Skip to contentSkip to conversation

staging: r8188eu: shorten and simplify calls chains

PATCH Superseded Fabio M. De Francesco 3 participants 39 messages Sep 19, 2021 – Sep 21, 2021 Staging

v8 of 10 Superseded Latest is v10 →

lore
Share & tools
lore
#0 lore raw

v8 cover staging: r8188eu: shorten and simplify calls chains

--- Preface ---

This is v8 of "shorten and simplify calls chain". The first 14 patches have already been applied to staging-testing, so we have been requested to reset the numbering of the remaining patches to 01/19, while discarding from this new submission the above-mentioned 14 patches (otherwise we would have submitted a series containing 33 patches).

The following commit message is provided as it was in v7, both for the purpose of presenting the whole picture to Maintainers, Reviewers, and to anybody else who may be interested in knowing the entire design and the evolution since v1 to the current v8.

--- Commit message ---

io_ops abstraction is useless in this driver, since there is only one ops registration. Without io_ops we can get rid of indirect calls mess and shorten the calls chain.

Shorten the calls chain of rtw_read8/16/32() down to the actual reads. For this purpose unify the three usb_read8/16/32 into the new usb_read(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_read() and use in it the new usb_control_msg_recv() API of USB Core.

Shorten the calls chain of rtw_write8/16/32() down to the actual writes. For this purpose unify the four usb_write8/16/32/N() into the new usb_write(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_write() and use in it the new usb_control_msg_send() API of USB Core.

The code with the modifications was thoroughly tested by Pavel Skripkin using a TP-Link TL-WN722N v2 / v3 [Realtek RTL8188EUS] and by Fabio M. De Francesco using a ASUSTek Computer, Inc. Realtek 8188EUS [USB-N10 Nano].

--- Changelog ---

v7->v8 (old numbering):
- 1-14:
Patches applied to staging-testing, so they are dropped from the current v8;

- 15-19:
Split into 19 patches. Numbering reset to 01. After this reset, 15-19/19 become 01-19/19 (so we have a total of 33 patches in this series.

v7->v8 (new numbering for the old 15-19/19):
- 1-15:
Old 15/19 and 16/19 are split into 15 patches (as it is required by Greg Kroah-Hartman) in order to make them more easily reviewable and for checking that the logic is not broken; no significant changes to the resulting code;
- 16-17:
Old 17/19 and 18/19 become 16/19 and 17/19: There are no significant changes to the code, with the sole exception of moving the acquiring of a mutex before the test for "bSurpriseRemoved" being 'true';
- 18-19:
Old 19/19 is split into new 18/19 and 19/19. The changes are split into a first patch that remove a shared buffer and a second that remove the mutex protecting the receiving (in usb_read()) and sending (in usb_write()) of the USB control messages.

v6->v7:
- 1-14:
Add Dan Carpenter's "Reviewed-by" tag. No code changes;
- 15:
Add a list of clean-up changes to the commit messages and restore an empty line inadvertently deleted in a previous version;
- 16:
Add a list of clean-up changes to the commit messages;
- 17-18:
No changes;
- 19:
Delete an unnecessary test.

v5->v6:
- 1-14:
Minimal changes to patch subjects to achieve consistent style;
- 15:
Fix a mistake in the checks of values returned by usb_control_msg();
- 16-19:
No changes.

v4->v5:
- 1-14:
No changes;
- 15-16:
Fix minor formatting issues and use "Reverse Xmas Tree" style, according to suggestions by David Laight and Dan Carpenter;
- 17-18:
Pavel Skripkin found logical errors in the checks of the values returned by usb_control_msg_{recv,send}(), so fix them; Dan Carpenter suggested to do error handling before success handling, so change the code accordingly;
- 19:
Add this patch in order to get rid of the shared buffer in usb_read() and usb_write() and remove this field from struct "dvobj_priv".

v3->v4:
- 1-14:
Split a patch into fourteen;
- 15-16:
Add these patches for clean-ups of the code that is going to be reused in 17-18/18;
- 17-18:
Make some changes according to a first review of Greg Kroah-Hartman; furthermore, remove the unnecessary while loop and a couple of if' test; handle the errors returned by usb_control_msg_recv().

v2->v3:
- 1-2:
No changes;
- 3:
Fix the version number of the patch.

v1->v2:
- 1-2:
No changes;
- 3:
Replace parameter REALTEK_USB_VENQT_READ with REALTEK_USB_VENQT_WRITE in usb_control_msg_send().

--- List of links to previous versions of the cover letters ---

v1: https://lore.kernel.org/lkml/[email protected]/
v2: https://lore.kernel.org/lkml/[email protected]/
v3: https://lore.kernel.org/lkml/[email protected]/
v4: https://lore.kernel.org/lkml/[email protected]/
v5: https://lore.kernel.org/lkml/[email protected]/
v6: https://lore.kernel.org/lkml/[email protected]/
v7: https://lore.kernel.org/lkml/[email protected]/

Fabio M. De Francesco (17):
staging: r8188eu: clean up symbols usbctrl_vendorreq()
staging: r8188eu: reorder declarations in usbctrl_vendorreq()
staging: r8188eu: remove unnecessary test in usbctrl_vendorreq()
staging: r8188eu: reorder comments in usbctrl_vendorreq()
staging: r8188eu: remove unnedeed parentheses in usbctrl_vendorreq()
staging: r8188eu: remove unnecessary space in usbctrl_vendorreq()
staging: r8188eu: remove unnecessary comment in usbctrl_vendorreq()
staging: r8188eu: fix grammar mistake in usbctrl_vendorreq()
staging: r8188eu: remove unnecessary braces in usbctrl_vendorreq()
staging: r8188eu: rename symbols in rtw_read*() and rtw_write*()
staging: r8188eu: remove unnecessary casts from rtw_{read,write}*()
staging: r8188eu: change the type of a variable in rtw_write16()
staging: r8188eu: remove an unneeded buffer from rtw_writeN()
staging: r8188eu: remove an unnecessary bit AND from rtw_writeN()
staging: r8188eu: change the type of a variable in rtw_read16()
staging: r8188eu: call the new usb_read() from rtw_read{8,16,32}()
staging: r8188eu: call the new usb_write() from rtw_write{8,16,32,N}()

Pavel Skripkin (2):
staging: r8188eu: remove shared buffer for usb requests
staging: r8188eu: remove usb_vendor_req_mutex

 drivers/staging/r8188eu/hal/usb_ops_linux.c | 213 ++++++++++----------
 drivers/staging/r8188eu/include/drv_types.h |   3 -
 drivers/staging/r8188eu/os_dep/usb_intf.c   |  34 +---
 3 files changed, 109 insertions(+), 141 deletions(-)
Signature · 1 line
2.33.0
#1 lore raw

PATCH v8 1/19 staging: r8188eu: clean up symbols usbctrl_vendorreq()

Clean up symbol names in usbctrl_vendorreq():

	pdata => data;
        pio_priv => io_priv;
        pintfhdl => intfhdl.
---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 28 ++++++++++-----------
 1 file changed, 14 insertions(+), 14 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +14 -14
index 117213c9f984..3bfae4bd4a1e 100644
@@ -8,14 +8,14 @@
 #include "../include/recv_osdep.h"
 #include "../include/rtl8188e_hal.h"
 
-static int usbctrl_vendorreq(struct intf_hdl *pintfhdl, u16 value, void *pdata, u16 len, u8 requesttype)
+static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u16 len, u8 requesttype)
 {
-	struct adapter	*adapt = pintfhdl->padapter;
-	struct dvobj_priv  *dvobjpriv = adapter_to_dvobj(adapt);
+	struct adapter *adapt = intfhdl->padapter;
+	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
 	struct usb_device *udev = dvobjpriv->pusbdev;
 	unsigned int pipe;
 	int status = 0;
-	u8 *pIo_buf;
+	u8 *io_buf;
 	int vendorreq_times = 0;
 
 	if ((adapt->bSurpriseRemoved) || (adapt->pwrctrlpriv.pnp_bstop_trx)) {
@@ -32,10 +32,10 @@ static int usbctrl_vendorreq(struct intf_hdl *pintfhdl, u16 value, void *pdata,
 	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
 
 	/*  Acquire IO memory for vendorreq */
-	pIo_buf = dvobjpriv->usb_vendor_req_buf;
+	io_buf = dvobjpriv->usb_vendor_req_buf;
 
-	if (!pIo_buf) {
-		DBG_88E("[%s] pIo_buf == NULL\n", __func__);
+	if (!io_buf) {
+		DBG_88E("[%s] io_buf == NULL\n", __func__);
 		status = -ENOMEM;
 		goto release_mutex;
 	}
@@ -47,22 +47,22 @@ static int usbctrl_vendorreq(struct intf_hdl *pintfhdl, u16 value, void *pdata,
 
 	while (++vendorreq_times <= MAX_USBCTRL_VENDORREQ_TIMES) {
 		if (requesttype == REALTEK_USB_VENQT_READ)
-			memset(pIo_buf, 0, len);
+			memset(io_buf, 0, len);
 		else
-			memcpy(pIo_buf, pdata, len);
+			memcpy(io_buf, data, len);
 
 		status = usb_control_msg(udev, pipe, REALTEK_USB_VENQT_CMD_REQ,
 					 requesttype, value, REALTEK_USB_VENQT_CMD_IDX,
-					 pIo_buf, len, RTW_USB_CONTROL_MSG_TIMEOUT);
+					 io_buf, len, RTW_USB_CONTROL_MSG_TIMEOUT);
 
 		if (status == len) {   /*  Success this control transfer. */
 			rtw_reset_continual_urb_error(dvobjpriv);
 			if (requesttype == REALTEK_USB_VENQT_READ)
-				memcpy(pdata, pIo_buf,  len);
+				memcpy(data, io_buf,  len);
 		} else { /*  error cases */
 			DBG_88E("reg 0x%x, usb %s %u fail, status:%d value=0x%x, vendorreq_times:%d\n",
 				value, (requesttype == REALTEK_USB_VENQT_READ) ? "read" : "write",
-				len, status, *(u32 *)pdata, vendorreq_times);
+				len, status, *(u32 *)data, vendorreq_times);
 
 			if (status < 0) {
 				if (status == (-ESHUTDOWN) || status == -ENODEV) {
@@ -74,8 +74,8 @@ static int usbctrl_vendorreq(struct intf_hdl *pintfhdl, u16 value, void *pdata,
 			} else { /*  status != len && status >= 0 */
 				if (status > 0) {
 					if (requesttype == REALTEK_USB_VENQT_READ) {
-						/*  For Control read transfer, we have to copy the read data from pIo_buf to pdata. */
-						memcpy(pdata, pIo_buf,  len);
+						/*  For Control read transfer, we have to copy the read data from io_buf to data. */
+						memcpy(data, io_buf,  len);
 					}
 				}
 			}
Signature · 1 line
2.33.0
#2 lore raw

Re: [PATCH v8 01/19] staging: r8188eu: clean up symbols usbctrl_vendorreq()

Quoted 5 lines · Fabio M. De Francesco

Clean up symbol names in usbctrl_vendorreq():

	pdata => data;
        pio_priv => io_priv;
        pintfhdl => intfhdl.

Odd mix of tabs and spaces :(

Anyway, what does intfhdl mean?

Ugh, that's a horrible structure name for a driver, but that's not your fault. However, isn't 'intf' a better name for this thing as that is how the driver does use it in other structures?

This isn't windows, where "handles" are used all over the places. We have real structures :)

thanks,

greg k-h

#3 lore raw

PATCH v8 2/19 staging: r8188eu: reorder declarations in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Reorder variables declarations according to the "Reverse Xmas Tree" style.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index 3bfae4bd4a1e..e06aea92c8d2 100644
@@ -13,10 +13,10 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 	struct adapter *adapt = intfhdl->padapter;
 	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
 	struct usb_device *udev = dvobjpriv->pusbdev;
+	int vendorreq_times = 0;
 	unsigned int pipe;
 	int status = 0;
 	u8 *io_buf;
-	int vendorreq_times = 0;
 
 	if ((adapt->bSurpriseRemoved) || (adapt->pwrctrlpriv.pnp_bstop_trx)) {
 		status = -EPERM;
Signature · 1 line
2.33.0
#4 lore raw

PATCH v8 3/19 staging: r8188eu: remove unnecessary test in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Remove unnecessary test for "!io_buf" in usbctrl_vendorreq(). Remove the no more used "release_mutex:" label.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -7
index e06aea92c8d2..d92bdcc3716d 100644
@@ -34,12 +34,6 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 	/*  Acquire IO memory for vendorreq */
 	io_buf = dvobjpriv->usb_vendor_req_buf;
 
-	if (!io_buf) {
-		DBG_88E("[%s] io_buf == NULL\n", __func__);
-		status = -ENOMEM;
-		goto release_mutex;
-	}
-
 	if (requesttype == REALTEK_USB_VENQT_READ)
 		pipe = usb_rcvctrlpipe(udev, 0);/* read_in */
 	else
@@ -91,7 +85,7 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 		if ((value >= FW_8188E_START_ADDRESS && value <= FW_8188E_END_ADDRESS) || status == len)
 			break;
 	}
-release_mutex:
+
 	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
 exit:
 	return status;
Signature · 1 line
2.33.0
#5 lore raw

Re: [PATCH v8 03/19] staging: r8188eu: remove unnecessary test in usbctrl_vendorreq()

Quoted 2 lines · Fabio M. De Francesco

Remove unnecessary test for "!io_buf" in usbctrl_vendorreq(). Remove the no more used "release_mutex:" label.

You need to explain _why_ it is unnecessary.

thanks,

greg k-h

#6 lore raw

PATCH v8 4/19 staging: r8188eu: reorder comments in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Reorder comments in usbctrl_vendorreq() to follow the Linux style.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +10 -5
index d92bdcc3716d..9138b730490f 100644
@@ -35,9 +35,11 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 	io_buf = dvobjpriv->usb_vendor_req_buf;
 
 	if (requesttype == REALTEK_USB_VENQT_READ)
-		pipe = usb_rcvctrlpipe(udev, 0);/* read_in */
+		/* read in */
+		pipe = usb_rcvctrlpipe(udev, 0);
 	else
-		pipe = usb_sndctrlpipe(udev, 0);/* write_out */
+		/* write out */
+		pipe = usb_sndctrlpipe(udev, 0);
 
 	while (++vendorreq_times <= MAX_USBCTRL_VENDORREQ_TIMES) {
 		if (requesttype == REALTEK_USB_VENQT_READ)
@@ -49,11 +51,13 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 					 requesttype, value, REALTEK_USB_VENQT_CMD_IDX,
 					 io_buf, len, RTW_USB_CONTROL_MSG_TIMEOUT);
 
-		if (status == len) {   /*  Success this control transfer. */
+		if (status == len) {
+			/*  success */
 			rtw_reset_continual_urb_error(dvobjpriv);
 			if (requesttype == REALTEK_USB_VENQT_READ)
 				memcpy(data, io_buf,  len);
-		} else { /*  error cases */
+		} else {
+			/* error cases */
 			DBG_88E("reg 0x%x, usb %s %u fail, status:%d value=0x%x, vendorreq_times:%d\n",
 				value, (requesttype == REALTEK_USB_VENQT_READ) ? "read" : "write",
 				len, status, *(u32 *)data, vendorreq_times);
@@ -65,7 +69,8 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 					struct hal_data_8188e	*haldata = GET_HAL_DATA(adapt);
 					haldata->srestpriv.wifi_error_status = USB_VEN_REQ_CMD_FAIL;
 				}
-			} else { /*  status != len && status >= 0 */
+			} else {
+				/* status != len && status >= 0 */
 				if (status > 0) {
 					if (requesttype == REALTEK_USB_VENQT_READ) {
 						/*  For Control read transfer, we have to copy the read data from io_buf to data. */
Signature · 1 line
2.33.0
#7 lore raw

Re: [PATCH v8 04/19] staging: r8188eu: reorder comments in usbctrl_vendorreq()

Quoted 14 lines · Fabio M. De Francesco

Reorder comments in usbctrl_vendorreq() to follow the Linux style.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -0
index d92bdcc3716d..9138b730490f 100644

drivers/staging/r8188eu/hal/usb_ops_linux.c static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1 · line 42

Quoted 10 lines
drivers/staging/r8188eu/hal/usb_ops_linux.c +4 -2
@@ -35,9 +35,11 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 	io_buf = dvobjpriv->usb_vendor_req_buf;
 
 	if (requesttype == REALTEK_USB_VENQT_READ)
-		pipe = usb_rcvctrlpipe(udev, 0);/* read_in */
+		/* read in */
+		pipe = usb_rcvctrlpipe(udev, 0);
 	else
-		pipe = usb_sndctrlpipe(udev, 0);/* write_out */
+		/* write out */
+		pipe = usb_sndctrlpipe(udev, 0);

For single-line if statements, that's not really needed. Just drop the comments entirely, the code here is "obvious".

thanks,

greg k-h

#8 lore raw

PATCH v8 5/19 staging: r8188eu: remove unnedeed parentheses in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Remove unneeded parentheses around a test for -ESHUTDOWN in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index 9138b730490f..0b48dde5657f 100644
@@ -63,7 +63,7 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 				len, status, *(u32 *)data, vendorreq_times);
 
 			if (status < 0) {
-				if (status == (-ESHUTDOWN) || status == -ENODEV) {
+				if (status == -ESHUTDOWN || status == -ENODEV) {
 					adapt->bSurpriseRemoved = true;
 				} else {
 					struct hal_data_8188e	*haldata = GET_HAL_DATA(adapt);
Signature · 1 line
2.33.0
#9 lore raw

PATCH v8 6/19 staging: r8188eu: remove unnecessary space in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Remove unnecessary extra space in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index 0b48dde5657f..a52aeb2558ad 100644
@@ -66,7 +66,7 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 				if (status == -ESHUTDOWN || status == -ENODEV) {
 					adapt->bSurpriseRemoved = true;
 				} else {
-					struct hal_data_8188e	*haldata = GET_HAL_DATA(adapt);
+					struct hal_data_8188e *haldata = GET_HAL_DATA(adapt);
 					haldata->srestpriv.wifi_error_status = USB_VEN_REQ_CMD_FAIL;
 				}
 			} else {
Signature · 1 line
2.33.0
#10 lore raw

PATCH v8 7/19 staging: r8188eu: remove unnecessary comment in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Remove an unnecessary comment in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 1 -
 1 file changed, 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -1
index a52aeb2558ad..fc3da0fbf474 100644
@@ -73,7 +73,6 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 				/* status != len && status >= 0 */
 				if (status > 0) {
 					if (requesttype == REALTEK_USB_VENQT_READ) {
-						/*  For Control read transfer, we have to copy the read data from io_buf to data. */
 						memcpy(data, io_buf,  len);
 					}
 				}
Signature · 1 line
2.33.0
#11 lore raw

Re: [PATCH v8 07/19] staging: r8188eu: remove unnecessary comment in usbctrl_vendorreq()

Quoted 14 lines · Fabio M. De Francesco

Remove an unnecessary comment in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 1 -
 1 file changed, 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -0
index a52aeb2558ad..fc3da0fbf474 100644

drivers/staging/r8188eu/hal/usb_ops_linux.c static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1 · line 77

Quoted 6 lines
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -1
@@ -73,7 +73,6 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 				/* status != len && status >= 0 */
 				if (status > 0) {
 					if (requesttype == REALTEK_USB_VENQT_READ) {
-						/*  For Control read transfer, we have to copy the read data from io_buf to data. */
 						memcpy(data, io_buf,  len);
 					}

And now you have a coding style issue :(

Don't try to fix up one and add another one please.

thanks,

greg k-h

#12 lore raw

PATCH v8 8/19 staging: r8188eu: fix grammar mistake in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Fix a grammar mistake in usbctrl_vendorreq(): "checksumed" => "checksummed".

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index fc3da0fbf474..3ca2959f4bcd 100644
@@ -85,7 +85,7 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 
 		}
 
-		/*  firmware download is checksumed, don't retry */
+		/*  firmware download is checksummed, don't retry */
 		if ((value >= FW_8188E_START_ADDRESS && value <= FW_8188E_END_ADDRESS) || status == len)
 			break;
 	}
Signature · 1 line
2.33.0
#13 lore raw

Re: [PATCH v8 08/19] staging: r8188eu: fix grammar mistake in usbctrl_vendorreq()

Quoted 15 lines · Fabio M. De Francesco

Fix a grammar mistake in usbctrl_vendorreq(): "checksumed" => "checksummed".

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -0
index fc3da0fbf474..3ca2959f4bcd 100644

drivers/staging/r8188eu/hal/usb_ops_linux.c static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1 · line 88

Quoted 5 lines
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
@@ -85,7 +85,7 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 
 		}
 
-		/*  firmware download is checksumed, don't retry */
+		/*  firmware download is checksummed, don't retry */

When touching this line, also remove the leading ' ' characters in the comment that are not needed please.

thanks,

greg k-h

#14 lore raw

PATCH v8 9/19 staging: r8188eu: remove unnecessary braces in usbctrl_vendorreq()

In reply to Fabio M. De Francesco

Remove unnecessary braces in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -2
index 3ca2959f4bcd..a270cb4249b5 100644
@@ -72,9 +72,8 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 			} else {
 				/* status != len && status >= 0 */
 				if (status > 0) {
-					if (requesttype == REALTEK_USB_VENQT_READ) {
+					if (requesttype == REALTEK_USB_VENQT_READ)
 						memcpy(data, io_buf,  len);
-					}
 				}
 			}
 
Signature · 1 line
2.33.0
#15 lore raw

Re: [PATCH v8 09/19] staging: r8188eu: remove unnecessary braces in usbctrl_vendorreq()

Quoted 14 lines · Fabio M. De Francesco

Remove unnecessary braces in usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +0 -0
index 3ca2959f4bcd..a270cb4249b5 100644

drivers/staging/r8188eu/hal/usb_ops_linux.c static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1 · line 77

Quoted 7 lines
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -2
@@ -72,9 +72,8 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 			} else {
 				/* status != len && status >= 0 */
 				if (status > 0) {
-					if (requesttype == REALTEK_USB_VENQT_READ) {
+					if (requesttype == REALTEK_USB_VENQT_READ)
 						memcpy(data, io_buf,  len);
-					}

This should have gone in the previous commit.

#16 lore raw

PATCH v8 10/19 staging: r8188eu: rename symbols in rtw_read*() and rtw_write*()

In reply to Fabio M. De Francesco

Rename names of variables in rtw_read{8.16.32}() and in rtw_write{8,16,32,N}() because of unnecessary 'p' (that probably stand for "pointer to") and 'w' (that probably stands for "word"):

        pio_priv => io_priv;
        pintfhdl => intfhdl;
        wvalue => value.
---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 60 ++++++++++-----------
 1 file changed, 30 insertions(+), 30 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +30 -30
index a270cb4249b5..efdf654e434a 100644
@@ -96,91 +96,91 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 
 u8 rtw_read8(struct adapter *adapter, u32 addr)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	u8 data;
 
-	usbctrl_vendorreq(pintfhdl, wvalue, &data, 1, REALTEK_USB_VENQT_READ);
+	usbctrl_vendorreq(intfhdl, value, &data, 1, REALTEK_USB_VENQT_READ);
 
 	return data;
 }
 
 u16 rtw_read16(struct adapter *adapter, u32 addr)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	__le32 data;
 
-	usbctrl_vendorreq(pintfhdl, wvalue, &data, 2, REALTEK_USB_VENQT_READ);
+	usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_READ);
 
 	return (u16)(le32_to_cpu(data) & 0xffff);
 }
 
 u32 rtw_read32(struct adapter *adapter, u32 addr)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	__le32 data;
 
-	usbctrl_vendorreq(pintfhdl, wvalue, &data, 4, REALTEK_USB_VENQT_READ);
+	usbctrl_vendorreq(intfhdl, value, &data, 4, REALTEK_USB_VENQT_READ);
 
 	return le32_to_cpu(data);
 }
 
 int rtw_write8(struct adapter *adapter, u32 addr, u8 val)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	int ret;
 
-	ret = usbctrl_vendorreq(pintfhdl, wvalue, &val, 1, REALTEK_USB_VENQT_WRITE);
+	ret = usbctrl_vendorreq(intfhdl, value, &val, 1, REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
 
 int rtw_write16(struct adapter *adapter, u32 addr, u16 val)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	__le32 data = cpu_to_le32(val & 0x0000ffff);
 	int ret;
 
-	ret = usbctrl_vendorreq(pintfhdl, wvalue, &data, 2, REALTEK_USB_VENQT_WRITE);
+	ret = usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
 
 int rtw_write32(struct adapter *adapter, u32 addr, u32 val)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	__le32 data = cpu_to_le32(val);
 	int ret;
 
-	ret = usbctrl_vendorreq(pintfhdl, wvalue, &data, 4, REALTEK_USB_VENQT_WRITE);
+	ret = usbctrl_vendorreq(intfhdl, value, &data, 4, REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
 
-int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *pdata)
+int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *data)
 {
-	struct io_priv *pio_priv = &adapter->iopriv;
-	struct intf_hdl *pintfhdl = &pio_priv->intf;
-	u16 wvalue = (u16)(addr & 0x0000ffff);
+	struct io_priv *io_priv = &adapter->iopriv;
+	struct intf_hdl *intfhdl = &io_priv->intf;
+	u16 value = (u16)(addr & 0x0000ffff);
 	u8 buf[VENDOR_CMD_MAX_DATA_LEN] = {0};
 	int ret;
 
 	if (length > VENDOR_CMD_MAX_DATA_LEN)
 		return _FAIL;
 
-	memcpy(buf, pdata, length);
-	ret = usbctrl_vendorreq(pintfhdl, wvalue, buf, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
+	memcpy(buf, data, length);
+	ret = usbctrl_vendorreq(intfhdl, value, buf, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
Signature · 1 line
2.33.0
#17 lore raw

Re: [PATCH v8 10/19] staging: r8188eu: rename symbols in rtw_read*() and rtw_write*()

Quoted 7 lines · Fabio M. De Francesco

Rename names of variables in rtw_read{8.16.32}() and in rtw_write{8,16,32,N}() because of unnecessary 'p' (that probably stand for "pointer to") and 'w' (that probably stands for "word"):

        pio_priv => io_priv;
        pintfhdl => intfhdl;
        wvalue => value.

All spaces, nice :)

Anyway, same 'intfhdl' issue here.

thanks,

greg k-h

#18 lore raw

PATCH v8 11/19 staging: r8188eu: remove unnecessary casts from rtw_{read,write}*()

In reply to Fabio M. De Francesco

Remove unnecessary casts from rtw_read{8,16,32}() and from rtw_write{8,16,32,N}().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +7 -7
index efdf654e434a..ba654db869f7 100644
@@ -98,7 +98,7 @@ u8 rtw_read8(struct adapter *adapter, u32 addr)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	u8 data;
 
 	usbctrl_vendorreq(intfhdl, value, &data, 1, REALTEK_USB_VENQT_READ);
@@ -110,7 +110,7 @@ u16 rtw_read16(struct adapter *adapter, u32 addr)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	__le32 data;
 
 	usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_READ);
@@ -122,7 +122,7 @@ u32 rtw_read32(struct adapter *adapter, u32 addr)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	__le32 data;
 
 	usbctrl_vendorreq(intfhdl, value, &data, 4, REALTEK_USB_VENQT_READ);
@@ -134,7 +134,7 @@ int rtw_write8(struct adapter *adapter, u32 addr, u8 val)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	int ret;
 
 	ret = usbctrl_vendorreq(intfhdl, value, &val, 1, REALTEK_USB_VENQT_WRITE);
@@ -146,7 +146,7 @@ int rtw_write16(struct adapter *adapter, u32 addr, u16 val)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	__le32 data = cpu_to_le32(val & 0x0000ffff);
 	int ret;
 
@@ -159,7 +159,7 @@ int rtw_write32(struct adapter *adapter, u32 addr, u32 val)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	__le32 data = cpu_to_le32(val);
 	int ret;
 
@@ -172,7 +172,7 @@ int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *data)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
-	u16 value = (u16)(addr & 0x0000ffff);
+	u16 value = addr & 0xffff;
 	u8 buf[VENDOR_CMD_MAX_DATA_LEN] = {0};
 	int ret;
 
Signature · 1 line
2.33.0
#19 lore raw

PATCH v8 12/19 staging: r8188eu: change the type of a variable in rtw_write16()

In reply to Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index ba654db869f7..b2a5b6ffbb46 100644
@@ -147,7 +147,7 @@ int rtw_write16(struct adapter *adapter, u32 addr, u16 val)
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
 	u16 value = addr & 0xffff;
-	__le32 data = cpu_to_le32(val & 0x0000ffff);
+	__le16 data = cpu_to_le16(val);
 	int ret;
 
 	ret = usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_WRITE);
Signature · 1 line
2.33.0
#20 lore raw

Re: [PATCH v8 12/19] staging: r8188eu: change the type of a variable in rtw_write16()

Quoted 1 line · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You are saying _what_ you are doing here, but it is not obvious at all _why_ you need to do this.

Please explain the why here when you redo this series.

thanks,

greg k-h

#21 lore raw

PATCH v8 13/19 staging: r8188eu: remove an unneeded buffer from rtw_writeN()

In reply to Fabio M. De Francesco

Remove the unnecessary buffer "u8 buf[VENDOR_CMD_MAX_DATA_LEN]" and pass directly "data" to usbctrl_vendorreq().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -3
index b2a5b6ffbb46..aa57bd7b8f10 100644
@@ -173,14 +173,12 @@ int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *data)
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
 	u16 value = addr & 0xffff;
-	u8 buf[VENDOR_CMD_MAX_DATA_LEN] = {0};
 	int ret;
 
 	if (length > VENDOR_CMD_MAX_DATA_LEN)
 		return _FAIL;
 
-	memcpy(buf, data, length);
-	ret = usbctrl_vendorreq(intfhdl, value, buf, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
+	ret = usbctrl_vendorreq(intfhdl, value, data, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
Signature · 1 line
2.33.0
#22 lore raw

PATCH v8 14/19 staging: r8188eu: remove an unnecessary bit AND from rtw_writeN()

In reply to Fabio M. De Francesco

Remove an unnecessary bitwise AND because "length" can never be greater than 0xffff since VENDOR_CMD_MAX_DATA_LEN is defined as (a weird) '254'.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +1 -1
index aa57bd7b8f10..625b29af9410 100644
@@ -178,7 +178,7 @@ int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *data)
 	if (length > VENDOR_CMD_MAX_DATA_LEN)
 		return _FAIL;
 
-	ret = usbctrl_vendorreq(intfhdl, value, data, (length & 0xffff), REALTEK_USB_VENQT_WRITE);
+	ret = usbctrl_vendorreq(intfhdl, value, data, length, REALTEK_USB_VENQT_WRITE);
 
 	return RTW_STATUS_CODE(ret);
 }
Signature · 1 line
2.33.0
#23 lore raw

PATCH v8 15/19 staging: r8188eu: change the type of a variable in rtw_read16()

In reply to Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +2 -2
index 625b29af9410..c378b5740353 100644
@@ -111,11 +111,11 @@ u16 rtw_read16(struct adapter *adapter, u32 addr)
 	struct io_priv *io_priv = &adapter->iopriv;
 	struct intf_hdl *intfhdl = &io_priv->intf;
 	u16 value = addr & 0xffff;
-	__le32 data;
+	__le16 data;
 
 	usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_READ);
 
-	return (u16)(le32_to_cpu(data) & 0xffff);
+	return le16_to_cpu(data);
 }
 
 u32 rtw_read32(struct adapter *adapter, u32 addr)
Signature · 1 line
2.33.0
#24 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 1 line · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

Again, why?

#25 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

In reply to Fabio M. De Francesco
Quoted 2 lines · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You should note in the commit message that:

The last two bytes of "data" are not initialized so the le32_to_cpu(data) technically reads uninitialized data. This can likely be detected by the KASan checker as reading uninitialized data. But because the bytes are discarded in the end so this will not affect runtime.

regards,
dan carpenter

#26 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 6 lines · CEST Dan Carpenter
Quoted 2 lines · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You should note in the commit message that:

Quoted 8 lines

The last two bytes of "data" are not initialized so the le32_to_cpu(data) technically reads uninitialized data. This can likely be detected by the KASan checker as reading uninitialized data. But because the bytes are discarded in the end so this will not affect runtime.

regards,
dan carpenter

Dear Dan,

Thanks for your suggestion about this specific topic.

We thought that, since "data" is in bitwise AND with 0xffff before being passed to the callee, it was enough to have reviewers know why we're doing that change of type with no further explanations. Actually it seems to be not enough to motivate that change.

We will surely use the note you provided.

However, since I'm not used to blindly follow suggestions (even if I trust your words with no doubts at all) without complete understanding of what I'm doing, I will need to understand what KASan is before copy-paste your note.

Thank you very much,

Fabio

#27 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 25 lines · Fabio M. De Francesco
Quoted 6 lines · CEST Dan Carpenter
Quoted 2 lines · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You should note in the commit message that:

Quoted 8 lines

The last two bytes of "data" are not initialized so the le32_to_cpu(data) technically reads uninitialized data. This can likely be detected by the KASan checker as reading uninitialized data. But because the bytes are discarded in the end so this will not affect runtime.

regards,
dan carpenter

Dear Dan,

Thanks for your suggestion about this specific topic.

We thought that, since "data" is in bitwise AND with 0xffff before being passed to the callee, it was enough to have reviewers know why we're doing that change of type with no further explanations. Actually it seems to be not enough to motivate that change.

Quoted 5 lines

We will surely use the note you provided.

However, since I'm not used to blindly follow suggestions (even if I trust your words with no doubts at all) without complete understanding of what I'm doing, I will need to understand what KASan is before copy-paste your note.

Google is your friend!

Either way reading uninitialized data is generally bad. The trickier thing is showing that your changes don't affect runtime. For both of these le32 to le16 changes.

regards,
dan carpenter

#28 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 9 lines · CEST Dan Carpenter
Quoted 8 lines · Fabio M. De Francesco
Quoted 7 lines · CEST Dan Carpenter
Quoted 2 lines · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You should note in the commit message that:

The last two bytes of "data" are not initialized so the

le32_to_cpu(data)

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

technically reads uninitialized data. This can likely be detected by the KASan checker as reading uninitialized data. But because the bytes are discarded in the end so this will not affect runtime.

regards,
dan carpenter

Quoted 6 lines
Quoted 6 lines

Dear Dan,

Thanks for your suggestion about this specific topic.

We thought that, since "data" is in bitwise AND with 0xffff before being passed to the callee, it was enough to have reviewers know why we're

doing

Quoted 1 line
Quoted 1 line

that change of type with no further explanations. Actually it seems to be

not

Quoted 5 lines
Quoted 5 lines

enough to motivate that change.

We will surely use the note you provided.

However, since I'm not used to blindly follow suggestions (even if I

trust

Quoted 1 line
Quoted 1 line

your words with no doubts at all) without complete understanding of what

I'm

Quoted 1 line
Quoted 1 line

doing, I will need to understand what KASan is before copy-paste your

note.

Quoted 2 lines

Google is your friend!

Yes, it is :)

I think you were referring to the KernelMemorySanitizer (KMSan), a detector of uses of uninitialized memory (but it seems to not be upstream):
https://github.com/google/kmsan

Instead you wrote about the The Kernel Address Sanitizer (KASan) that seems to be a dynamic memory error detector designed to find out-of-bound and use- after-free bugs (this is upstream):
https://www.kernel.org/doc/html/v5.0/dev-tools/kasan.html

Can you please confirm?

Back to the code... uninitialised data is not a problem in the old code, it's just bad design. The new code cannot affect runtime, it's just better design.

There's no change in runtime behaviour because of different protection nets:
Aside from the bitwise AND that truncate that variable two the size of two bytes and set the higher bytes to 0, memcpy() inside usbctrl_vendorreq(), the new usb_read() and usb_write uses memcpy() with count = size (and size is checked also in rtw_writeN()).

I can't see any bugs. Just bad design, that we fix and possible sanitizer's warning, that disappear with our fixes. Am I right?

Thanks,

Fabio

Quoted 9 lines

Either way reading uninitialized data is generally bad. The trickier thing is showing that your changes don't affect runtime. For both of these le32 to le16 changes.

regards,
dan carpenter

#29 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 47 lines · Fabio M. De Francesco
Quoted 9 lines · CEST Dan Carpenter
Quoted 8 lines · Fabio M. De Francesco
Quoted 7 lines · CEST Dan Carpenter
Quoted 2 lines · Fabio M. De Francesco

Change the type of "data" from __le32 to __le16.

You should note in the commit message that:

The last two bytes of "data" are not initialized so the

le32_to_cpu(data)

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

technically reads uninitialized data. This can likely be detected by the KASan checker as reading uninitialized data. But because the bytes are discarded in the end so this will not affect runtime.

regards,
dan carpenter

Quoted 6 lines
Quoted 6 lines

Dear Dan,

Thanks for your suggestion about this specific topic.

We thought that, since "data" is in bitwise AND with 0xffff before being passed to the callee, it was enough to have reviewers know why we're

doing

Quoted 1 line
Quoted 1 line

that change of type with no further explanations. Actually it seems to be

not

Quoted 5 lines
Quoted 5 lines

enough to motivate that change.

We will surely use the note you provided.

However, since I'm not used to blindly follow suggestions (even if I

trust

Quoted 1 line
Quoted 1 line

your words with no doubts at all) without complete understanding of what

I'm

Quoted 1 line
Quoted 1 line

doing, I will need to understand what KASan is before copy-paste your

note.

Quoted 2 lines

Google is your friend!

Yes, it is :)

I think you were referring to the KernelMemorySanitizer (KMSan), a detector of uses of uninitialized memory (but it seems to not be upstream):
https://github.com/google/kmsan

Quoted 6 lines

Instead you wrote about the The Kernel Address Sanitizer (KASan) that seems to be a dynamic memory error detector designed to find out-of-bound and use- after-free bugs (this is upstream):
https://www.kernel.org/doc/html/v5.0/dev-tools/kasan.html

Can you please confirm?

That sounds probably right.

Quoted 3 lines

Back to the code... uninitialised data is not a problem in the old code, it's just bad design. The new code cannot affect runtime, it's just better design.

It would be a problem for KMSan and the kbuild-bot will email you about it. From your commit message "Change the type of "data" from __le32 to __le16." it's not clear you understand why the kbuild-bot will email you and why it's correct to do so.

regards,
dan carpenter

#30 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 4 lines · CEST Dan Carpenter
Quoted 3 lines · Fabio M. De Francesco

On Monday, September 20, 2021 3:10:36 PM CEST Dan Carpenter wrote:

Back to the code... uninitialised data is not a problem in the old code,

it's

Quoted 1 line
Quoted 1 line

just bad design. The new code cannot affect runtime, it's just better

design.

Quoted 5 lines

It would be a problem for KMSan and the kbuild-bot will email you about it. From your commit message "Change the type of "data" from __le32 to __le16." it's not clear you understand why the kbuild-bot will email you and why it's correct to do so.

I guess that for sure you are right in saying that it is not entirely clear why the kbuild-bot will email me. However, since the change is correct, but the commit message is not, I confirm that we'll use your note with only a slight change (only if you agree): replace KASan with KMSan in your text.

Thanks,

Fabio

Quoted 2 lines

regards,
dan carpenter

#31 lore raw

Re: [PATCH v8 15/19] staging: r8188eu: change the type of a variable in rtw_read16()

Quoted 14 lines · Fabio M. De Francesco
Quoted 4 lines · CEST Dan Carpenter
Quoted 3 lines · Fabio M. De Francesco

On Monday, September 20, 2021 3:10:36 PM CEST Dan Carpenter wrote:

Back to the code... uninitialised data is not a problem in the old code,

it's

Quoted 1 line
Quoted 1 line

just bad design. The new code cannot affect runtime, it's just better

design.

Quoted 5 lines

It would be a problem for KMSan and the kbuild-bot will email you about it. From your commit message "Change the type of "data" from __le32 to __le16." it's not clear you understand why the kbuild-bot will email you and why it's correct to do so.

Quoted 5 lines

I guess that for sure you are right in saying that it is not entirely clear why the kbuild-bot will email me. However, since the change is correct, but the commit message is not, I confirm that we'll use your note with only a slight change (only if you agree): replace KASan with KMSan in your text.

Perfect. Thanks.

regards,
dan carpenter

#32 lore raw

PATCH v8 16/19 staging: r8188eu: call the new usb_read() from rtw_read{8,16,32}()

In reply to Fabio M. De Francesco

Create and call new usb_read() instead of usbctrl_vendorreq() from inside rtw_read{8,16,32}().

In old code, rtw_read{8,16,32}() called usbctrl_vendorreq() which in turn uses usb_control_msg() from within a "while" loop to build a control URB, send it off and wait for completion. usbctrl_vendorreq() was used for both receiving and sending messages, depending on the "requesttype" argument which was passed by callers.

Compared to usbctrl_vendorreq(), which managed both reads and writes from and to the USB endpoint, the new usb_read() manages only reads. For this purpose it uses the newer USB Core usb_control_msg_recv() API. The latter is preferred according both to a suggestion by Greg Kroah-Hartman and also to its actual design.

Two noteworthy features of usb_control_msg_recv() are that (1) the data pointer can be made to a reference on the stack because it does not have the restriction that usb_control_msg() has where the data pointer must be to dynamically allocated memory, and that (2) the whole message must be properly received from the device in order for this function to be successfuli (if a device returns less than the expected amount of data, then the function will fail).

usbctrl_vendorreq() uses a "while" loop that we considered unnecessary so that it is not in the new usb_read(). Furthermore, the latter has no redundant checking, less obvious comments, and it manages errors before success cases. All in all, usb_read() is simpler than usbctrl_vendorreq() and uses less lines of code.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 59 +++++++++++++++++++--
 1 file changed, 56 insertions(+), 3 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +56 -3
index c378b5740353..53704b7c1059 100644
@@ -94,6 +94,59 @@ static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u1
 	return status;
 }
 
+static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
+{
+	struct adapter *adapt = intfhdl->padapter;
+	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
+	struct usb_device *udev = dvobjpriv->pusbdev;
+	int status;
+	u8 *io_buf; /* Pointer to I/O buffer */
+
+	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
+
+	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
+		return -EPERM;
+
+	io_buf = dvobjpriv->usb_vendor_req_buf;
+
+	status = usb_control_msg_recv(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
+				      REALTEK_USB_VENQT_READ, value,
+				      REALTEK_USB_VENQT_CMD_IDX, io_buf,
+				      size, RTW_USB_CONTROL_MSG_TIMEOUT,
+				      GFP_KERNEL);
+
+	if (status == -ESHUTDOWN ||
+	    status == -ENODEV ||
+	    status == -ENOENT) {
+		/*
+		 * device or controller has been disabled due to
+		 * some problem that could not be worked around,
+		 * device or bus doesn’t exist, endpoint does not
+		 * exist or is not enabled.
+		 */
+		adapt->bSurpriseRemoved = true;
+		goto mutex_unlock;
+	}
+
+	if (status < 0) {
+		GET_HAL_DATA(adapt)->srestpriv.wifi_error_status =
+			USB_VEN_REQ_CMD_FAIL;
+
+		if (rtw_inc_and_chk_continual_urb_error(dvobjpriv))
+			adapt->bSurpriseRemoved = true;
+
+		goto mutex_unlock;
+	}
+
+	rtw_reset_continual_urb_error(dvobjpriv);
+	memcpy(data, io_buf, size);
+
+mutex_unlock:
+	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
+
+	return status;
+}
+
 u8 rtw_read8(struct adapter *adapter, u32 addr)
 {
 	struct io_priv *io_priv = &adapter->iopriv;
@@ -101,7 +154,7 @@ u8 rtw_read8(struct adapter *adapter, u32 addr)
 	u16 value = addr & 0xffff;
 	u8 data;
 
-	usbctrl_vendorreq(intfhdl, value, &data, 1, REALTEK_USB_VENQT_READ);
+	usb_read(intfhdl, value, &data, 1);
 
 	return data;
 }
@@ -113,7 +166,7 @@ u16 rtw_read16(struct adapter *adapter, u32 addr)
 	u16 value = addr & 0xffff;
 	__le16 data;
 
-	usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_READ);
+	usb_read(intfhdl, value, &data, 2);
 
 	return le16_to_cpu(data);
 }
@@ -125,7 +178,7 @@ u32 rtw_read32(struct adapter *adapter, u32 addr)
 	u16 value = addr & 0xffff;
 	__le32 data;
 
-	usbctrl_vendorreq(intfhdl, value, &data, 4, REALTEK_USB_VENQT_READ);
+	usb_read(intfhdl, value, &data, 4);
 
 	return le32_to_cpu(data);
 }
Signature · 1 line
2.33.0
#33 lore raw

PATCH v8 17/19 staging: r8188eu: call the new usb_write() from rtw_write{8,16,32,N}()

In reply to Fabio M. De Francesco

Create and call new usb_write() instead of usbctrl_vendorreq() from inside rtw_write{8,16,32,N}().

In old code, rtw_write{8,16,32,N}() called usbctrl_vendorreq() which in turn uses usb_control_msg() from within a "while" loop to build a control URB, send it off and wait for completion. usbctrl_vendorreq() was used for both receiving and sending messages, depending on the "requesttype" argument which is passed by callers.

Compared to usbctrl_vendorreq(), which manages both reads and writes from and to the USB endpoint, the new usb_write() manages only writes. For this purpose it uses the newer USB Core usb_control_msg_send() API. The latter is preferred according both to suggestions by Greg Kroah-Hartman and also to its actual design.

A noteworthy feature of usb_control_msg_send() is that the data pointer can be made to a reference on the stack because it does not have the restriction that usb_control_msg() has where the data pointer must be to dynamically allocated memory.

usbctrl_vendorreq() used a "while" loop that we considered unnecessary so that it is not in the new usb_write(). Furthermore, the latter has no redundant checking, less obvious comments, no debug prints, and it manages errors before success case. All in all, usb_write() is simpler than usbctrl_vendorreq() and uses less lines of code.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 117 +++++++-------------
 1 file changed, 42 insertions(+), 75 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +42 -75
index 53704b7c1059..2552450ab152 100644
@@ -8,93 +8,60 @@
 #include "../include/recv_osdep.h"
 #include "../include/rtl8188e_hal.h"
 
-static int usbctrl_vendorreq(struct intf_hdl *intfhdl, u16 value, void *data, u16 len, u8 requesttype)
+static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 {
 	struct adapter *adapt = intfhdl->padapter;
 	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
 	struct usb_device *udev = dvobjpriv->pusbdev;
-	int vendorreq_times = 0;
-	unsigned int pipe;
-	int status = 0;
-	u8 *io_buf;
-
-	if ((adapt->bSurpriseRemoved) || (adapt->pwrctrlpriv.pnp_bstop_trx)) {
-		status = -EPERM;
-		goto exit;
-	}
-
-	if (len > MAX_VENDOR_REQ_CMD_SIZE) {
-		DBG_88E("[%s] Buffer len error ,vendor request failed\n", __func__);
-		status = -EINVAL;
-		goto exit;
-	}
+	int status;
+	u8 *io_buf; /* Pointer to I/O buffer */
 
 	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
 
-	/*  Acquire IO memory for vendorreq */
-	io_buf = dvobjpriv->usb_vendor_req_buf;
-
-	if (requesttype == REALTEK_USB_VENQT_READ)
-		/* read in */
-		pipe = usb_rcvctrlpipe(udev, 0);
-	else
-		/* write out */
-		pipe = usb_sndctrlpipe(udev, 0);
+	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
+		return -EPERM;
 
-	while (++vendorreq_times <= MAX_USBCTRL_VENDORREQ_TIMES) {
-		if (requesttype == REALTEK_USB_VENQT_READ)
-			memset(io_buf, 0, len);
-		else
-			memcpy(io_buf, data, len);
+	io_buf = dvobjpriv->usb_vendor_req_buf;
 
-		status = usb_control_msg(udev, pipe, REALTEK_USB_VENQT_CMD_REQ,
-					 requesttype, value, REALTEK_USB_VENQT_CMD_IDX,
-					 io_buf, len, RTW_USB_CONTROL_MSG_TIMEOUT);
+	status = usb_control_msg_recv(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
+				      REALTEK_USB_VENQT_READ, value,
+				      REALTEK_USB_VENQT_CMD_IDX, io_buf,
+				      size, RTW_USB_CONTROL_MSG_TIMEOUT,
+				      GFP_KERNEL);
 
-		if (status == len) {
-			/*  success */
-			rtw_reset_continual_urb_error(dvobjpriv);
-			if (requesttype == REALTEK_USB_VENQT_READ)
-				memcpy(data, io_buf,  len);
-		} else {
-			/* error cases */
-			DBG_88E("reg 0x%x, usb %s %u fail, status:%d value=0x%x, vendorreq_times:%d\n",
-				value, (requesttype == REALTEK_USB_VENQT_READ) ? "read" : "write",
-				len, status, *(u32 *)data, vendorreq_times);
-
-			if (status < 0) {
-				if (status == -ESHUTDOWN || status == -ENODEV) {
-					adapt->bSurpriseRemoved = true;
-				} else {
-					struct hal_data_8188e *haldata = GET_HAL_DATA(adapt);
-					haldata->srestpriv.wifi_error_status = USB_VEN_REQ_CMD_FAIL;
-				}
-			} else {
-				/* status != len && status >= 0 */
-				if (status > 0) {
-					if (requesttype == REALTEK_USB_VENQT_READ)
-						memcpy(data, io_buf,  len);
-				}
-			}
+	if (status == -ESHUTDOWN ||
+	    status == -ENODEV ||
+	    status == -ENOENT) {
+		/*
+		 * device or controller has been disabled due to
+		 * some problem that could not be worked around,
+		 * device or bus doesn’t exist, endpoint does not
+		 * exist or is not enabled.
+		 */
+		adapt->bSurpriseRemoved = true;
+		goto mutex_unlock;
+	}
 
-			if (rtw_inc_and_chk_continual_urb_error(dvobjpriv)) {
-				adapt->bSurpriseRemoved = true;
-				break;
-			}
+	if (status < 0) {
+		GET_HAL_DATA(adapt)->srestpriv.wifi_error_status =
+			USB_VEN_REQ_CMD_FAIL;
 
-		}
+		if (rtw_inc_and_chk_continual_urb_error(dvobjpriv))
+			adapt->bSurpriseRemoved = true;
 
-		/*  firmware download is checksummed, don't retry */
-		if ((value >= FW_8188E_START_ADDRESS && value <= FW_8188E_END_ADDRESS) || status == len)
-			break;
+		goto mutex_unlock;
 	}
 
+	rtw_reset_continual_urb_error(dvobjpriv);
+	memcpy(data, io_buf, size);
+
+mutex_unlock:
 	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
-exit:
+
 	return status;
 }
 
-static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
+static int usb_write(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 {
 	struct adapter *adapt = intfhdl->padapter;
 	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
@@ -109,8 +76,9 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 
 	io_buf = dvobjpriv->usb_vendor_req_buf;
 
-	status = usb_control_msg_recv(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
-				      REALTEK_USB_VENQT_READ, value,
+	memcpy(io_buf, data, size);
+	status = usb_control_msg_send(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
+				      REALTEK_USB_VENQT_WRITE, value,
 				      REALTEK_USB_VENQT_CMD_IDX, io_buf,
 				      size, RTW_USB_CONTROL_MSG_TIMEOUT,
 				      GFP_KERNEL);
@@ -139,7 +107,6 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 	}
 
 	rtw_reset_continual_urb_error(dvobjpriv);
-	memcpy(data, io_buf, size);
 
 mutex_unlock:
 	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
@@ -190,7 +157,7 @@ int rtw_write8(struct adapter *adapter, u32 addr, u8 val)
 	u16 value = addr & 0xffff;
 	int ret;
 
-	ret = usbctrl_vendorreq(intfhdl, value, &val, 1, REALTEK_USB_VENQT_WRITE);
+	ret = usb_write(intfhdl, value, &val, 1);
 
 	return RTW_STATUS_CODE(ret);
 }
@@ -203,7 +170,7 @@ int rtw_write16(struct adapter *adapter, u32 addr, u16 val)
 	__le16 data = cpu_to_le16(val);
 	int ret;
 
-	ret = usbctrl_vendorreq(intfhdl, value, &data, 2, REALTEK_USB_VENQT_WRITE);
+	ret = usb_write(intfhdl, value, &data, 2);
 
 	return RTW_STATUS_CODE(ret);
 }
@@ -216,7 +183,7 @@ int rtw_write32(struct adapter *adapter, u32 addr, u32 val)
 	__le32 data = cpu_to_le32(val);
 	int ret;
 
-	ret = usbctrl_vendorreq(intfhdl, value, &data, 4, REALTEK_USB_VENQT_WRITE);
+	ret = usb_write(intfhdl, value, &data, 4);
 
 	return RTW_STATUS_CODE(ret);
 }
@@ -231,7 +198,7 @@ int rtw_writeN(struct adapter *adapter, u32 addr, u32 length, u8 *data)
 	if (length > VENDOR_CMD_MAX_DATA_LEN)
 		return _FAIL;
 
-	ret = usbctrl_vendorreq(intfhdl, value, data, length, REALTEK_USB_VENQT_WRITE);
+	ret = usb_write(intfhdl, value, data, length);
 
 	return RTW_STATUS_CODE(ret);
 }
Signature · 1 line
2.33.0
#34 lore raw

PATCH v8 18/19 staging: r8188eu: remove shared buffer for USB requests

In reply to Fabio M. De Francesco

From: Pavel Skripkin <[email protected]>

Remove the shared buffer for USB requests because it is not necessary and use a local on stack variable since the new Core USB API does not have the constraints of usb_control_msg().

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c |  8 ++------
 drivers/staging/r8188eu/include/drv_types.h |  3 ---
 drivers/staging/r8188eu/os_dep/usb_intf.c   | 14 +-------------
 3 files changed, 3 insertions(+), 22 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +2 -6
index 2552450ab152..75500c28d6e6 100644
@@ -14,15 +14,13 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
 	struct usb_device *udev = dvobjpriv->pusbdev;
 	int status;
-	u8 *io_buf; /* Pointer to I/O buffer */
+	u8 io_buf[4];
 
 	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
 
 	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
 		return -EPERM;
 
-	io_buf = dvobjpriv->usb_vendor_req_buf;
-
 	status = usb_control_msg_recv(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
 				      REALTEK_USB_VENQT_READ, value,
 				      REALTEK_USB_VENQT_CMD_IDX, io_buf,
@@ -67,15 +65,13 @@ static int usb_write(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 	struct dvobj_priv *dvobjpriv = adapter_to_dvobj(adapt);
 	struct usb_device *udev = dvobjpriv->pusbdev;
 	int status;
-	u8 *io_buf; /* Pointer to I/O buffer */
+	u8 io_buf[VENDOR_CMD_MAX_DATA_LEN];
 
 	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
 
 	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
 		return -EPERM;
 
-	io_buf = dvobjpriv->usb_vendor_req_buf;
-
 	memcpy(io_buf, data, size);
 	status = usb_control_msg_send(udev, 0, REALTEK_USB_VENQT_CMD_REQ,
 				      REALTEK_USB_VENQT_WRITE, value,
drivers/staging/r8188eu/include/drv_types.h +0 -3
index 626c6273be6f..7907dd7b5bf0 100644
@@ -170,9 +170,6 @@ struct dvobj_priv {
 	struct semaphore usb_suspend_sema;
 	struct mutex  usb_vendor_req_mutex;
 
-	u8 *usb_alloc_vendor_req_buf;
-	u8 *usb_vendor_req_buf;
-
 	struct usb_interface *pusbintf;
 	struct usb_device *pusbdev;
 
drivers/staging/r8188eu/os_dep/usb_intf.c +1 -13
index 11a584cbe9d8..5ab42d55207f 100644
@@ -75,24 +75,12 @@ static struct rtw_usb_drv *usb_drv = &rtl8188e_usb_drv;
 
 static u8 rtw_init_intf_priv(struct dvobj_priv *dvobj)
 {
-	u8 rst = _SUCCESS;
-
 	mutex_init(&dvobj->usb_vendor_req_mutex);
-
-	dvobj->usb_alloc_vendor_req_buf = kzalloc(MAX_USB_IO_CTL_SIZE, GFP_KERNEL);
-	if (!dvobj->usb_alloc_vendor_req_buf) {
-		DBG_88E("alloc usb_vendor_req_buf failed... /n");
-		rst = _FAIL;
-		goto exit;
-	}
-	dvobj->usb_vendor_req_buf = (u8 *)N_BYTE_ALIGMENT((size_t)(dvobj->usb_alloc_vendor_req_buf), ALIGNMENT_UNIT);
-exit:
-	return rst;
+	return _SUCCESS;
 }
 
 static void rtw_deinit_intf_priv(struct dvobj_priv *dvobj)
 {
-	kfree(dvobj->usb_alloc_vendor_req_buf);
 	mutex_destroy(&dvobj->usb_vendor_req_mutex);
 }
 
Signature · 1 line
2.33.0
#35 lore raw

PATCH v8 19/19 staging: r8188eu: remove usb_vendor_req_mutex

In reply to Fabio M. De Francesco

From: Pavel Skripkin <[email protected]>

This mutex was used to protect shared buffer for USB requests. Since buffer was removed in previous patch we can remove this mutex as well.

---
 drivers/staging/r8188eu/hal/usb_ops_linux.c | 18 ++++-------------
 drivers/staging/r8188eu/os_dep/usb_intf.c   | 22 ++-------------------
 2 files changed, 6 insertions(+), 34 deletions(-)
drivers/staging/r8188eu/hal/usb_ops_linux.c +4 -14
index 75500c28d6e6..19a37fd6a4b4 100644
@@ -16,8 +16,6 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 	int status;
 	u8 io_buf[4];
 
-	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
-
 	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
 		return -EPERM;
 
@@ -37,7 +35,7 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 		 * exist or is not enabled.
 		 */
 		adapt->bSurpriseRemoved = true;
-		goto mutex_unlock;
+		return status;
 	}
 
 	if (status < 0) {
@@ -47,15 +45,12 @@ static int usb_read(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 		if (rtw_inc_and_chk_continual_urb_error(dvobjpriv))
 			adapt->bSurpriseRemoved = true;
 
-		goto mutex_unlock;
+		return status;
 	}
 
 	rtw_reset_continual_urb_error(dvobjpriv);
 	memcpy(data, io_buf, size);
 
-mutex_unlock:
-	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
-
 	return status;
 }
 
@@ -67,8 +62,6 @@ static int usb_write(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 	int status;
 	u8 io_buf[VENDOR_CMD_MAX_DATA_LEN];
 
-	mutex_lock(&dvobjpriv->usb_vendor_req_mutex);
-
 	if (adapt->bSurpriseRemoved || adapt->pwrctrlpriv.pnp_bstop_trx)
 		return -EPERM;
 
@@ -89,7 +82,7 @@ static int usb_write(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 		 * exist or is not enabled.
 		 */
 		adapt->bSurpriseRemoved = true;
-		goto mutex_unlock;
+		return status;
 	}
 
 	if (status < 0) {
@@ -99,14 +92,11 @@ static int usb_write(struct intf_hdl *intfhdl, u16 value, void *data, u8 size)
 		if (rtw_inc_and_chk_continual_urb_error(dvobjpriv))
 			adapt->bSurpriseRemoved = true;
 
-		goto mutex_unlock;
+		return status;
 	}
 
 	rtw_reset_continual_urb_error(dvobjpriv);
 
-mutex_unlock:
-	mutex_unlock(&dvobjpriv->usb_vendor_req_mutex);
-
 	return status;
 }
 
drivers/staging/r8188eu/os_dep/usb_intf.c +2 -20
index 5ab42d55207f..2e6e6070c304 100644
@@ -73,21 +73,9 @@ static struct rtw_usb_drv rtl8188e_usb_drv = {
 
 static struct rtw_usb_drv *usb_drv = &rtl8188e_usb_drv;
 
-static u8 rtw_init_intf_priv(struct dvobj_priv *dvobj)
-{
-	mutex_init(&dvobj->usb_vendor_req_mutex);
-	return _SUCCESS;
-}
-
-static void rtw_deinit_intf_priv(struct dvobj_priv *dvobj)
-{
-	mutex_destroy(&dvobj->usb_vendor_req_mutex);
-}
-
 static struct dvobj_priv *usb_dvobj_init(struct usb_interface *usb_intf)
 {
 	int	i;
-	int	status = _FAIL;
 	struct dvobj_priv *pdvobjpriv;
 	struct usb_host_config		*phost_conf;
 	struct usb_config_descriptor	*pconf_desc;
@@ -146,19 +134,13 @@ static struct dvobj_priv *usb_dvobj_init(struct usb_interface *usb_intf)
 		DBG_88E("NON USB_SPEED_HIGH\n");
 	}
 
-	if (rtw_init_intf_priv(pdvobjpriv) == _FAIL)
-		goto free_dvobj;
-
 	/* 3 misc */
 	sema_init(&pdvobjpriv->usb_suspend_sema, 0);
 	rtw_reset_continual_urb_error(pdvobjpriv);
 
 	usb_get_dev(pusbd);
 
-	status = _SUCCESS;
-
-free_dvobj:
-	if (status != _SUCCESS && pdvobjpriv) {
+	if (pdvobjpriv) {
 		usb_set_intfdata(usb_intf, NULL);
 		kfree(pdvobjpriv);
 		pdvobjpriv = NULL;
@@ -188,7 +170,7 @@ static void usb_dvobj_deinit(struct usb_interface *usb_intf)
 				usb_reset_device(interface_to_usbdev(usb_intf));
 			}
 		}
-		rtw_deinit_intf_priv(dvobj);
+
 		kfree(dvobj);
 	}
 
Signature · 1 line
2.33.0
#36 lore raw
In reply to Fabio M. De Francesco
Quoted 42 lines · Fabio M. De Francesco

--- Preface ---

This is v8 of "shorten and simplify calls chain". The first 14 patches have already been applied to staging-testing, so we have been requested to reset the numbering of the remaining patches to 01/19, while discarding from this new submission the above-mentioned 14 patches (otherwise we would have submitted a series containing 33 patches).

The following commit message is provided as it was in v7, both for the purpose of presenting the whole picture to Maintainers, Reviewers, and to anybody else who may be interested in knowing the entire design and the evolution since v1 to the current v8.

--- Commit message ---

io_ops abstraction is useless in this driver, since there is only one ops registration. Without io_ops we can get rid of indirect calls mess and shorten the calls chain.

Shorten the calls chain of rtw_read8/16/32() down to the actual reads. For this purpose unify the three usb_read8/16/32 into the new usb_read(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_read() and use in it the new usb_control_msg_recv() API of USB Core.

Shorten the calls chain of rtw_write8/16/32() down to the actual writes. For this purpose unify the four usb_write8/16/32/N() into the new usb_write(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_write() and use in it the new usb_control_msg_send() API of USB Core.

The code with the modifications was thoroughly tested by Pavel Skripkin using a TP-Link TL-WN722N v2 / v3 [Realtek RTL8188EUS] and by Fabio M. De Francesco using a ASUSTek Computer, Inc. Realtek 8188EUS [USB-N10 Nano].

--- Changelog ---

v7->v8 (old numbering):
- 1-14:
Patches applied to staging-testing, so they are dropped from the current v8;

Quoted 4 lines

- 15-19:
Split into 19 patches. Numbering reset to 01. After this reset, 15-19/19 become 01-19/19 (so we have a total of 33 patches in this series.

Better, still needs a bit more work. I took 2 of these to shorten your load a bit :)

thanks,

greg k-h

#37 lore raw
Quoted 6 lines · CEST Greg Kroah-Hartman
Quoted 5 lines · Fabio M. De Francesco

--- Preface ---

This is v8 of "shorten and simplify calls chain". The first 14 patches have already been applied to staging-testing, so we have been requested to reset the numbering of the remaining patches to 01/19, while

discarding

Quoted 1 line
Quoted 1 line

from this new submission the above-mentioned 14 patches (otherwise we

would

Quoted 25 lines
Quoted 25 lines

have submitted a series containing 33 patches).

The following commit message is provided as it was in v7, both for the purpose of presenting the whole picture to Maintainers, Reviewers, and to anybody else who may be interested in knowing the entire design and the evolution since v1 to the current v8.

--- Commit message ---

io_ops abstraction is useless in this driver, since there is only one ops registration. Without io_ops we can get rid of indirect calls mess and shorten the calls chain.

Shorten the calls chain of rtw_read8/16/32() down to the actual reads. For this purpose unify the three usb_read8/16/32 into the new usb_read(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_read() and use in it the new usb_control_msg_recv() API of USB Core.

Shorten the calls chain of rtw_write8/16/32() down to the actual writes. For this purpose unify the four usb_write8/16/32/N() into the new usb_write(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_write() and use in it the new usb_control_msg_send() API of USB Core.

Quoted 3 lines
Quoted 3 lines

The code with the modifications was thoroughly tested by Pavel Skripkin using a TP-Link TL-WN722N v2 / v3 [Realtek RTL8188EUS] and by Fabio M. De Francesco using a ASUSTek Computer, Inc. Realtek 8188EUS [USB-N10

Nano].

Quoted 6 lines
Quoted 6 lines

--- Changelog ---

v7->v8 (old numbering):
- 1-14:
Patches applied to staging-testing, so they are

dropped

Quoted 4 lines
Quoted 4 lines

from the current v8;

- 15-19:
Split into 19 patches. Numbering reset to 01. After

this

Quoted 1 line
Quoted 1 line
		reset, 15-19/19 become 01-19/19 (so we have a total of 

33

Quoted 5 lines
Quoted 1 line

patches in this series.

Better, still needs a bit more work. I took 2 of these to shorten your load a bit :)

Thanks, Greg. I've already seen that you took those other 2 from the series. We are about half way to destination: 14 + 2 already taken. :)

I've not replied to the other messages of yours because they are easy to understand, fix or change commit messages, and resend as v9.

Instead I'm replying to this particular message because I'm not able to see where "a bit more work" is needed.

Maybe that you don't like that in paragraph "v7->v8 (old numbering" I only said that the old 15-19/19 (4 patches) are now split into 19 without further information about each logical group within those 19?

Please notice that, immediately after the above-mentioned section there is also the "v7->v8 (new numbering for the old 15-19/19):" new section where everything is detailed to a level that we considered sufficient.

For your convenience I copy-paste the lines that immediately follow the "old numbering" section:

v7->v8 (new numbering for the old 15-19/19):
- 1-15:
Old 15/19 and 16/19 are split into 15 patches (as it is required by Greg Kroah-Hartman) in order to make them more easily reviewable and for checking that the logic is not broken; no significant changes to the resulting code;
- 16-17:
Old 17/19 and 18/19 become 16/19 and 17/19: There are no significant changes to the code, with the sole exception of moving the acquiring of a mutex before the test for "bSurpriseRemoved" being 'true';
- 18-19:
Old 19/19 is split into new 18/19 and 19/19. The changes are split into a first patch that remove a shared buffer and a second that remove the mutex protecting the receiving (in usb_read()) and sending (in usb_write()) of the USB control messages.

If it is not what you require, can you please provide some more hints about what you would consider the "perfect" approach to rework this changelog?

It would help a lot, because at the moment I cannot see what I'm missing and I've been unable to reach Pavel and ask him if he saw something wrong that I can't see.

Thanks,

Fabio

Quoted 5 lines

thanks,

greg k-h

#38 lore raw
Quoted 58 lines · Fabio M. De Francesco
Quoted 6 lines · CEST Greg Kroah-Hartman
Quoted 5 lines · Fabio M. De Francesco

--- Preface ---

This is v8 of "shorten and simplify calls chain". The first 14 patches have already been applied to staging-testing, so we have been requested to reset the numbering of the remaining patches to 01/19, while

discarding

Quoted 1 line
Quoted 1 line

from this new submission the above-mentioned 14 patches (otherwise we

would

Quoted 25 lines
Quoted 25 lines

have submitted a series containing 33 patches).

The following commit message is provided as it was in v7, both for the purpose of presenting the whole picture to Maintainers, Reviewers, and to anybody else who may be interested in knowing the entire design and the evolution since v1 to the current v8.

--- Commit message ---

io_ops abstraction is useless in this driver, since there is only one ops registration. Without io_ops we can get rid of indirect calls mess and shorten the calls chain.

Shorten the calls chain of rtw_read8/16/32() down to the actual reads. For this purpose unify the three usb_read8/16/32 into the new usb_read(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_read() and use in it the new usb_control_msg_recv() API of USB Core.

Shorten the calls chain of rtw_write8/16/32() down to the actual writes. For this purpose unify the four usb_write8/16/32/N() into the new usb_write(); make the latter parameterizable with 'size'; embed most of the code of usbctrl_vendorreq() into usb_write() and use in it the new usb_control_msg_send() API of USB Core.

Quoted 3 lines
Quoted 3 lines

The code with the modifications was thoroughly tested by Pavel Skripkin using a TP-Link TL-WN722N v2 / v3 [Realtek RTL8188EUS] and by Fabio M. De Francesco using a ASUSTek Computer, Inc. Realtek 8188EUS [USB-N10

Nano].

Quoted 6 lines
Quoted 6 lines

--- Changelog ---

v7->v8 (old numbering):
- 1-14:
Patches applied to staging-testing, so they are

dropped

Quoted 4 lines
Quoted 4 lines

from the current v8;

- 15-19:
Split into 19 patches. Numbering reset to 01. After

this

Quoted 1 line
Quoted 1 line
		reset, 15-19/19 become 01-19/19 (so we have a total of 

33

Quoted 5 lines
Quoted 1 line

patches in this series.

Better, still needs a bit more work. I took 2 of these to shorten your load a bit :)

Quoted 8 lines

Thanks, Greg. I've already seen that you took those other 2 from the series. We are about half way to destination: 14 + 2 already taken. :)

I've not replied to the other messages of yours because they are easy to understand, fix or change commit messages, and resend as v9.

Instead I'm replying to this particular message because I'm not able to see where "a bit more work" is needed.

The "bit more work" was my review comments on the patches.

thanks,

greg k-h