()with proper error check. Hence use the wrappers instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>Nacked-by Oliver Neukum <[email protected]>Quoted 9 lines
--- drivers/net/usb/rtl8150.c | 32 ++++++-------------------------- 1 file changed, 6 insertions(+), 26 deletions(-)drivers/net/usb/rtl8150.c +0
-0drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 158
Quoted 7 lines
drivers/net/usb/rtl8150.c +0
-4@@ -152,36 +152,16 @@ static const char driver_name [] = "rtl8150";*/ static int get_registers (rtl8150_t * dev,u16 indx,u16 size,void *data) { -void *buf; -int ret; - - buf =kmalloc (size,GFP_NOIO );
GFP_NOIOis used here for a reason. You need to use this helper while in contexts of error recovery and runtime PM.Quoted 10 lines
drivers/net/usb/rtl8150.c +0
-10-if (!buf) -return -ENOMEM ; - - ret =usb_control_msg (dev->udev,usb_rcvctrlpipe (dev->udev,0 ), -RTL8150_REQ_GET_REGS ,RTL8150_REQT_READ , - indx,0 , buf, size,500 ); -if (ret >0 && ret <= size) -memcpy (data, buf, ret); -kfree (buf); -return ret;
Quoted 8 lines
drivers/net/usb/rtl8150.c static const char driver_name [] = "rtl8150"; · line 157
Quoted 3 lines
drivers/net/usb/rtl8150.c +3
-0+return usb_control_msg_recv (dev->udev,0 ,RTL8150_REQ_GET_REGS , +RTL8150_REQT_READ , indx,0 , data, + size,500 );This internally uses
kmemdup()withGFP_KERNEL.
You cannot make this change. The API does not support it. I am afraid we will have to change the API first, before more such changes are done.
One possible fix is to add yet another argument to usb_control_msg_recv(), which would be the GFP_XYZ flag to pass on to kmemdup(). Up to Greg, of course.
cheers,
Petko
Oliver Neukum
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Am Mittwoch, den 23.09.2020, 17:48 +0300 schrieb Petko Manolov:
Quoted 7 lines
Quoted 4 lines
This internally uses
kmemdup()withGFP_KERNEL.
You cannot make this change. The API does not support it. I am afraid we will have to change the API first, before more such changes are done.One possible fix is to add yet another argument to
usb_control_msg_recv(), which would be theGFP_XYZflag to pass on tokmemdup(). Up to Greg, of course.
Hi,
submitted. The problem is those usages that are very hard to trace. I'd dislike to just slab GFP_NOIO on them for no obvious reason.
Regards
Oliver
Petko Manolov
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 8 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 17:48 +0300 schrieb Petko Manolov:
Quoted 3 lines
One possible fix is to add yet another argument to
usb_control_msg_recv(), which would be theGFP_XYZflag to pass on tokmemdup(). Up to Greg, of course.submitted. The problem is those usages that are very hard to trace. I'd dislike to just slab
GFP_NOIOon them for no obvious reason.
Do you mean you submitted a patch for usb_control_msg_recv() (because i don't see it on linux-netdev) or i'm reading this all wrong?
Petko
Greg KH
maintainerRe: [PATCH 3/4] net: usb: rtl8150: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 7 lines · Petko Manolov
Quoted 6 lines · Oliver Neukum
Am Mittwoch, den 23.09.2020, 17:48 +0300 schrieb Petko Manolov:
Quoted 3 lines
One possible fix is to add yet another argument to
usb_control_msg_recv(), which would be theGFP_XYZflag to pass on tokmemdup(). Up to Greg, of course.
Quoted 5 lines
Quoted 2 lines
submitted. The problem is those usages that are very hard to trace. I'd dislike to just slab
GFP_NOIOon them for no obvious reason.Do you mean you submitted a patch for
usb_control_msg_recv()(because i don't see it on linux-netdev) or i'm reading this all wrong?
It's on the linux-usb list:
https://lore.kernel.org/r/[email protected]
PATCH 4/4 net: rndis_host: use usb_control_msg_recv() and usb_control_msg_send()
In reply to Himadri PandyaThe new usb_control_msg_recv() and usb_control_msg_send() nicely wraps usb_control_msg() with proper error check. Hence use the wrappers instead of calling usb_control_msg() directly.
--- drivers/net/usb/rndis_host.c | 44 ++++++++++++++---------------------- 1 file changed, 17 insertions(+), 27 deletions(-)
drivers/net/usb/rndis_host.c +17 -27
@@ -113,14 +113,13 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) buf->request_id = (__force __le32 ) xid; } master_ifnum = info->control->cur_altsetting->desc.bInterfaceNumber; - retval =usb_control_msg (dev->udev, -usb_sndctrlpipe (dev->udev,0 ), -USB_CDC_SEND_ENCAPSULATED_COMMAND , -USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf,le32_to_cpu (buf->msg_len), -RNDIS_CONTROL_TIMEOUT_MS ); -if (unlikely (retval <0 || xid ==0 )) + retval =usb_control_msg_send (dev->udev,0 , +USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, +le32_to_cpu (buf->msg_len), +RNDIS_CONTROL_TIMEOUT_MS ); +if (unlikely (xid ==0 ))return retval;/* Some devices don't respond on the control channel until
@@ -139,14 +138,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen)/* Poll the control channel; the request probably completed immediately */ rsp =le32_to_cpu (buf->msg_type) |RNDIS_MSG_COMPLETION ;for (count =0 ; count <10 ; count++) { -memset (buf,0 ,CONTROL_BUFFER_SIZE ); - retval =usb_control_msg (dev->udev, -usb_rcvctrlpipe (dev->udev,0 ), -USB_CDC_GET_ENCAPSULATED_RESPONSE , -USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf, buflen, -RNDIS_CONTROL_TIMEOUT_MS ); + retval =usb_control_msg_recv (dev->udev,0 , +USB_CDC_GET_ENCAPSULATED_RESPONSE , +USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, buflen, +RNDIS_CONTROL_TIMEOUT_MS );if (likely (retval >=8 )) { msg_type =le32_to_cpu (buf->msg_type); msg_len =le32_to_cpu (buf->msg_len);
@@ -178,17 +174,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) msg->msg_type =cpu_to_le32 (RNDIS_MSG_KEEPALIVE_C ); msg->msg_len =cpu_to_le32 (sizeof *msg); msg->status =cpu_to_le32 (RNDIS_STATUS_SUCCESS ); - retval =usb_control_msg (dev->udev, -usb_sndctrlpipe (dev->udev,0 ), -USB_CDC_SEND_ENCAPSULATED_COMMAND , -USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - msg,sizeof *msg, -RNDIS_CONTROL_TIMEOUT_MS ); -if (unlikely (retval <0 )) -dev_dbg (&info->control->dev, -"rndis keepalive err %d\n" , - retval); + retval =usb_control_msg_send (dev->udev,0 , +USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, msg,sizeof (*msg), +RNDIS_CONTROL_TIMEOUT_MS ); }break ;default :
Signature · 1 line
2.17.1
Greg KH
maintainerRe: [PATCH 4/4] net: rndis_host: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 43 lines · Himadri Pandya
The new
usb_control_msg_recv()andusb_control_msg_send()nicely wrapsusb_control_msg()with proper error check. Hence use the wrappers instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>--- drivers/net/usb/rndis_host.c | 44 ++++++++++++++---------------------- 1 file changed, 17 insertions(+), 27 deletions(-)drivers/net/usb/rndis_host.c +7
-12@@ -113,14 +113,13 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) buf->request_id = (__force __le32 ) xid; } master_ifnum = info->control->cur_altsetting->desc.bInterfaceNumber; - retval =usb_control_msg (dev->udev, -usb_sndctrlpipe (dev->udev,0 ), -USB_CDC_SEND_ENCAPSULATED_COMMAND , -USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf,le32_to_cpu (buf->msg_len), -RNDIS_CONTROL_TIMEOUT_MS ); -if (unlikely (retval <0 || xid ==0 )) + retval =usb_control_msg_send (dev->udev,0 , +USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, +le32_to_cpu (buf->msg_len), +RNDIS_CONTROL_TIMEOUT_MS ); +if (unlikely (xid ==0 ))return retval;/* Some devices don't respond on the control channel until @@ -139,14 +138,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen)/* Poll the control channel; the request probably completed immediately */ rsp =le32_to_cpu (buf->msg_type) |RNDIS_MSG_COMPLETION ;for (count =0 ; count <10 ; count++) { -memset (buf,0 ,CONTROL_BUFFER_SIZE ); - retval = usb_control_msg(dev->udev, - usb_rcvctrlpipe(dev->udev, 0), - USB_CDC_GET_ENCAPSULATED_RESPONSE,
drivers/net/usb/rndis_host.c int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) · line 146
Quoted 10 lines
drivers/net/usb/rndis_host.c +5
-4-USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf, buflen, - RNDIS_CONTROL_TIMEOUT_MS); + retval =usb_control_msg_recv (dev->udev,0 , +USB_CDC_GET_ENCAPSULATED_RESPONSE , +USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, buflen, +RNDIS_CONTROL_TIMEOUT_MS );if (likely (retval >=8 )) {
retval here is never going to be positive, right? So I don't think this patch is correct :(
Quoted 12 lines
drivers/net/usb/rndis_host.c +0
-6msg_type =le32_to_cpu (buf->msg_type); msg_len =le32_to_cpu (buf->msg_len);@@ -178,17 +174,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) msg->msg_type =cpu_to_le32 (RNDIS_MSG_KEEPALIVE_C ); msg->msg_len =cpu_to_le32 (sizeof *msg); msg->status =cpu_to_le32 (RNDIS_STATUS_SUCCESS ); - retval = usb_control_msg(dev->udev, - usb_sndctrlpipe(dev->udev, 0), - USB_CDC_SEND_ENCAPSULATED_COMMAND, - USB_TYPE_CLASS | USB_RECIP_INTERFACE, - 0, master_ifnum, - msg, sizeof *msg,
drivers/net/usb/rndis_host.c int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) · line 181
Quoted 10 lines
drivers/net/usb/rndis_host.c +5
-5- RNDIS_CONTROL_TIMEOUT_MS); -if (unlikely (retval <0 )) -dev_dbg (&info->control->dev, -"rndis keepalive err %d\n" , - retval); + retval =usb_control_msg_send (dev->udev,0 , +USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, msg,sizeof (*msg), +RNDIS_CONTROL_TIMEOUT_MS );
You lost the error message that the previous call had if something went wrong. Don't know if it's really needed, but there's no reason to remove it here.
thanks,
greg k-h
Re: [PATCH 4/4] net: rndis_host: use usb_control_msg_recv() and usb_control_msg_send()
Quoted 50 lines · Greg KH
Quoted 48 lines · Himadri Pandya
The new
usb_control_msg_recv()andusb_control_msg_send()nicely wrapsusb_control_msg()with proper error check. Hence use the wrappers instead of callingusb_control_msg()directly.Signed-off-by Himadri Pandya <[email protected]>--- drivers/net/usb/rndis_host.c | 44 ++++++++++++++---------------------- 1 file changed, 17 insertions(+), 27 deletions(-)drivers/net/usb/rndis_host.c +8
-16@@ -113,14 +113,13 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) buf->request_id = (__force __le32 ) xid; } master_ifnum = info->control->cur_altsetting->desc.bInterfaceNumber; - retval =usb_control_msg (dev->udev, -usb_sndctrlpipe (dev->udev,0 ), -USB_CDC_SEND_ENCAPSULATED_COMMAND , -USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf,le32_to_cpu (buf->msg_len), -RNDIS_CONTROL_TIMEOUT_MS ); -if (unlikely (retval <0 || xid ==0 )) + retval =usb_control_msg_send (dev->udev,0 , +USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, +le32_to_cpu (buf->msg_len), +RNDIS_CONTROL_TIMEOUT_MS ); +if (unlikely (xid ==0 ))return retval;/* Some devices don't respond on the control channel until @@ -139,14 +138,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen)/* Poll the control channel; the request probably completed immediately */ rsp =le32_to_cpu (buf->msg_type) |RNDIS_MSG_COMPLETION ;for (count =0 ; count <10 ; count++) { -memset (buf,0 ,CONTROL_BUFFER_SIZE ); - retval =usb_control_msg (dev->udev, -usb_rcvctrlpipe (dev->udev,0 ), -USB_CDC_GET_ENCAPSULATED_RESPONSE , -USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - buf, buflen, -RNDIS_CONTROL_TIMEOUT_MS ); + retval = usb_control_msg_recv(dev->udev, 0,
Quoted 9 lines
drivers/net/usb/rndis_host.c int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) · line 146
Quoted 5 lines
drivers/net/usb/rndis_host.c +4
-0+USB_CDC_GET_ENCAPSULATED_RESPONSE , +USB_DIR_IN |USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, buf, buflen, + RNDIS_CONTROL_TIMEOUT_MS);if (likely (retval >=8 )) {retval here is never going to be positive, right? So I don't think this patch is correct :(
Yes :(.
Quoted 18 lines
Quoted 18 lines
drivers/net/usb/rndis_host.c +1
-11msg_type =le32_to_cpu (buf->msg_type); msg_len =le32_to_cpu (buf->msg_len);@@ -178,17 +174,11 @@ int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) msg->msg_type =cpu_to_le32 (RNDIS_MSG_KEEPALIVE_C ); msg->msg_len =cpu_to_le32 (sizeof *msg); msg->status =cpu_to_le32 (RNDIS_STATUS_SUCCESS ); - retval =usb_control_msg (dev->udev, -usb_sndctrlpipe (dev->udev,0 ), -USB_CDC_SEND_ENCAPSULATED_COMMAND , -USB_TYPE_CLASS |USB_RECIP_INTERFACE , -0 , master_ifnum, - msg,sizeof *msg, -RNDIS_CONTROL_TIMEOUT_MS ); -if (unlikely (retval <0 )) -dev_dbg (&info->control->dev, -"rndis keepalive err %d\n" , - retval); + retval = usb_control_msg_send(dev->udev, 0,
Quoted 9 lines
drivers/net/usb/rndis_host.c int rndis_command(struct usbnet *dev, struct rndis_msg_hdr *buf, int buflen) · line 181
Quoted 4 lines
drivers/net/usb/rndis_host.c +4
-0+USB_CDC_SEND_ENCAPSULATED_COMMAND , +USB_TYPE_CLASS |USB_RECIP_INTERFACE , +0 , master_ifnum, msg,sizeof (*msg), + RNDIS_CONTROL_TIMEOUT_MS);You lost the error message that the previous call had if something went wrong. Don't know if it's really needed, but there's no reason to remove it here.
The wrapper returns the error so thought that might work instead. But yes, the old msg is better.
Anyways, this series is dropped :).
Thanks for the review,
Himadri
Quoted 3 lines
thanks,
greg k-h
Greg KH
maintainerQuoted 13 lines · Himadri Pandya
A recent bug-fix shaded light on possible incorrect use of
usb_control_msg()without proper error check. This resulted in introducing new usb core api wrapper functionsusb_control_msg_send()andusb_control_msg_recv()by Greg KH. This patch series continue the clean-up using the new functions.Himadri Pandya (4):
net: usbnet: useusb_control_msg_recv()andusb_control_msg_send()
net:sierra_net: useusb_control_msg_recv()
net: usb: rtl8150: useusb_control_msg_recv()and
usb_control_msg_send()
net:rndis_host: useusb_control_msg_recv()andusb_control_msg_send()
Quoted 6 lines
drivers/net/usb/rndis_host.c | 44 +++++++++++++--------------------- drivers/net/usb/rtl8150.c | 32 +++++-------------------- drivers/net/usb/sierra_net.c | 17 ++++++------- drivers/net/usb/usbnet.c | 46 ++++-------------------------------- 4 files changed, 34 insertions(+), 105 deletions(-)
Note, all of these depend on a set of patches in my usb tree, so they should probably wait to be sent to the networking developers until after 5.10-rc1 is out.
thanks,
greg k-h
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees