Skip to contentSkip to conversation

usb: gadget: f_mass_storage: cahnge wait_event to wait_event_timeout

PATCH In review Oh Eomji 4 participants 4 messages Jan 21, 2021 1 message not in the archive USB

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): CGME20210121070836epcas2p130c0f62d82aa3fcd2e021a1ef88a7ebd@epcas2p1.samsung.com

Oh Eomji

author
#1 lore raw

PATCH 1/1 usb: gadget: f_mass_storage: cahnge wait_event to wait_event_timeout

Changed to return a timeout error if there is no response for a certain period of time in order to solve the problem of waiting for a event complete while executing unbind.

---
 drivers/usb/gadget/function/f_mass_storage.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/gadget/function/f_mass_storage.c +1 -1
index 950c943..b474840 100644
@@ -3000,7 +3000,7 @@ static void fsg_unbind(struct usb_configuration *c, struct usb_function *f)
 	if (fsg->common->fsg == fsg) {
 		__raise_exception(fsg->common, FSG_STATE_CONFIG_CHANGE, NULL);
 		/* FIXME: make interruptible or killable somehow? */
-		wait_event(common->fsg_wait, common->fsg != fsg);
+		wait_event_timeout(common->fsg_wait, common->fsg != fsg, HZ / 4);
 	}
 
 	usb_free_all_descriptors(&fsg->function);
Signature · 1 line
2.7.4
#2 lore raw
Quoted 3 lines · Oh Eomji

Changed to return a timeout error if there is no response for a certain period of time in order to solve the problem of waiting for a event complete while executing unbind.

Can you shed a light on the choice of the timeout length?

Quoted 21 lines
---
 drivers/usb/gadget/function/f_mass_storage.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/gadget/function/f_mass_storage.c +1 -1
index 950c943..b474840 100644
@@ -3000,7 +3000,7 @@ static void fsg_unbind(struct usb_configuration *c, struct usb_function *f)
 	if (fsg->common->fsg == fsg) {
 		__raise_exception(fsg->common, FSG_STATE_CONFIG_CHANGE, NULL);
 		/* FIXME: make interruptible or killable somehow? */
-		wait_event(common->fsg_wait, common->fsg != fsg);
+		wait_event_timeout(common->fsg_wait, common->fsg != fsg, HZ / 4);
 	}
 
 	usb_free_all_descriptors(&fsg->function);
Signature · 1 line
2.7.4
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#3 lore raw
In reply to Oh Eomji
Quoted 14 lines · Oh Eomji

Changed to return a timeout error if there is no response for a certain period of time in order to solve the problem of waiting for a event complete while executing unbind.

---
 drivers/usb/gadget/function/f_mass_storage.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/gadget/function/f_mass_storage.c +0 -0
index 950c943..b474840 100644

drivers/usb/gadget/function/f_mass_storage.c static void fsg_unbind(struct usb_configuration *c, struct usb_function *f) · line 3003

Quoted 5 lines
drivers/usb/gadget/function/f_mass_storage.c +1 -1
@@ -3000,7 +3000,7 @@ static void fsg_unbind(struct usb_configuration *c, struct usb_function *f)
 	if (fsg->common->fsg == fsg) {
 		__raise_exception(fsg->common, FSG_STATE_CONFIG_CHANGE, NULL);
 		/* FIXME: make interruptible or killable somehow? */
-		wait_event(common->fsg_wait, common->fsg != fsg);
+		wait_event_timeout(common->fsg_wait, common->fsg != fsg, HZ / 4);

That's a random choice of a timeout value.

Please document this really really really well as to why you picked this number, and what it means.

Also, is the commet above this line still correct now?

thanks,

greg k-h

#4 lore raw
In reply to Oh Eomji
Quoted 14 lines · Oh Eomji

Changed to return a timeout error if there is no response for a certain period of time in order to solve the problem of waiting for a event complete while executing unbind.

---
 drivers/usb/gadget/function/f_mass_storage.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/gadget/function/f_mass_storage.c +0 -0
index 950c943..b474840 100644

drivers/usb/gadget/function/f_mass_storage.c static void fsg_unbind(struct usb_configuration *c, struct usb_function *f) · line 3006

Quoted 8 lines
drivers/usb/gadget/function/f_mass_storage.c +1 -1
@@ -3000,7 +3000,7 @@ static void fsg_unbind(struct usb_configuration *c, struct usb_function *f)
 	if (fsg->common->fsg == fsg) {
 		__raise_exception(fsg->common, FSG_STATE_CONFIG_CHANGE, NULL);
 		/* FIXME: make interruptible or killable somehow? */
-		wait_event(common->fsg_wait, common->fsg != fsg);
+		wait_event_timeout(common->fsg_wait, common->fsg != fsg, HZ / 4);
 	}
 
 	usb_free_all_descriptors(&fsg->function);

No, no, no!

This patch completely misunderstands the purpose of the wait_event call. The reason it isn't interruptible is not because that would be difficult -- all it takes is adding a timeout argument, as you did here.

The reason is because the driver will get invalid memory accesses if fsg_unbind returns too early. fsg will be deallocated, but fsg_set_interface will continue to use it. This patch completely ignores that issue.

Was there any real reason for writing this patch? Does it solve a real-world problem? Did you encounter a situation where the wait_event call would wait for more than 1/4 second?

Alan Stern