- 22 Sep, 2006 40 commits
-
-
David S. Miller authored
First of all, if the xfrm_state only gets used for input packets this entropy is a complete waste. Secondly, it is often the case that a configuration loads many rules (perhaps even dynamically) and they don't all necessarily ever get used. This get_random_bytes() call was showing up in the profiles for xfrm_state inserts which is how I noticed this. Signed-off-by:David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Signed-off-by: -
David S. Miller authored
This idea is from Alexey Kuznetsov. It is common for policies to be non-prefixed. And for that case we can optimize lookups, insert, etc. quite a bit. For each direction, we have a dynamically sized policy hash table for non-prefixed policies. We also have a hash table on policy->index. For prefixed policies, we have a list per-direction which we will consult on lookups when a non-prefix hashtable lookup fails. This still isn't as efficient as I would like it. There are four immediate problems: 1) Lots of excessive refcounting, which can be fixed just like xfrm_state was 2) We do 2 hash probes on insert, one to look for dups and one to allocate a unique policy->index. Althought I wonder how much this matters since xfrm_state inserts do up to 3 hash probes and that seems to perform fine. 3) xfrm_policy_insert() is very complex because of the priority ordering and entry replacement logic. 4) Lots of counter bumping, in addition to policy refcounts, in the form of xfrm_policy_count[]. This is merely used to let code path(s) know that some IPSEC rules exist. So this count is indexed per-direction, maybe that is overkill. Signed-off-by: -
David S. Miller authored
The source address is always non-prefixed so we should use it to help give entropy to the bydst hash. Signed-off-by: -
David S. Miller authored
The refcounting done for timers and hash table insertions are just wasted cycles. We can eliminate all of this refcounting because: 1) The implicit refcount when the xfrm_state object is active will always be held while the object is in the hash tables. We never kfree() the xfrm_state until long after we've made sure that it has been unhashed. 2) Timers are even easier. Once we mark that x->km.state as anything other than XFRM_STATE_VALID (__xfrm_state_delete sets it to XFRM_STATE_DEAD), any timer that fires will do nothing and return without rearming the timer. Therefore we can defer the del_timer calls until when the object is about to be freed up during GC. We have to use del_timer_sync() and defer it to GC because we can't do a del_timer_sync() while holding x->lock which all callers of __xfrm_state_delete hold. This makes SA changes even more light-weight. Signed-off-by: -
David S. Miller authored
Just let GC and other normal mechanisms take care of getting rid of DST cache references to deleted xfrm_state objects instead of walking all the policy bundles. Signed-off-by: -
David S. Miller authored
Instead, simply set all potentially aliasing existing xfrm_state objects to have the current generation counter value. This will make routes get relooked up the next time an existing route mentioning these aliased xfrm_state objects gets used, via xfrm_dst_check(). Signed-off-by: -
David S. Miller authored
It can use __xfrm{4,6}_addr_hash(). Signed-off-by: -
David S. Miller authored
Besides the daddr, key the hash on family and reqid too. Signed-off-by: -
David S. Miller authored
Each xfrm_state inserted gets a new generation counter value. When a bundle is created, the xfrm_dst objects get the current generation counter of the xfrm_state they will attach to at dst->xfrm. xfrm_bundle_ok() will return false if it sees an xfrm_dst with a generation count different from the generation count of the xfrm_state that dst points to. This provides a facility by which to passively and cheaply invalidate cached IPSEC routes during SA database changes. Signed-off-by: -
David S. Miller authored
The grow algorithm is simple, we grow if: 1) we see a hash chain collision at insert, and 2) we haven't hit the hash size limit (currently 1*1024*1024 slots), and 3) the number of xfrm_state objects is > the current hash mask All of this needs some tweaking. Remove __initdata from "hashdist" so we can use it safely at run time. Signed-off-by: -
David S. Miller authored
Signed-off-by: -
David S. Miller authored
Signed-off-by: -
David S. Miller authored
Signed-off-by: -
Masahide NAKAMURA authored
Support Mobile IPv6 extension headers sorting for two transformation policies. Mobile IPv6 extension headers should be placed after IPsec transport mode, but before transport AH when outbound. Signed-off-by:
Masahide NAKAMURA <nakam@linux-ipv6.org> Signed-off-by:
YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> Signed-off-by:
-
Masahide NAKAMURA authored
Add sort functions to combine templates/states for IPsec. Think of outbound transformation order we should be careful with transport AH which must be the last of all transport ones. Signed-off-by:
-
Masahide NAKAMURA authored
Sub policy can be used through netlink socket. PF_KEY uses main only and it is TODO to support sub. Signed-off-by:
-
Masahide NAKAMURA authored
Under two transformation policies it is required to merge them. This is a platform to sort state for outbound and templates for inbound respectively. It will be used when Mobile IPv6 and IPsec are used at the same time. Signed-off-by:
-
Masahide NAKAMURA authored
Sub policy is introduced. Main and sub policy are applied the same flow. (Policy that current kernel uses is named as main.) It is required another transformation policy management to keep IPsec and Mobile IPv6 lives separate. Policy which lives shorter time in kernel should be a sub i.e. normally main is for IPsec and sub is for Mobile IPv6. (Such usage as two IPsec policies on different database can be used, too.) Limitation or TODOs: - Sub policy is not supported for per socket one (it is always inserted as main). - Current kernel makes cached outbound with flowi to skip searching database. However this patch makes it disabled only when "two policies are used and the first matched one is bypass case" because neither flowi nor bundle information knows about transformation template size. Signed-off-by:
-
Masahide NAKAMURA authored
Add Kconfig to support sub policy. Signed-off-by:
-
Masahide NAKAMURA authored
Ignore to report user-space for known mobility headers rejected by destination options header transformation. Mobile IPv6 specification (RFC3775) says that mobility header is used with destination options header carrying home address option only for binding update message. Other type message cannot be used and node must drop it silently (and must not send binding error) if receving such packet. To achieve it, (1) application should use transformation policy and wild-card states to catch binding update message prior other packets (2) kernel doesn't report the reject to user-space not to send binding error message by application. This patch is for (2). Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
-
Masahide NAKAMURA authored
Report to user-space when home address option is rejected. In receiving this message user-space application will send Mobile IPv6 binding error. It is rate-limited by kernel. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
-
Masahide NAKAMURA authored
XFRM_MSG_REPORT is a message as notification of state protocol and selector from kernel to user-space. Mobile IPv6 will use it when inbound reject is occurred at route optimization to make user-space know a binding error requirement. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
For Mobile IPv6 usage, it is required to trace which secpath state is reject factor in order to notify it to user space (to know the address which cannot be used route optimized communication). Based on MIPL2 kernel patch. This patch was also written by: Henrik Petander <petander@tcs.hut.fi> Signed-off-by:
-
Masahide NAKAMURA authored
Transformation support mobility header. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
Mobility header is built by user-space and sent through raw socket. Kernel just extracts its type to flow. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
Like ICMPv6, mobility header is handled through raw socket. In inbound case, check only whether ICMPv6 error should be sent as a reply or not by kernel. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> This patch was also written by: Antti Tuominen <anttit@tcs.hut.fi> Signed-off-by:
-
Masahide NAKAMURA authored
Add Mobility header definition for Mobile IPv6. Based on MIPL2 kernel patch. This patch was also written by: Antti Tuominen <anttit@tcs.hut.fi> Signed-off-by:
-
Masahide NAKAMURA authored
Add Mobile IPv6 route optimization protocols to netlink interface. Route optimization states carry care-of address. Based on MIPL2 kernel patch. Signed-off-by:
-
Noriaki TAKAMIYA authored
Add destination options header transformation for Mobile IPv6. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
Noriaki TAKAMIYA <takamiya@po.ntts.co.jp> Signed-off-by:
-
Noriaki TAKAMIYA authored
Add routing header type 2 transformation for Mobile IPv6. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
Mobile IPv6 defines home address option as an option of destination options header. It is placed before fragment header then ip6_find_1stfragopt() is fixed to know about it. Home address option also carries final source address of the flow, then outbound AH calculation should take care of it like routing header case. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
IPv6 source address is replaced in receiving packet with home address option carried by destination options header. To send ICMPv6 error back, original address which is received one on wire should be used. This function checks such header is included and reverts them. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
-
David S. Miller authored
Signed-off-by: -
Masahide NAKAMURA authored
Add inbound function of home address option by registering it to TLV table for destination options header. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
-
Noriaki TAKAMIYA authored
Add home address option definition for Mobile IPv6. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
In receiving Mobile IPv6 home address option which is a TLV carried by destination options header, kernel will try to mangle source adderss of packet. Think of cloned skbuff it is required to replace it by the parser just like routing header case. This is a framework to achieve that to allow TLV parser to replace inbound skbuff pointer. Signed-off-by:
-
Masahide NAKAMURA authored
This is a helper to search option offset from extension header which can carry TLV option like destination options header. Mobile IPv6 home address option will use it. Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
Add socket option and ancillary data interface of routing header type 2. Mobile IPv6 application will use this to send binding acknowledgement with the header without relation of confirmed route optimization (binding). Based on MIPL2 kernel patch. Signed-off-by:
-
Masahide NAKAMURA authored
Add inbound interface of routing header type 2 for Mobile IPv6. Based on MIPL2 kernel patch. This patch was also written by: Ville Nuorvala <vnuorval@tcs.hut.fi> Signed-off-by:
-
