Regression: system doesn't start if no DVD drive is connected #17
Comments
|
@RattletraPM it is a kernel bug and it was already there in DeltaResero's version. Thanks for clarifying out the exact way it affects the system. I am not sure how this could be fixed though? It requires some kernel patch? |
|
I've done a few more tests since making this issue and found a dirty fix. The issue is indeed caused by the rvl-di driver, however trying to blacklist it simply wouldn't work. I tried both adding it to Being the lazy person I am, I didn't want to mess with the source code and recompile the whole kernel, but before giving up I had an idea. What I thought about doing was to open the kernel's zImage in a hexeditor, find the device tree, find the string hollywood-di (offset 0x1B191) and change it to something else so it wouldn't be found at boot, and thus it would get ignored. I told you it was a dirty fix but hey, it worked! So, I don't know how helpful this would be to make a proper fix, unlike mine, but at least there is a way to make Linux boot without a DVD drive for now. Most likely, giving an option to boot without DVD support via Gumboot would be enough if you don't want to touch the driver's source, so considering I know this is a niche issue, I'd say it could be an okay fix. |
[ Upstream commit 67d7d8a ] Hulk Robot reported a issue: ================================================================== BUG: KASAN: use-after-free in ext4_xattr_set_entry+0x18ab/0x3500 Write of size 4105 at addr ffff8881675ef5f4 by task syz-executor.0/7092 CPU: 1 PID: 7092 Comm: syz-executor.0 Not tainted 4.19.90-dirty neagix#17 Call Trace: [...] memcpy+0x34/0x50 mm/kasan/kasan.c:303 ext4_xattr_set_entry+0x18ab/0x3500 fs/ext4/xattr.c:1747 ext4_xattr_ibody_inline_set+0x86/0x2a0 fs/ext4/xattr.c:2205 ext4_xattr_set_handle+0x940/0x1300 fs/ext4/xattr.c:2386 ext4_xattr_set+0x1da/0x300 fs/ext4/xattr.c:2498 __vfs_setxattr+0x112/0x170 fs/xattr.c:149 __vfs_setxattr_noperm+0x11b/0x2a0 fs/xattr.c:180 __vfs_setxattr_locked+0x17b/0x250 fs/xattr.c:238 vfs_setxattr+0xed/0x270 fs/xattr.c:255 setxattr+0x235/0x330 fs/xattr.c:520 path_setxattr+0x176/0x190 fs/xattr.c:539 __do_sys_lsetxattr fs/xattr.c:561 [inline] __se_sys_lsetxattr fs/xattr.c:557 [inline] __x64_sys_lsetxattr+0xc2/0x160 fs/xattr.c:557 do_syscall_64+0xdf/0x530 arch/x86/entry/common.c:298 entry_SYSCALL_64_after_hwframe+0x44/0xa9 RIP: 0033:0x459fe9 RSP: 002b:00007fa5e54b4c08 EFLAGS: 00000246 ORIG_RAX: 00000000000000bd RAX: ffffffffffffffda RBX: 000000000051bf60 RCX: 0000000000459fe9 RDX: 00000000200003c0 RSI: 0000000020000180 RDI: 0000000020000140 RBP: 000000000051bf60 R08: 0000000000000001 R09: 0000000000000000 R10: 0000000000001009 R11: 0000000000000246 R12: 0000000000000000 R13: 00007ffc73c93fc0 R14: 000000000051bf60 R15: 00007fa5e54b4d80 [...] ================================================================== Above issue may happen as follows: ------------------------------------- ext4_xattr_set ext4_xattr_set_handle ext4_xattr_ibody_find >> s->end < s->base >> no EXT4_STATE_XATTR >> xattr_check_inode is not executed ext4_xattr_ibody_set ext4_xattr_set_entry >> size_t min_offs = s->end - s->base >> UAF in memcpy we can easily reproduce this problem with the following commands: mkfs.ext4 -F /dev/sda mount -o debug_want_extra_isize=128 /dev/sda /mnt touch /mnt/file setfattr -n user.cat -v `seq -s z 4096|tr -d '[:digit:]'` /mnt/file In ext4_xattr_ibody_find, we have the following assignment logic: header = IHDR(inode, raw_inode) = raw_inode + EXT4_GOOD_OLD_INODE_SIZE + i_extra_isize is->s.base = IFIRST(header) = header + sizeof(struct ext4_xattr_ibody_header) is->s.end = raw_inode + s_inode_size In ext4_xattr_set_entry min_offs = s->end - s->base = s_inode_size - EXT4_GOOD_OLD_INODE_SIZE - i_extra_isize - sizeof(struct ext4_xattr_ibody_header) last = s->first free = min_offs - ((void *)last - s->base) - sizeof(__u32) = s_inode_size - EXT4_GOOD_OLD_INODE_SIZE - i_extra_isize - sizeof(struct ext4_xattr_ibody_header) - sizeof(__u32) In the calculation formula, all values except s_inode_size and i_extra_size are fixed values. When i_extra_size is the maximum value s_inode_size - EXT4_GOOD_OLD_INODE_SIZE, min_offs is -4 and free is -8. The value overflows. As a result, the preceding issue is triggered when memcpy is executed. Therefore, when finding xattr or setting xattr, check whether there is space for storing xattr in the inode to resolve this issue. Cc: stable@kernel.org Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220616021358.2504451-3-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> [uli: backport to 4.4] Signed-off-by: Ulrich Hecht <uli+cip@fpond.eu>
(I have opened a separate issue because it's unrelated from the previous one I've opened)
I've noticed that the OS won't start properly if no DVD drive is connected. This happens in MINI mode and seems to be a regression since mikep4, as that kernel is able to boot even with no DVD drive at all. mikep5 and DeltaResero's kernels seem to be affected aswell. I haven't tested any of the newer kernels in IOS mode, but mikep5 started up fine under IOS, which makes me think it's indeed related to the rvl-di module.
As for what happens, after booting via Gumboot the Tux logo shows up, but that's about it. Nothing happens, init doesn't seem to run at all. With older, non-gumboot kernels you'd just get a black screen. Connecting even just a DVD logic board will fix the issue, and the OS will boot just fine. This seems to happen on multiple Wii consoles and motherboards.
The text was updated successfully, but these errors were encountered: