On Fri, Jun 14, 2019 at 9:13 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
>
> On Fri, 2019-06-14 at 16:05 +0800, Yan, Zheng wrote:
> > On Fri, Jun 7, 2019 at 11:38 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
> > > Add a way to decode an entity_addr_t. Once CEPH_FEATURE_MSG_ADDR2 is
> > > enabled, the server daemons will start encoding entity_addr_t
> > > differently.
> > >
> > > Add a new helper function that can handle either format.
> > >
> > > Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > ---
> > > include/linux/ceph/decode.h | 2 +
> > > net/ceph/Makefile | 2 +-
> > > net/ceph/decode.c | 75 +++++++++++++++++++++++++++++++++++++
> > > 3 files changed, 78 insertions(+), 1 deletion(-)
> > > create mode 100644 net/ceph/decode.c
> > >
> > > diff --git a/include/linux/ceph/decode.h b/include/linux/ceph/decode.h
> > > index a6c2a48d42e0..1c0a665bfc03 100644
> > > --- a/include/linux/ceph/decode.h
> > > +++ b/include/linux/ceph/decode.h
> > > @@ -230,6 +230,8 @@ static inline void ceph_decode_addr(struct ceph_entity_addr *a)
> > > WARN_ON(a->in_addr.ss_family == 512);
> > > }
> > >
> > > +extern int ceph_decode_entity_addr(void **p, void *end,
> > > + struct ceph_entity_addr *addr);
> > > /*
> > > * encoders
> > > */
> > > diff --git a/net/ceph/Makefile b/net/ceph/Makefile
> > > index db09defe27d0..59d0ba2072de 100644
> > > --- a/net/ceph/Makefile
> > > +++ b/net/ceph/Makefile
> > > @@ -5,7 +5,7 @@
> > > obj-$(CONFIG_CEPH_LIB) += libceph.o
> > >
> > > libceph-y := ceph_common.o messenger.o msgpool.o buffer.o pagelist.o \
> > > - mon_client.o \
> > > + mon_client.o decode.o \
> > > cls_lock_client.o \
> > > osd_client.o osdmap.o crush/crush.o crush/mapper.o crush/hash.o \
> > > striper.o \
> > > diff --git a/net/ceph/decode.c b/net/ceph/decode.c
> > > new file mode 100644
> > > index 000000000000..27edf5d341ec
> > > --- /dev/null
> > > +++ b/net/ceph/decode.c
> > > @@ -0,0 +1,75 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +
> > > +#include <linux/ceph/decode.h>
> > > +
> > > +int
> > > +ceph_decode_entity_addr(void **p, void *end, struct ceph_entity_addr *addr)
> > > +{
> > > + u8 marker, v, compat;
> >
> > It's better to use name struct_v, struct_compat
> >
> >
> > > + u32 len;
> > > +
> > > + ceph_decode_8_safe(p, end, marker, bad);
> > > + if (marker == 1) {
> > > + ceph_decode_8_safe(p, end, v, bad);
> > > + ceph_decode_8_safe(p, end, compat, bad);
> > > + if (!v || compat != 1)
> > > + goto bad;
> > > + /* FIXME: sanity check? */
> > > + ceph_decode_32_safe(p, end, len, bad);
> > > + /* type is __le32, so we must copy into place as-is */
> > > + ceph_decode_copy_safe(p, end, &addr->type,
> > > + sizeof(addr->type), bad);
> > > +
> > > + /*
> > > + * TYPE_NONE == 0
> > > + * TYPE_LEGACY == 1
> > > + *
> > > + * Clients that don't support ADDR2 always send TYPE_NONE.
> > > + * For now, since all we support is msgr1, just set this to 0
> > > + * when we get a TYPE_LEGACY type.
> > > + */
> > > + if (addr->type == cpu_to_le32(1))
> > > + addr->type = 0;
> > > + } else if (marker == 0) {
> > > + addr->type = 0;
> > > + /* Skip rest of type field */
> > > + ceph_decode_skip_n(p, end, 3, bad);
> > > + } else {
> >
> > versioned encoding has forward compatibility. The code should looks like
> >
> > if (struct_v == 0) {
> > /* old format */
> > return;
> > }
> >
> > if (struct_compat != 1)
> > goto bad
> >
> > end = *p + struct_len;
> >
> > if (struct_v == 1) {
> > ....
> > }
> >
> > if (struct_v == 2) {
> > ...
> > }
> >
> > *p = end;
> >
> >
> >
> >
> > > + goto bad;
> > > + }
> > > +
> > > + ceph_decode_need(p, end, sizeof(addr->nonce), bad);
> > > + ceph_decode_copy(p, &addr->nonce, sizeof(addr->nonce));
> > > +
> > > + /* addr length */
> > > + if (marker == 1) {
> > > + ceph_decode_32_safe(p, end, len, bad);
> > > + if (len > sizeof(addr->in_addr))
> > > + goto bad;
> > > + } else {
> > > + len = sizeof(addr->in_addr);
> > > + }
> > > +
> > > + memset(&addr->in_addr, 0, sizeof(addr->in_addr));
> > > +
> > > + if (len) {
> > > + ceph_decode_need(p, end, len, bad);
> > > + ceph_decode_copy(p, &addr->in_addr, len);
> > > +
> > > + /*
> > > + * Fix up sa_family. Legacy encoding sends it in BE, addr2
> > > + * encoding uses LE.
> > > + */
> > > + if (marker == 1)
> > > + addr->in_addr.ss_family =
> > > + le16_to_cpu((__force __le16)addr->in_addr.ss_family);
> > > + else
> > > + addr->in_addr.ss_family =
> > > + be16_to_cpu((__force __be16)addr->in_addr.ss_family);
> > > + }
> > > + return 0;
> > > +bad:
> > > + return -EINVAL;
> > > +}
> > > +EXPORT_SYMBOL(ceph_decode_entity_addr);
> > > +
> > > --
> > > 2.21.0
>
>
> (Dropping dev@xxxxxxx from cc list since they evidently _really_ don't
> want to see kernel patches there)
>
> Something like this then on top of the original patch?
>
> SQUASH: address Zheng's comments
>
> Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> ---
> net/ceph/decode.c | 21 ++++++++++++---------
> 1 file changed, 12 insertions(+), 9 deletions(-)
>
> diff --git a/net/ceph/decode.c b/net/ceph/decode.c
> index 27edf5d341ec..5a008567d018 100644
> --- a/net/ceph/decode.c
> +++ b/net/ceph/decode.c
> @@ -5,18 +5,20 @@
> int
> ceph_decode_entity_addr(void **p, void *end, struct ceph_entity_addr *addr)
> {
> - u8 marker, v, compat;
> + u8 marker, struct_v, struct_compat;
> u32 len;
>
> ceph_decode_8_safe(p, end, marker, bad);
> if (marker == 1) {
> - ceph_decode_8_safe(p, end, v, bad);
> - ceph_decode_8_safe(p, end, compat, bad);
> - if (!v || compat != 1)
> + ceph_decode_8_safe(p, end, struct_v, bad);
> + ceph_decode_8_safe(p, end, struct_compat, bad);
> + if (!struct_v || struct_compat != 1)
> goto bad;
> +
> /* FIXME: sanity check? */
> ceph_decode_32_safe(p, end, len, bad);
> - /* type is __le32, so we must copy into place as-is */
> +
> + /* type is defined as __le32, copy into place as-is */
> ceph_decode_copy_safe(p, end, &addr->type,
> sizeof(addr->type), bad);
>
> @@ -32,17 +34,18 @@ ceph_decode_entity_addr(void **p, void *end, struct ceph_entity_addr *addr)
> addr->type = 0;
> } else if (marker == 0) {
> addr->type = 0;
> + struct_v = 0;
> + struct_compat = 0;
> /* Skip rest of type field */
> ceph_decode_skip_n(p, end, 3, bad);
> } else {
> goto bad;
> }
>
> - ceph_decode_need(p, end, sizeof(addr->nonce), bad);
> - ceph_decode_copy(p, &addr->nonce, sizeof(addr->nonce));
> + ceph_decode_copy_safe(p, end, &addr->nonce, sizeof(addr->nonce), bad);
>
> /* addr length */
> - if (marker == 1) {
> + if (struct_v > 0) {
> ceph_decode_32_safe(p, end, len, bad);
> if (len > sizeof(addr->in_addr))
> goto bad;
> @@ -60,7 +63,7 @@ ceph_decode_entity_addr(void **p, void *end, struct ceph_entity_addr *addr)
> * Fix up sa_family. Legacy encoding sends it in BE, addr2
> * encoding uses LE.
> */
> - if (marker == 1)
> + if (struct_v > 0)
> addr->in_addr.ss_family =
> le16_to_cpu((__force __le16)addr->in_addr.ss_family);
> else
> --
> 2.21.0
>
>
still missing code that updates (*p) at the very end.
if (struct_compat != 1)
goto bad
end = *p + struct_len;
...
*p = end;
I think It's better to define separate functions for legacy encoding
and new format.
