diff mbox series

ethdev: fix variable length flow elements support

Message ID 20211026090554.9128-1-getelson@nvidia.com (mailing list archive)
State Superseded, archived
Delegated to: Ferruh Yigit
Headers
Series ethdev: fix variable length flow elements support |

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/github-robot: build success github build: passed
ci/iol-broadcom-Functional success Functional Testing PASS
ci/iol-broadcom-Performance success Performance Testing PASS
ci/iol-x86_64-unit-testing success Testing PASS
ci/iol-x86_64-compile-testing success Testing PASS
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-intel-Functional success Functional Testing PASS
ci/iol-aarch64-compile-testing success Testing PASS
ci/Intel-compilation success Compilation OK
ci/intel-Testing success Testing PASS
ci/iol-aarch64-unit-testing success Testing PASS

Commit Message

Gregory Etelson Oct. 26, 2021, 9:05 a.m. UTC 
  RTE flow API defines two flow items and actions types - common
and PMD private. Common RTE flow types are defined in rte_flow.h
while PMD private types restricted to specific PMD only.
RTE flow API allows PMD private types in flow rule,
but it must not try to interpret private item or acton properties.

Current implementation tried to locate PMD private element, item or
action, in common flow elements records.

The patch restricts access to common flow elements records for
non-private PMD types only.

Fixes: 6cf72047332b ("ethdev: support flow elements with variable length")

Signed-off-by: Gregory Etelson <getelson@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
---
 lib/ethdev/rte_flow.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Comments

Ferruh Yigit Nov. 3, 2021, 9:04 p.m. UTC | #1 
On 10/26/2021 10:05 AM, Gregory Etelson wrote:
> RTE flow API defines two flow items and actions types - common
> and PMD private. Common RTE flow types are defined in rte_flow.h
> while PMD private types restricted to specific PMD only.
> RTE flow API allows PMD private types in flow rule,
> but it must not try to interpret private item or acton properties.
> 
> Current implementation tried to locate PMD private element, item or
> action, in common flow elements records.
> 
> The patch restricts access to common flow elements records for
> non-private PMD types only.
> 
> Fixes: 6cf72047332b ("ethdev: support flow elements with variable length")
> 
> Signed-off-by: Gregory Etelson <getelson@nvidia.com>
> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
> ---
>   lib/ethdev/rte_flow.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/ethdev/rte_flow.c b/lib/ethdev/rte_flow.c
> index d268784532..a93f68abbc 100644
> --- a/lib/ethdev/rte_flow.c
> +++ b/lib/ethdev/rte_flow.c
> @@ -54,11 +54,13 @@ rte_flow_conv_copy(void *buf, const void *data, const size_t size,
>   	/**
>   	 * Allow PMD private flow item
>   	 */
> -	size_t sz = type >= 0 ? desc[type].size : sizeof(void *);
> +	bool rte_type = type >= 0;
> +
> +	size_t sz = rte_type ? desc[type].size : sizeof(void *);
>   	if (buf == NULL || data == NULL)
>   		return 0;
>   	rte_memcpy(buf, data, (size > sz ? sz : size));
> -	if (desc[type].desc_fn)

Was this (possible) negative array index intentional, or are you fixing it?

> +	if (rte_type && desc[type].desc_fn)
>   		sz += desc[type].desc_fn(size > 0 ? buf : NULL, data);
>   	return sz;
>   }
>
Gregory Etelson Nov. 4, 2021, 5:33 a.m. UTC | #2 
Hello Ferruh,

> >   lib/ethdev/rte_flow.c | 6 ++++--
> >   1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/lib/ethdev/rte_flow.c
> b/lib/ethdev/rte_flow.c
> > index d268784532..a93f68abbc 100644
> > --- a/lib/ethdev/rte_flow.c
> > +++ b/lib/ethdev/rte_flow.c
> > @@ -54,11 +54,13 @@
> rte_flow_conv_copy(void *buf, const void *data,
> const size_t size,
> >       /**
> >        * Allow PMD private flow item
> >        */
> > -     size_t sz = type >= 0 ? desc[type].size :
> sizeof(void *);
> > +     bool rte_type = type >= 0;
> > +
> > +     size_t sz = rte_type ? desc[type].size :
> sizeof(void *);
> >       if (buf == NULL || data == NULL)
> >               return 0;
> >       rte_memcpy(buf, data, (size > sz ? sz :
> size));
> > -     if (desc[type].desc_fn)
> 
> Was this (possible) negative array index
> intentional, or are you fixing it?
> 

Negative type values assigned to PMD private items and actions. 
RTE allows private PMD types in flow rules since 
5d1bff8fe2 ethdev: allow negative values in flow rule types
We construct flow rules with private PMD types
to implement tunnel offload.
However, negative type must not be used as index in
rte_flow_desc_item[] and rte_flow_desc_action[] arrays.

> > +     if (rte_type && desc[type].desc_fn)
> >               sz += desc[type].desc_fn(size > 0 ? buf
> : NULL, data);
> >       return sz;
> >   }
> >
Ferruh Yigit Nov. 4, 2021, 9:10 a.m. UTC | #3 
On 11/4/2021 5:33 AM, Gregory Etelson wrote:
> Hello Ferruh,
> 
>>>    lib/ethdev/rte_flow.c | 6 ++++--
>>>    1 file changed, 4 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/lib/ethdev/rte_flow.c
>> b/lib/ethdev/rte_flow.c
>>> index d268784532..a93f68abbc 100644
>>> --- a/lib/ethdev/rte_flow.c
>>> +++ b/lib/ethdev/rte_flow.c
>>> @@ -54,11 +54,13 @@
>> rte_flow_conv_copy(void *buf, const void *data,
>> const size_t size,
>>>        /**
>>>         * Allow PMD private flow item
>>>         */
>>> -     size_t sz = type >= 0 ? desc[type].size :
>> sizeof(void *);
>>> +     bool rte_type = type >= 0;
>>> +
>>> +     size_t sz = rte_type ? desc[type].size :
>> sizeof(void *);
>>>        if (buf == NULL || data == NULL)
>>>                return 0;
>>>        rte_memcpy(buf, data, (size > sz ? sz :
>> size));
>>> -     if (desc[type].desc_fn)
>>
>> Was this (possible) negative array index
>> intentional, or are you fixing it?
>>
> 
> Negative type values assigned to PMD private items and actions.
> RTE allows private PMD types in flow rules since
> 5d1bff8fe2 ethdev: allow negative values in flow rule types
> We construct flow rules with private PMD types
> to implement tunnel offload.
> However, negative type must not be used as index in
> rte_flow_desc_item[] and rte_flow_desc_action[] arrays.
> 

That is what I assumed, but wasn't clear from commit log, commit log
reads as it is updating an intended usage.

>>> +     if (rte_type && desc[type].desc_fn)
>>>                sz += desc[type].desc_fn(size > 0 ? buf
>> : NULL, data);
>>>        return sz;
>>>    }
>>>
>
Gregory Etelson Nov. 4, 2021, 9:20 a.m. UTC | #4 
Hello Ferruh,

..snip..
> >
> > Negative type values assigned to PMD private
> items and actions.
> > RTE allows private PMD types in flow rules
> since
> > 5d1bff8fe2 ethdev: allow negative values in
> flow rule types
> > We construct flow rules with private PMD
> types
> > to implement tunnel offload.
> > However, negative type must not be used as
> index in
> > rte_flow_desc_item[] and
> rte_flow_desc_action[] arrays.
> >
> 
> That is what I assumed, but wasn't clear from
> commit log, commit log
> reads as it is updating an intended usage.
> 

I'll update the commit message in v2.

> >>> +     if (rte_type && desc[type].desc_fn)
> >>>                sz += desc[type].desc_fn(size > 0 ?
> buf
> >> : NULL, data);
> >>>        return sz;
> >>>    }
> >>>
> >

Regards,
Gregory
diff mbox series

Patch

diff --git a/lib/ethdev/rte_flow.c b/lib/ethdev/rte_flow.c
index d268784532..a93f68abbc 100644
--- a/lib/ethdev/rte_flow.c
+++ b/lib/ethdev/rte_flow.c
@@ -54,11 +54,13 @@  rte_flow_conv_copy(void *buf, const void *data, const size_t size,
 	/**
 	 * Allow PMD private flow item
 	 */
-	size_t sz = type >= 0 ? desc[type].size : sizeof(void *);
+	bool rte_type = type >= 0;
+
+	size_t sz = rte_type ? desc[type].size : sizeof(void *);
 	if (buf == NULL || data == NULL)
 		return 0;
 	rte_memcpy(buf, data, (size > sz ? sz : size));
-	if (desc[type].desc_fn)
+	if (rte_type && desc[type].desc_fn)
 		sz += desc[type].desc_fn(size > 0 ? buf : NULL, data);
 	return sz;
 }