> -----Original Message----- > From: Jarkko Sakkinen [mailto:jarkko.sakkinen@xxxxxxxxxxxxxxx] > Sent: Tuesday, February 05, 2019 16:36 > To: Winkler, Tomas <tomas.winkler@xxxxxxxxx> > Cc: linux-integrity@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; linux- > security-module@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx; James Morris > <jmorris@xxxxxxxxx>; Jerry Snitselaar <jsnitsel@xxxxxxxxxx> > Subject: Re: [PATCH v3] tpm/tpm_crb: Avoid unaligned reads in crb_recv() > > On Tue, Feb 05, 2019 at 11:07:16AM +0000, Winkler, Tomas wrote: > > > The current approach to read first 6 bytes from the response and > > > then tail of the response, can cause the 2nd memcpy_fromio() to do > > > an unaligned read (e.g. read 32-bit word from address aligned to a > > > 16-bits), depending on how > > > memcpy_fromio() is implemented. If this happens, the read will fail > > > and the memory controller will fill the read with 1's. > > > > > > This was triggered by 170d13ca3a2f, which should be probably refined > > > to check and react to the address alignment. Before that commit, on > > > x86 > > > memcpy_fromio() turned out to be memcpy(). By a luck GCC has done > > > the right thing (from tpm_crb's perspective) for us so far, but we should not > rely on that. > > > Thus, it makes sense to fix this also in tpm_crb, not least because > > > the fix can be then backported to stable kernels and make them more > > > robust when compiled in differing environments. > > > > > > Cc: stable@xxxxxxxxxxxxxxx > > > Cc: James Morris <jmorris@xxxxxxxxx> > > > Cc: Tomas Winkler <tomas.winkler@xxxxxxxxx> > > > Cc: Jerry Snitselaar <jsnitsel@xxxxxxxxxx> > > > Fixes: 30fc8d138e91 ("tpm: TPM 2.0 CRB Interface") > > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> > > > Reviewed-by: Jerry Snitselaar <jsnitsel@xxxxxxxxxx> > > > --- > > > v3: > > > * Fix typo i.e. %s/reminding/remaining/g > > > > Why you haven't fixed all the typos I've pointed out? I think you missed that. > > I saw only comment about remaining. Was there something else? Can fix. https://www.spinics.net/lists/stable/msg283648.html 1. unrecovable -> unrecoverable 2. /* Read 8 bytes (not just 6 bytes, which would cover the tag and the response length > + * fields) in order to make sure that the remaining memory accesses */ Thanks Tomas
