The fbtft_framebuffer_alloc() calls fb_deferred_io_init() before
initializing info->fix.smem_len.  It is set to zero by the
framebuffer_alloc() function.  It will trigger a WARN_ON() at the
start of fb_deferred_io_init() and the function will not do anything.

Fixes: 856082f021a2 ("fbdev: defio: fix the pagelist corruption")

Signed-off-by: Peter Suti <peter.suti@streamunlimited.com>
---
 V2 -> V3: Add patch changelog 
 V1 -> V2: Change commit message and base on top of linux-next

 drivers/staging/fbtft/fbtft-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/fbtft/fbtft-core.c b/drivers/staging/fbtft/fbtft-core.c
index 9b3eaed80cdd..afaba94d1d1c 100644
--- a/drivers/staging/fbtft/fbtft-core.c
+++ b/drivers/staging/fbtft/fbtft-core.c
@@ -654,7 +654,6 @@ struct fb_info *fbtft_framebuffer_alloc(struct fbtft_display *display,
 	fbdefio->delay =            HZ / fps;
 	fbdefio->sort_pagereflist = true;
 	fbdefio->deferred_io =      fbtft_deferred_io;
-	fb_deferred_io_init(info);
 
 	snprintf(info->fix.id, sizeof(info->fix.id), "%s", dev->driver->name);
 	info->fix.type =           FB_TYPE_PACKED_PIXELS;
@@ -665,6 +664,7 @@ struct fb_info *fbtft_framebuffer_alloc(struct fbtft_display *display,
 	info->fix.line_length =    width * bpp / 8;
 	info->fix.accel =          FB_ACCEL_NONE;
 	info->fix.smem_len =       vmem_size;
+	fb_deferred_io_init(info);
 
 	info->var.rotate =         pdata->rotate;
 	info->var.xres =           width;
-- 
2.25.1

Hi Peter,

> -----Original Message-----
> From: Peter Suti <peter.suti@streamunlimited.com>
> Sent: Wednesday, July 27, 2022 3:36 PM
> To: Liu, Chuansheng <chuansheng.liu@intel.com>; dan.carpenter@oracle.com;
> Greg Kroah-Hartman <gregkh@linuxfoundation.org>; Thomas Zimmermann
> <tzimmermann@suse.de>; Javier Martinez Canillas <javierm@redhat.com>
> Cc: Peter Suti <peter.suti@streamunlimited.com>; dri-
> devel@lists.freedesktop.org; linux-fbdev@vger.kernel.org; linux-
> staging@lists.linux.dev; linux-kernel@vger.kernel.org
> Subject: [PATCH v3] staging: fbtft: core: set smem_len before
> fb_deferred_io_init call
> 
> The fbtft_framebuffer_alloc() calls fb_deferred_io_init() before
> initializing info->fix.smem_len.  It is set to zero by the
> framebuffer_alloc() function.  It will trigger a WARN_ON() at the
> start of fb_deferred_io_init() and the function will not do anything.

Please show us the WARN_ON() call stack, then everything is clear. It is the driver itself issue to be fixed.
I saw Thomas made such WARN_ON().

> 
> Fixes: 856082f021a2 ("fbdev: defio: fix the pagelist corruption")

Don't agree with such description.

Best Regards
Chuansheng

On Thu, Jul 28, 2022 at 12:49:15PM +0000, Liu, Chuansheng wrote:
> Hi Peter,
> 
> > 
> > The fbtft_framebuffer_alloc() calls fb_deferred_io_init() before
> > initializing info->fix.smem_len.  It is set to zero by the
> > framebuffer_alloc() function.  It will trigger a WARN_ON() at the
> > start of fb_deferred_io_init() and the function will not do anything.
> 
> Please show us the WARN_ON() call stack, then everything is clear.
> It is the driver itself issue to be fixed. I saw Thomas made such
> WARN_ON().

I don't understand the confusion here.  The code is very simple and the
description seems very clear.  No need to redo the patch.  I think
Peter tested it before the WARN_ON() was added so he probably didn't
see the WARN_ON().  I told him to add that because it's pretty obvious
what will happen just from reading the code.

> 
> > 
> > Fixes: 856082f021a2 ("fbdev: defio: fix the pagelist corruption")
> 
> Don't agree with such description.

I don't see how you can disagree?  Before your patch the
fb_deferred_io_init() did not use info->fix.smem_len and now it does.

regards,
dan carpenter