2013-10-01 13:35:58

by Manoj Chourasia

[permalink] [raw]

Subject: FW: [PATCH] hidraw: close underlying device at removal of last reader

---

Hi Jiri,

Mika Westerberg [email protected] reported a bug introduced in hid driver after
commit 212a871a393 (HID: hidraw: correctly deallocate memory on device disconnect)
that hidraw doesn't close the underlying hid device when the device node is closed last time.

I am proposing following patch as fix. ?Please review it. ?Mika will help in testing.

----------------------------------------
>From b37519b65be0d51b90c5fc177c0dc0cd2f6358b1 Mon Sep 17 00:00:00 2001
From: Manoj Chourasia <[email protected]>
Date: Tue, 1 Oct 2013 15:39:00 +0530
Subject: [PATCH] HID: hidraw: close underlying device at removal of last reader

Even though device exist bit is set the underlying
HW device should be closed when the last reader
of the device is closed i.e. open count drops to zero.

Signed-off-by: Manoj Chourasia <[email protected]>
---
drivers/hid/hidraw.c | 21 +++++++++++++++------
1 file changed, 15 insertions(+), 6 deletions(-)

diff --git a/drivers/hid/hidraw.c b/drivers/hid/hidraw.c
index dbfe300..3c0dd44 100644
--- a/drivers/hid/hidraw.c
+++ b/drivers/hid/hidraw.c
@@ -305,19 +305,28 @@ static int hidraw_fasync(int fd, struct file *file, int on)
static void drop_ref(struct hidraw *hidraw, int exists_bit)
{
if (exists_bit) {
- hid_hw_close(hidraw->hid);
hidraw->exist = 0;
- if (hidraw->open)
+ if (hidraw->open) {
+ hid_hw_close(hidraw->hid);
wake_up_interruptible(&hidraw->wait);
+ }
} else {
--hidraw->open;
}

- if (!hidraw->open && !hidraw->exist) {
- device_destroy(hidraw_class, MKDEV(hidraw_major, hidraw->minor));
- hidraw_table[hidraw->minor] = NULL;
- kfree(hidraw);
+ if (!hidraw->open) {
+ if (!hidraw->exist) {
+ device_destroy(hidraw_class,
+ MKDEV(hidraw_major, hidraw->minor));
+ hidraw_table[hidraw->minor] = NULL;
+ kfree(hidraw);
+ } else {
+ /* close device for last reader */
+ hid_hw_power(hidraw->hid, PM_HINT_NORMAL);
+ hid_hw_close(hidraw->hid);
+ }
}
+
}

static int hidraw_release(struct inode * inode, struct file * file)
--
1.7.9.5


-Regards
Manoj

---

Attachments:

PATCH-hidraw-close-underlying-device-at-removal-of-las.patch (1.60 kB)
PATCH-hidraw-close-underlying-device-at-removal-of-las.patch

2013-10-02 15:02:52

by Jiri Kosina

[permalink] [raw]

Subject: Re: FW: [PATCH] hidraw: close underlying device at removal of last reader

---

On Tue, 1 Oct 2013, Manoj Chourasia wrote:

> Hi Jiri,
>
> Mika Westerberg [email protected] reported a bug introduced in hid driver after
> commit 212a871a393 (HID: hidraw: correctly deallocate memory on device disconnect)
> that hidraw doesn't close the underlying hid device when the device node is closed last time.
>
> I am proposing following patch as fix. ?Please review it. ?Mika will help in testing.
>
> ----------------------------------------
> >From b37519b65be0d51b90c5fc177c0dc0cd2f6358b1 Mon Sep 17 00:00:00 2001
> From: Manoj Chourasia <[email protected]>
> Date: Tue, 1 Oct 2013 15:39:00 +0530
> Subject: [PATCH] HID: hidraw: close underlying device at removal of last reader
>
> Even though device exist bit is set the underlying
> HW device should be closed when the last reader
> of the device is closed i.e. open count drops to zero.
>
> Signed-off-by: Manoj Chourasia <[email protected]>

Applied, thanks a lot to both of you.

--
Jiri Kosina
SUSE Labs