On Wed, Dec 27, 2023 at 7:38 PM Chengming Zhou
<zhouchengming@xxxxxxxxxxxxx> wrote:
>
> On 2023/12/27 14:25, Barry Song wrote:
> > On Wed, Dec 27, 2023 at 4:51 PM Chengming Zhou
> > <zhouchengming@xxxxxxxxxxxxx> wrote:
> >>
> >> On 2023/12/27 08:23, Nhat Pham wrote:
> >>> On Tue, Dec 26, 2023 at 3:30 PM Chris Li <chrisl@xxxxxxxxxx> wrote:
> >>>>
> >>>> Again, sorry I was looking at the decompression side rather than the
> >>>> compression side. The compression side does not even offer a safe
> >>>> version of the compression function.
> >>>> That seems to be dangerous. It seems for now we should make the zswap
> >>>> roll back to 2 page buffer until we have a safe way to do compression
> >>>> without overwriting the output buffers.
> >>>
> >>> Unfortunately, I think this is the way - at least until we rework the
> >>> crypto/compression API (if that's even possible?).
> >>> I still think the 2 page buffer is dumb, but it is what it is :(
> >>
> >> Hi,
> >>
> >> I think it's a bug in `scomp_acomp_comp_decomp()`, which doesn't use
> >> the caller passed "src" and "dst" scatterlist. Instead, it uses its own
> >> per-cpu "scomp_scratch", which have 128KB src and dst.
> >>
> >> When compression done, it uses the output req->dlen to copy scomp_scratch->dst
> >> to our dstmem, which has only one page now, so this problem happened.
> >>
> >> I still don't know why the alg->compress(src, slen, dst, &dlen) doesn't
> >> check the dlen? It seems an obvious bug, right?
> >>
> >> As for this problem in `scomp_acomp_comp_decomp()`, this patch below
> >> should fix it. I will set up a few tests to check later.
> >>
> >> Thanks!
> >>
> >> diff --git a/crypto/scompress.c b/crypto/scompress.c
> >> index 442a82c9de7d..e654a120ae5a 100644
> >> --- a/crypto/scompress.c
> >> +++ b/crypto/scompress.c
> >> @@ -117,6 +117,7 @@ static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
> >> struct crypto_scomp *scomp = *tfm_ctx;
> >> void **ctx = acomp_request_ctx(req);
> >> struct scomp_scratch *scratch;
> >> + unsigned int dlen;
> >> int ret;
> >>
> >> if (!req->src || !req->slen || req->slen > SCOMP_SCRATCH_SIZE)
> >> @@ -128,6 +129,8 @@ static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
> >> if (!req->dlen || req->dlen > SCOMP_SCRATCH_SIZE)
> >> req->dlen = SCOMP_SCRATCH_SIZE;
> >>
> >> + dlen = req->dlen;
> >> +
> >> scratch = raw_cpu_ptr(&scomp_scratch);
> >> spin_lock(&scratch->lock);
> >>
> >> @@ -145,6 +148,9 @@ static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
> >> ret = -ENOMEM;
> >> goto out;
> >> }
> >> + } else if (req->dlen > dlen) {
> >> + ret = -ENOMEM;
> >> + goto out;
> >> }
> >
> > This can't fix the problem as crypto_scomp_compress() has written overflow data.
>
> No, crypto_scomp_compress() writes to its own scomp_scratch->dst memory, then copy
> to our dstmem.
Hi Chengming,
I still feel these two memcpys are too big and unnecessary, so i sent
a RFC[1] to remove
them as well as another one removing memcpy in zswap[2].
but unfortunately I don't have real hardware to run and collect data,
I wonder if you are
interested in testing and collecting data as you are actively
contributing to zswap.
[1] https://lore.kernel.org/linux-mm/20240103053134.564457-1-21cnbao@xxxxxxxxx/
[2]
https://lore.kernel.org/linux-mm/20240103025759.523120-1-21cnbao@xxxxxxxxx/
https://lore.kernel.org/linux-mm/20240103025759.523120-2-21cnbao@xxxxxxxxx/
Thanks
Barry
