Message ID | 20210121220044.22361-2-justin.iurman@uliege.be (mailing list archive) |
---|---|
State | Accepted |
Commit | 07d46d93c9acdfe0614071d73c415dd5f745cc6e |
Delegated to: | Netdev Maintainers |
Headers | show |
Series | Fix big endian definition of ipv6_rpl_sr_hdr | expand |
Context | Check | Description |
---|---|---|
netdev/cover_letter | success | Link |
netdev/fixes_present | fail | Series targets non-next tree, but doesn't contain any Fixes tags |
netdev/patch_count | success | Link |
netdev/tree_selection | success | Clearly marked for net |
netdev/subject_prefix | success | Link |
netdev/cc_maintainers | success | CCed 3 of 3 maintainers |
netdev/source_inline | success | Was 0 now: 0 |
netdev/verify_signedoff | success | Link |
netdev/module_param | success | Was 0 now: 0 |
netdev/build_32bit | success | Errors and warnings before: 0 this patch: 0 |
netdev/kdoc | success | Errors and warnings before: 0 this patch: 0 |
netdev/verify_fixes | success | Link |
netdev/checkpatch | success | total: 0 errors, 0 warnings, 0 checks, 13 lines checked |
netdev/build_allmodconfig_warn | success | Errors and warnings before: 0 this patch: 0 |
netdev/header_inline | success | Link |
netdev/stable | success | Stable not CCed |
On Thu, 21 Jan 2021 23:00:44 +0100 Justin Iurman wrote: > Following RFC 6554 [1], the current order of fields is wrong for big > endian definition. Indeed, here is how the header looks like: > > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > | Next Header | Hdr Ext Len | Routing Type | Segments Left | > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > | CmprI | CmprE | Pad | Reserved | > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > > This patch reorders fields so that big endian definition is now correct. > > [1] https://tools.ietf.org/html/rfc6554#section-3 > > Signed-off-by: Justin Iurman <justin.iurman@uliege.be> Are you sure? This looks right to me. > diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h > index 1dccb55cf8c6..708adddf9f13 100644 > --- a/include/uapi/linux/rpl.h > +++ b/include/uapi/linux/rpl.h > @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { > pad:4, > reserved1:16; > #elif defined(__BIG_ENDIAN_BITFIELD) > - __u32 reserved:20, > + __u32 cmpri:4, > + cmpre:4, > pad:4, > - cmpri:4, > - cmpre:4; > + reserved:20; > #else > #error "Please fix <asm/byteorder.h>" > #endif
> De: "Jakub Kicinski" <kuba@kernel.org> > À: "Justin Iurman" <justin.iurman@uliege.be> > Cc: netdev@vger.kernel.org, davem@davemloft.net, "alex aring" <alex.aring@gmail.com> > Envoyé: Dimanche 24 Janvier 2021 05:54:44 > Objet: Re: [PATCH net 1/1] uapi: fix big endian definition of ipv6_rpl_sr_hdr > On Thu, 21 Jan 2021 23:00:44 +0100 Justin Iurman wrote: >> Following RFC 6554 [1], the current order of fields is wrong for big >> endian definition. Indeed, here is how the header looks like: >> >> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >> | Next Header | Hdr Ext Len | Routing Type | Segments Left | >> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >> | CmprI | CmprE | Pad | Reserved | >> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >> >> This patch reorders fields so that big endian definition is now correct. >> >> [1] https://tools.ietf.org/html/rfc6554#section-3 >> >> Signed-off-by: Justin Iurman <justin.iurman@uliege.be> > > Are you sure? This looks right to me. AFAIK, yes. Did you mean the old (current) one looks right, or the new one? If you meant the old/current one, well, I don't understand why the big endian definition would look like this: #elif defined(__BIG_ENDIAN_BITFIELD) __u32 reserved:20, pad:4, cmpri:4, cmpre:4; When the RFC defines the header as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | CmprI | CmprE | Pad | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The little endian definition looks fine. But, when it comes to big endian, you define fields as you see them on the wire with the same order, right? So the current big endian definition makes no sense. It looks like it was a wrong mix with the little endian conversion. >> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h >> index 1dccb55cf8c6..708adddf9f13 100644 >> --- a/include/uapi/linux/rpl.h >> +++ b/include/uapi/linux/rpl.h >> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { >> pad:4, >> reserved1:16; >> #elif defined(__BIG_ENDIAN_BITFIELD) >> - __u32 reserved:20, >> + __u32 cmpri:4, >> + cmpre:4, >> pad:4, >> - cmpri:4, >> - cmpre:4; >> + reserved:20; >> #else >> #error "Please fix <asm/byteorder.h>" > > #endif
On 1/24/21 2:57 AM, Justin Iurman wrote: >> De: "Jakub Kicinski" <kuba@kernel.org> >> À: "Justin Iurman" <justin.iurman@uliege.be> >> Cc: netdev@vger.kernel.org, davem@davemloft.net, "alex aring" <alex.aring@gmail.com> >> Envoyé: Dimanche 24 Janvier 2021 05:54:44 >> Objet: Re: [PATCH net 1/1] uapi: fix big endian definition of ipv6_rpl_sr_hdr > >> On Thu, 21 Jan 2021 23:00:44 +0100 Justin Iurman wrote: >>> Following RFC 6554 [1], the current order of fields is wrong for big >>> endian definition. Indeed, here is how the header looks like: >>> >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>> | Next Header | Hdr Ext Len | Routing Type | Segments Left | >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>> | CmprI | CmprE | Pad | Reserved | >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>> >>> This patch reorders fields so that big endian definition is now correct. >>> >>> [1] https://tools.ietf.org/html/rfc6554#section-3 >>> >>> Signed-off-by: Justin Iurman <justin.iurman@uliege.be> >> >> Are you sure? This looks right to me. > > AFAIK, yes. Did you mean the old (current) one looks right, or the new one? If you meant the old/current one, well, I don't understand why the big endian definition would look like this: > > #elif defined(__BIG_ENDIAN_BITFIELD) > __u32 reserved:20, > pad:4, > cmpri:4, > cmpre:4; > > When the RFC defines the header as follows: > > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > | CmprI | CmprE | Pad | Reserved | > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > > The little endian definition looks fine. But, when it comes to big endian, you define fields as you see them on the wire with the same order, right? So the current big endian definition makes no sense. It looks like it was a wrong mix with the little endian conversion. > >>> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h >>> index 1dccb55cf8c6..708adddf9f13 100644 >>> --- a/include/uapi/linux/rpl.h >>> +++ b/include/uapi/linux/rpl.h >>> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { >>> pad:4, >>> reserved1:16; >>> #elif defined(__BIG_ENDIAN_BITFIELD) >>> - __u32 reserved:20, >>> + __u32 cmpri:4, >>> + cmpre:4, >>> pad:4, >>> - cmpri:4, >>> - cmpre:4; >>> + reserved:20; >>> #else >>> #error "Please fix <asm/byteorder.h>" >>> #endif cross-checking with other headers - tcp and vxlan-gpe - this patch looks correct.
On Sun, 24 Jan 2021 11:57:03 -0700 David Ahern wrote: > On 1/24/21 2:57 AM, Justin Iurman wrote: > >> De: "Jakub Kicinski" <kuba@kernel.org> > >> À: "Justin Iurman" <justin.iurman@uliege.be> > >> Cc: netdev@vger.kernel.org, davem@davemloft.net, "alex aring" <alex.aring@gmail.com> > >> Envoyé: Dimanche 24 Janvier 2021 05:54:44 > >> Objet: Re: [PATCH net 1/1] uapi: fix big endian definition of ipv6_rpl_sr_hdr > > > >> On Thu, 21 Jan 2021 23:00:44 +0100 Justin Iurman wrote: > >>> Following RFC 6554 [1], the current order of fields is wrong for big > >>> endian definition. Indeed, here is how the header looks like: > >>> > >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > >>> | Next Header | Hdr Ext Len | Routing Type | Segments Left | > >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > >>> | CmprI | CmprE | Pad | Reserved | > >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > >>> > >>> This patch reorders fields so that big endian definition is now correct. > >>> > >>> [1] https://tools.ietf.org/html/rfc6554#section-3 > >>> > >>> Signed-off-by: Justin Iurman <justin.iurman@uliege.be> > >> > >> Are you sure? This looks right to me. > > > > AFAIK, yes. Did you mean the old (current) one looks right, or the new one? Old one / existing is correct. > > If you meant the old/current one, well, I don't understand why the big endian definition would look like this: > > > > #elif defined(__BIG_ENDIAN_BITFIELD) > > __u32 reserved:20, > > pad:4, > > cmpri:4, > > cmpre:4; > > > > When the RFC defines the header as follows: > > > > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > > | CmprI | CmprE | Pad | Reserved | > > +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > > > > The little endian definition looks fine. But, when it comes to big endian, you define fields as you see them on the wire with the same order, right? So the current big endian definition makes no sense. It looks like it was a wrong mix with the little endian conversion. Well, you don't list the bit positions in the quote from the RFC, and I'm not familiar with the IETF parlor. I'm only comparing the LE definition with the BE. If you claim the BE is wrong, then the LE is wrong, too. > >>> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h > >>> index 1dccb55cf8c6..708adddf9f13 100644 > >>> --- a/include/uapi/linux/rpl.h > >>> +++ b/include/uapi/linux/rpl.h > >>> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { > >>> pad:4, > >>> reserved1:16; > >>> #elif defined(__BIG_ENDIAN_BITFIELD) > >>> - __u32 reserved:20, > >>> + __u32 cmpri:4, > >>> + cmpre:4, > >>> pad:4, > >>> - cmpri:4, > >>> - cmpre:4; > >>> + reserved:20; > >>> #else > >>> #error "Please fix <asm/byteorder.h>" > >>> #endif > > cross-checking with other headers - tcp and vxlan-gpe - this patch looks > correct. What are you cross-checking?
On 1/25/21 12:32 PM, Jakub Kicinski wrote: >>>>> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h >>>>> index 1dccb55cf8c6..708adddf9f13 100644 >>>>> --- a/include/uapi/linux/rpl.h >>>>> +++ b/include/uapi/linux/rpl.h >>>>> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { >>>>> pad:4, >>>>> reserved1:16; >>>>> #elif defined(__BIG_ENDIAN_BITFIELD) >>>>> - __u32 reserved:20, >>>>> + __u32 cmpri:4, >>>>> + cmpre:4, >>>>> pad:4, >>>>> - cmpri:4, >>>>> - cmpre:4; >>>>> + reserved:20; >>>>> #else >>>>> #error "Please fix <asm/byteorder.h>" >>>>> #endif >> >> cross-checking with other headers - tcp and vxlan-gpe - this patch looks >> correct. > > What are you cross-checking? > https://tools.ietf.org/html/draft-ietf-nvo3-vxlan-gpe-10, Section 3.1 header definition and vxlanhdr_gpe in include/net/vxlan.h. The __BIG_ENDIAN_BITFIELD part follows the definition in the spec. Similarly for the TCP header - RFC header definition and tcphdr in include/uapi/linux/tcp.h. TCP header shows doff + res1 order which is comparable to cmpri + cpmre in this header as both sets are 4-bits and start a word.
On Mon, 25 Jan 2021 13:12:42 -0700 David Ahern wrote: > On 1/25/21 12:32 PM, Jakub Kicinski wrote: > >>>>> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h > >>>>> index 1dccb55cf8c6..708adddf9f13 100644 > >>>>> --- a/include/uapi/linux/rpl.h > >>>>> +++ b/include/uapi/linux/rpl.h > >>>>> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { > >>>>> pad:4, > >>>>> reserved1:16; > >>>>> #elif defined(__BIG_ENDIAN_BITFIELD) > >>>>> - __u32 reserved:20, > >>>>> + __u32 cmpri:4, > >>>>> + cmpre:4, > >>>>> pad:4, > >>>>> - cmpri:4, > >>>>> - cmpre:4; > >>>>> + reserved:20; > >>>>> #else > >>>>> #error "Please fix <asm/byteorder.h>" > >>>>> #endif > >> > >> cross-checking with other headers - tcp and vxlan-gpe - this patch looks > >> correct. > > > > What are you cross-checking? > > > > https://tools.ietf.org/html/draft-ietf-nvo3-vxlan-gpe-10, Section 3.1 > header definition and vxlanhdr_gpe in include/net/vxlan.h. The > __BIG_ENDIAN_BITFIELD part follows the definition in the spec. > > Similarly for the TCP header - RFC header definition and tcphdr in > include/uapi/linux/tcp.h. TCP header shows doff + res1 order which is > comparable to cmpri + cpmre in this header as both sets are 4-bits and > start a word. Ack, thanks for the pointers. The LE definition is broken as well, then, right?
> Le 25 janv. 2021 à 20:32, Jakub Kicinski <kuba@kernel.org> a écrit : > >> On Sun, 24 Jan 2021 11:57:03 -0700 David Ahern wrote: >> On 1/24/21 2:57 AM, Justin Iurman wrote: >>>> De: "Jakub Kicinski" <kuba@kernel.org> >>>> À: "Justin Iurman" <justin.iurman@uliege.be> >>>> Cc: netdev@vger.kernel.org, davem@davemloft.net, "alex aring" <alex.aring@gmail.com> >>>> Envoyé: Dimanche 24 Janvier 2021 05:54:44 >>>> Objet: Re: [PATCH net 1/1] uapi: fix big endian definition of ipv6_rpl_sr_hdr >>> >>>>> On Thu, 21 Jan 2021 23:00:44 +0100 Justin Iurman wrote: >>>>> Following RFC 6554 [1], the current order of fields is wrong for big >>>>> endian definition. Indeed, here is how the header looks like: >>>>> >>>>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>>>> | Next Header | Hdr Ext Len | Routing Type | Segments Left | >>>>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>>>> | CmprI | CmprE | Pad | Reserved | >>>>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>>>> >>>>> This patch reorders fields so that big endian definition is now correct. >>>>> >>>>> [1] https://tools.ietf.org/html/rfc6554#section-3 >>>>> >>>>> Signed-off-by: Justin Iurman <justin.iurman@uliege.be> >>>> >>>> Are you sure? This looks right to me. >>> >>> AFAIK, yes. Did you mean the old (current) one looks right, or the new one? > > Old one / existing is correct. > >>> If you meant the old/current one, well, I don't understand why the big endian definition would look like this: >>> >>> #elif defined(__BIG_ENDIAN_BITFIELD) >>> __u32 reserved:20, >>> pad:4, >>> cmpri:4, >>> cmpre:4; >>> >>> When the RFC defines the header as follows: >>> >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>> | CmprI | CmprE | Pad | Reserved | >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ >>> >>> The little endian definition looks fine. But, when it comes to big endian, you define fields as you see them on the wire with the same order, right? So the current big endian definition makes no sense. It looks like it was a wrong mix with the little endian conversion. > > Well, you don't list the bit positions in the quote from the RFC, and > I'm not familiar with the IETF parlor. I'm only Indeed, sorry for that. Bit positions are available if you follow the link to the RFC I referenced in the patch. It is always defined as network byte order by default (=BE). > comparing the LE > definition with the BE. If you claim the BE is wrong, then the LE is > wrong, too. Actually, no, it’s not. If you have a look at the header definition from the RFC, you can see that the LE is correct (valid translation from BE, the *new* BE in this patch). >>>>> diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h >>>>> index 1dccb55cf8c6..708adddf9f13 100644 >>>>> --- a/include/uapi/linux/rpl.h >>>>> +++ b/include/uapi/linux/rpl.h >>>>> @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { >>>>> pad:4, >>>>> reserved1:16; >>>>> #elif defined(__BIG_ENDIAN_BITFIELD) >>>>> - __u32 reserved:20, >>>>> + __u32 cmpri:4, >>>>> + cmpre:4, >>>>> pad:4, >>>>> - cmpri:4, >>>>> - cmpre:4; >>>>> + reserved:20; >>>>> #else >>>>> #error "Please fix <asm/byteorder.h>" >>>>> #endif >> >> cross-checking with other headers - tcp and vxlan-gpe - this patch looks >> correct. > > What are you cross-checking?
On Mon, 25 Jan 2021 22:11:35 +0100 (CET) Justin Iurman wrote: > >>> If you meant the old/current one, well, I don't understand why the big endian definition would look like this: > >>> > >>> #elif defined(__BIG_ENDIAN_BITFIELD) > >>> __u32 reserved:20, > >>> pad:4, > >>> cmpri:4, > >>> cmpre:4; > >>> > >>> When the RFC defines the header as follows: > >>> > >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > >>> | CmprI | CmprE | Pad | Reserved | > >>> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > >>> > >>> The little endian definition looks fine. But, when it comes to big endian, you define fields as you see them on the wire with the same order, right? So the current big endian definition makes no sense. It looks like it was a wrong mix with the little endian conversion. > > > > Well, you don't list the bit positions in the quote from the RFC, and > > I'm not familiar with the IETF parlor. I'm only > > Indeed, sorry for that. Bit positions are available if you follow the link to the RFC I referenced in the patch. It is always defined as network byte order by default (=BE). > > > comparing the LE > > definition with the BE. If you claim the BE is wrong, then the LE is > > wrong, too. > > Actually, no, it’s not. If you have a look at the header definition from the RFC, you can see that the LE is correct (valid translation from BE, the *new* BE in this patch). Sigh, I see it now. Thanks!
diff --git a/include/uapi/linux/rpl.h b/include/uapi/linux/rpl.h index 1dccb55cf8c6..708adddf9f13 100644 --- a/include/uapi/linux/rpl.h +++ b/include/uapi/linux/rpl.h @@ -28,10 +28,10 @@ struct ipv6_rpl_sr_hdr { pad:4, reserved1:16; #elif defined(__BIG_ENDIAN_BITFIELD) - __u32 reserved:20, + __u32 cmpri:4, + cmpre:4, pad:4, - cmpri:4, - cmpre:4; + reserved:20; #else #error "Please fix <asm/byteorder.h>" #endif
Following RFC 6554 [1], the current order of fields is wrong for big endian definition. Indeed, here is how the header looks like: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Header | Hdr Ext Len | Routing Type | Segments Left | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | CmprI | CmprE | Pad | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ This patch reorders fields so that big endian definition is now correct. [1] https://tools.ietf.org/html/rfc6554#section-3 Signed-off-by: Justin Iurman <justin.iurman@uliege.be> --- include/uapi/linux/rpl.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)