With a 6.10 based kernel, no changes to filesystem/MM code, I'm seeing a reliable BUG_ON() within minutes of booting on one of my VMs. I don't have a complete oops dump, but this is what I do have, cobbled together from what was logged by journald, and what syslogd was able to splat on the terminals before the VM died. Sep 04 15:51:46 lists kernel: kernel BUG at fs/ext4/inode.c:1967! [ 1346.494848] Call trace: [ 1346.495409] [<c04b4f90>] (mpage_process_page_bufs) from [<c04b938c>] (mpage_prepare_extent_to_map+0x410/0x51c) [ 1346.499202] [<c04b938c>] (mpage_prepare_extent_to_map) from [<c04bbc40>] (ext4_do_writepages+0x320/0xb94) [ 1346.502113] [<c04bbc40>] (ext4_do_writepages) from [<c04bc5dc>] (ext4_writepages+0xc0/0x1b4) [ 1346.504662] [<c04bc5dc>] (ext4_writepages) from [<c0361154>] (do_writepages+0x68/0x220) [ 1346.506974] [<c0361154>] (do_writepages) from [<c0354868>] (filemap_fdatawrite_wbc+0x64/0x84) [ 1346.509165] [<c0354868>] (filemap_fdatawrite_wbc) from [<c035706c>] (__filemap_fdatawrite_range+0x50/0x58) [ 1346.511414] [<c035706c>] (__filemap_fdatawrite_range) from [<c035709c>] (filemap_flush+0x28/0x30) [ 1346.513518] [<c035709c>] (filemap_flush) from [<c04a8834>] (ext4_release_file+0x70/0xac) [ 1346.515312] [<c04a8834>] (ext4_release_file) from [<c03f8088>] (__fput+0xd4/0x2cc) [ 1346.517219] [<c03f8088>] (__fput) from [<c03f3e64>] (sys_close+0x28/0x5c) [ 1346.518720] [<c03f3e64>] (sys_close) from [<c0200060>] (ret_fast_syscall+0x0/0x5c)
