[PATCH] Input: evdev: Fall back to vmalloc for client event buffer

[PATCH] Input: evdev: Fall back to vmalloc for client event buffer

[Daniel Stone]

evdev always tries to allocate the event buffer for clients using
kzalloc rather than vmalloc, presumably to avoid mapping overhead where
possible.  However, drivers like bcm5974, which claims support for
reporting 16 fingers simultaneously, can have an extraordinarily large
buffer.  The resultant contiguous order-4 allocation attempt fails due
to fragmentation, and the device is thus unusable until reboot.

Try kzalloc if we can to avoid the mapping overhead, but if that fails,
fall back to vzalloc.

Signed-off-by: Daniel Stone <daniels@collabora.com>
---
 drivers/input/evdev.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c
index f0f8928..edfca4a 100644
--- a/drivers/input/evdev.c
+++ b/drivers/input/evdev.c
@@ -18,6 +18,7 @@
 #include <linux/poll.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
+#include <linux/mm.h>
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/input/mt.h>
@@ -289,7 +290,11 @@ static int evdev_release(struct inode *inode, struct file *file)
 	mutex_unlock(&evdev->mutex);
 
 	evdev_detach_client(evdev, client);
-	kfree(client);
+
+	if (is_vmalloc_addr(client))
+		vfree(client);
+	else
+		kfree(client);
 
 	evdev_close_device(evdev);
 
@@ -311,10 +316,12 @@ static int evdev_open(struct inode *inode, struct file *file)
 	unsigned int bufsize = evdev_compute_buffer_size(evdev->handle.dev);
 	struct evdev_client *client;
 	int error;
+	int size =
+		sizeof(struct evdev_client) + bufsize * sizeof(struct input_event);
 
-	client = kzalloc(sizeof(struct evdev_client) +
-				bufsize * sizeof(struct input_event),
-			 GFP_KERNEL);
+	client = kzalloc(size, GFP_KERNEL | __GFP_NOWARN);
+	if (!client)
+		client = vzalloc(size);
 	if (!client)
 		return -ENOMEM;
 
-- 
1.8.3.1

Re: [PATCH] Input: evdev: Fall back to vmalloc for client event buffer

[Daniel Stone]

Hi,

On 7 October 2013 14:06, Daniel Stone <daniel@fooishbar.org> wrote:
> evdev always tries to allocate the event buffer for clients using
> kzalloc rather than vmalloc, presumably to avoid mapping overhead where
> possible.  However, drivers like bcm5974, which claims support for
> reporting 16 fingers simultaneously, can have an extraordinarily large
> buffer.  The resultant contiguous order-4 allocation attempt fails due
> to fragmentation, and the device is thus unusable until reboot.
>
> Try kzalloc if we can to avoid the mapping overhead, but if that fails,
> fall back to vzalloc.
>
> Signed-off-by: Daniel Stone <daniels@collabora.com>

Ping? This bit me again yesterday, and is pretty much guaranteed to
happen eventually, if you suspend and resume enough times.

Cheers,
Daniel

Re: [PATCH] Input: evdev: Fall back to vmalloc for client event buffer

[Dmitry Torokhov]

On Mon, Oct 07, 2013 at 02:06:53PM +0100, Daniel Stone wrote:
> evdev always tries to allocate the event buffer for clients using
> kzalloc rather than vmalloc, presumably to avoid mapping overhead where
> possible.  However, drivers like bcm5974, which claims support for
> reporting 16 fingers simultaneously, can have an extraordinarily large
> buffer.  The resultant contiguous order-4 allocation attempt fails due
> to fragmentation, and the device is thus unusable until reboot.
> 
> Try kzalloc if we can to avoid the mapping overhead, but if that fails,
> fall back to vzalloc.
> 
> Signed-off-by: Daniel Stone <daniels@collabora.com>

Applied, thank you Daniel.

-- 
Dmitry