On 2024-04-18 Harshit Mogalapalli wrote:

While fuzzing 5.15.y kernel with Syzkaller, we noticed a INFO: task hung 
bug in fb_deferred_io_work()

I think the problem is because of improper offset address calculation.
The kernel calculate address offset with:
	offset = vmf->address - vmf->vma->vm_start

Now the problem is that your C program mmap the framebuffer at 2
different offsets:
	mmap(ptr, 4096, PROT_WRITE, MAP_FIXED | MAP_SHARED, fd, 0xff000);
	mmap(ptr, 4096, PROT_WRITE, MAP_FIXED | MAP_SHARED, fd, 0);

but the kernel doesn't take these different offsets into account.
So, 2 different pages are mistakenly recognized as the same page.

Can you try the following patch?

Best regards,
Nam

diff --git a/drivers/video/fbdev/core/fb_defio.c b/drivers/video/fbdev/core/fb_defio.c
index dae96c9f61cf..d5d6cd9e8b29 100644
--- a/drivers/video/fbdev/core/fb_defio.c
+++ b/drivers/video/fbdev/core/fb_defio.c

@@ -196,7 +196,8 @@ static vm_fault_t fb_deferred_io_track_page(struct fb_info *info, unsigned long
  */
 static vm_fault_t fb_deferred_io_page_mkwrite(struct fb_info *info, struct vm_fault *vmf)
 {
-	unsigned long offset = vmf->address - vmf->vma->vm_start;
+	unsigned long offset = vmf->address - vmf->vma->vm_start
+			+ (vmf->vma->vm_pgoff << PAGE_SHIFT);
 	struct page *page = vmf->page;
 
 	file_update_time(vmf->vma->vm_file);