[U-Boot] [PATCH v1] ubifs: avoid assert failed in ubifs.c

Richard Weinberger richard at nod.at
Tue May 22 11:23:13 UTC 2018


Am Dienstag, 22. Mai 2018, 12:56:48 CEST schrieb Marek Vasut:
> On 05/10/2018 10:57 PM, Marek Vasut wrote:
> > On 04/27/2018 03:51 PM, Patrice Chotard wrote:
> >> This patch solves assert failed displayed in the console during a boot.
> >> The root cause is that the ubifs_inode is not already allocated when
> >> ubifs_printdir and ubifs_finddir functions are called.
> >>
> >> Trace showing the issue:
> >> feed 'boot.scr.uimg', ino 94, new f_pos 0x17b40ece
> >> dent->ch.sqnum '7132', creat_sqnum 3886945402880
> >> UBIFS assert failed in ubifs_finddir at 436
> >> INODE ALLOCATION: creat_sqnum '7129'
> >> Found U-Boot script /boot.scr.uimg
> >>
> >> Signed-off-by: Christophe Kerello <christophe.kerello at st.com>
> >> Signed-off-by: Patrice Chotard <patrice.chotard at st.com>
> > 
> > I ran into this too, but what I do not quite understand from the commit
> > message is how hiding the error actually solves the problem that the
> > assert points to.
> > 
> > Why does the assert trigger in the first place ?
> > 
> > What is the root cause of the issue that is being hidden by this patch?
> 
> Bump?

I had a look, the bug is deeper, ubifs_finddir() allocates a vfs inode manually
and ignores UBIFS internals. ubifs_inode() will read beyond the allocated buffer.
In best case the assert triggers because ->creat_sqnum is garbage, in worst case, U-Boot will
just crash.

AFAICT, the correct solution is to use ubifs_iget().
Then we can keep the assert and it will check for the right thing.

Thanks,
//richard


More information about the U-Boot mailing list