On Sun, May 14, 2023 at 08:14:06PM +0200, Stefano Brivio wrote:
...
  Fixes: fde8004ab0b4 ("netlink: Use 8 KiB *
netlink message header size as response buffer")
 Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> 
Reviewed-by: David Gibson <david(a)gibson.dropbear.id.au>

Makes sense regardless of the rest of the series.

...
  ---
  netlink.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
 
 diff --git a/netlink.c b/netlink.c
 index b99af85..c07a13c 100644
 --- a/netlink.c
 +++ b/netlink.c
 @@ -99,7 +99,7 @@ fail:
  /**
   * nl_req() - Send netlink request and read response
   * @ns:		Use netlink socket in namespace
 - * @buf:	Buffer for response (at least BUFSIZ long)
 + * @buf:	Buffer for response (at least NLBUFSIZ long)
   * @req:	Request with netlink header
   * @len:	Request length
   * 
-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

On Sun, May 14, 2023 at 08:14:07PM +0200, Stefano Brivio wrote:
...
  In pasta_wait_for_ns(), open() failing with ENOENT is
expected: we're
 busy-looping until the network namespace appears. But any other
 failure is not something we're going to recover from: return right
 away if we don't get either success or ENOENT.
 
 Now that pasta_wait_for_ns() can actually fail, handle that in
 pasta_start_ns() by reporting the issue and exiting.
 
 Looping on EPERM, when pasta doesn't actually have the permissions to
 join a given namespace, isn't exactly a productive thing to do.
 
 Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> 
Reviewed-by: David Gibson <david(a)gibson.dropbear.id.au>

Also makes sense regardless of the rest of the series.

...
  ---
  pasta.c | 9 +++++++--
  1 file changed, 7 insertions(+), 2 deletions(-)
 
 diff --git a/pasta.c b/pasta.c
 index 3a4d704..2fa0168 100644
 --- a/pasta.c
 +++ b/pasta.c
 @@ -94,8 +94,11 @@ static int pasta_wait_for_ns(void *arg)
  	char ns[PATH_MAX];
  
  	snprintf(ns, PATH_MAX, "/proc/%i/ns/net", pasta_child_pid);
 -	do
 -		while ((c->pasta_netns_fd = open(ns, flags)) < 0);
 +	while ((c->pasta_netns_fd = open(ns, flags)) < 0) {
 +		if (errno != ENOENT)
 +			return 0;
 +	}
 +
  	while (setns(c->pasta_netns_fd, CLONE_NEWNET) &&
  	       !close(c->pasta_netns_fd));
  
 @@ -252,6 +255,8 @@ void pasta_start_ns(struct ctx *c, uid_t uid, gid_t gid,
  	}
  
  	NS_CALL(pasta_wait_for_ns, c);
 +	if (c->pasta_netns_fd < 0)
 +		die("Failed to join network namespace");
  }
  
  /** 
-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

On Sun, May 14, 2023 at 08:14:11PM +0200, Stefano Brivio wrote:
...
  This reverts commit
7656a6f8888237b9e23d63666e921528b6aaf950: now, by
 default, we copy all the routes associated to the outbound interface
 into the routing table of the container, so there's no need for this
 horrible workaround anymore.
 
 Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> 
Reviewed-by: David Gibson <david(a)gibson.dropbear.id.au>

This was never the right solution - even without the other
considerations raised by this series we probably should have
advertized an explicit host route to the gateway instead of mangling
the netmask.

...
  ---
  conf.c | 25 +------------------------
  1 file changed, 1 insertion(+), 24 deletions(-)
 
 diff --git a/conf.c b/conf.c
 index c2a745e..3a2fc2d 100644
 --- a/conf.c
 +++ b/conf.c
 @@ -634,9 +634,6 @@ static int conf_ip4_prefix(const char *arg)
  static unsigned int conf_ip4(unsigned int ifi,
  			     struct ip4_ctx *ip4, unsigned char *mac)
  {
 -	in_addr_t addr, gw;
 -	int shift;
 -
  	if (!ifi)
  		ifi = nl_get_ext_if(AF_INET);
  
 @@ -651,10 +648,8 @@ static unsigned int conf_ip4(unsigned int ifi,
  	if (IN4_IS_ADDR_UNSPECIFIED(&ip4->addr))
  		nl_addr(0, ifi, AF_INET, &ip4->addr, &ip4->prefix_len, NULL);
  
 -	addr = ntohl(ip4->addr.s_addr);
 -	gw = ntohl(ip4->gw.s_addr);
 -
  	if (!ip4->prefix_len) {
 +		in_addr_t addr = ntohl(ip4->addr.s_addr);
  		if (IN_CLASSA(addr))
  			ip4->prefix_len = (32 - IN_CLASSA_NSHIFT);
  		else if (IN_CLASSB(addr))
 @@ -665,24 +660,6 @@ static unsigned int conf_ip4(unsigned int ifi,
  			ip4->prefix_len = 32;
  	}
  
 -	/* We might get an address with a netmask that makes the default
 -	 * gateway unreachable, and in that case we would fail to configure
 -	 * the default route, with --config-net, or presumably a DHCP client
 -	 * in the guest or container would face the same issue.
 -	 *
 -	 * The host might have another route, to the default gateway itself,
 -	 * fixing the situation, but we only read default routes.
 -	 *
 -	 * Fix up the mask to allow reaching the default gateway from our
 -	 * configured address, if needed, and only if we find a non-zero
 -	 * mask that makes the gateway reachable.
 -	 */
 -	shift = 32 - ip4->prefix_len;
 -	while (shift < 32 && addr >> shift != gw >> shift)
 -		shift++;
 -	if (shift < 32)
 -		ip4->prefix_len = 32 - shift;
 -
  	memcpy(&ip4->addr_seen, &ip4->addr, sizeof(ip4->addr_seen));
  
  	if (MAC_IS_ZERO(mac)) 
-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

On Tue, May 16, 2023 at 11:42:09PM +0200, Stefano Brivio wrote:
...
  On Tue, 16 May 2023 15:06:29 +1000
 David Gibson <david(a)gibson.dropbear.id.au> wrote:
 
  On Sun, May 14, 2023 at 08:14:05PM +0200, Stefano
Brivio wrote:
  This series, along with pseudo-related fixes,
enables:
 
 - optional copy of all routes from selected interface in outer
   namespace, to (hopefully!) fix the issue reported by Callum at:
     https://github.com/containers/podman/issues/18539
 
 - optional copy of all addresses, mostly for consistency. It doesn't,
   however, enable assignment of multiple addresses in the sense
   requested at:
     https://bugs.passt.top/show_bug.cgi?id=47
 
   because the addresses still need to be present on the host, and
   the "outer" address isn't selected depending on the address used
   inside the container
 
 - operation without a gateway address, to (again, hopefully) support
   usage of Wireguard endpoints established outside the container,
     https://bugs.passt.top/show_bug.cgi?id=49
 
 I tested the single functionalities introduced here, but I didn't
 try to reproduce the setups where the issues were reported, so some
 help with testing is definitely fundamental here. Thanks.   
 
 I've sent reviews for some of the simpler patches in this series which
 make sense even without the context of the overall aim.  I think those
 can be applied immediately. 
 
 Those are actually the least important patches for users 
Well, granted.

...
  -- and I can't
 apply 6/10 without breaking Podman's CI plus probably a number of
 deployments (that's why it comes after 5/10)... so, no, I would rather
 not apply the rest for the moment. 
Uh.. true, 6/10 is problematic, but I think the other easy ones could
be applied safely enough.

...
 
  For the rest
of the series, I want to address the generalities before
 doing detailed review of the implementation.
 
 I think the basic idea here is sound: we want to expose anything
 routable to the host as routable to the guest, even when the host has
 a more complex routing setup that just a netmask on the "main"
 interface and a default gateway within that prefix. 
 
 The intentions behind this series are actually slightly different:
 
 - we have a complete breakage in a seemingly common use case (I would
   even say cloud-init setups in general), and I'd like to fix that
   sooner rather than later 
Well, sure, but we should at least think about where we're going with
this longer term, so we don't box ourselves in.

...
  - this concerns only the direct configuration pasta
does, with
   --config-net. What we advertise is definitely related, but not the
   same topic... to the point that the issues fixed by this series don't
   even occur with a DHCP client:
     https://github.com/containers/podman/issues/18539#issuecomment-1545023424 
Ah, interesting.  It looks like dhclient (or rather dhclient-script, I
expect) is adding an explicit /32 route to the default gateway.  It
seems to me the best quick fix for --config-net is to do the same
thing.  Basically rather than expanding the netmask as we did in 6/10,
if the gateway address is not in the interface's netmask add a /32 or
/128 route to the gateway.

...
    And, in general, we can't advertise everything
we can configure (say,
   a route without router over DHCP). 
Ah, true.  The DHCP options for static routes are even more limited
than I realized.  Ok, that nixes option B.3.

...
    I'd be much more careful about what we
advertise. We have direct
   control of what we configure via netlink, but for DHCP, NDP, DHCPv6,
   we need to think of possible interpretations and common half-bugs as
   well.
 
  But I think we want to think a bit more deeply
about exactly what we
 need/want to expose here.
 
 Even with the current code, the default gateway address we advertise
 to the guest is kind of meaningless: the guest cannot directly access
 that gateway, everything really goes through passt on the host. 
 
 In the simplest, probably most common network setups, that's actually
 the gateway that connects our guest to other nodes. 
I don't understand what you mean by this.  Yes, we have the same IP
for the gateway that the host sees, but the NAT to host means that we
can't even talk to the gateway at L4.  Literally the only thing the
guest kernel will do with that gateway address is put it into ARP and
neighbour discovery packets, which passt will resolve to its own MAC,
like nearly every other IP.

...
  For other cases, I think we should eventually
implement
 https://bugs.passt.top/show_bug.cgi?id=47 anyway, and it goes without
 saying that, then, we can't just use the same host route no matter what
 the container chooses. We'll need to match them. 
Oh.. I'm wondering if I've been confusing by using "host route" in
two
different ways: one being "a route taken from the passt host system"
and the other meaning "a route to a single network host, that is /32
or /128".

I agree that we should move to allowing multiple IPs on the guest
side, but I don't see how that conflicts with the routing issue here.

...
  I mean, I'm not saying that the behaviour from
this series is complete
 and self-consistent, just that it works around obvious, urgent issues
 and at the same time it looks like we'll probably need something
 similar to support further use cases. 
Adding a /32 or /128 route to the gateway seems a simpler way to do
that to me.  Plus it matches the behaviour that DHCP seems to be doing
anyway.

...
 
  This works
because the gateway address (like everything) will ARP/NDP
 to passt's host side MAC address and once the packets hit passt it
 doesn't matter what the guest thought the routing was going to be.
 
 I think we have a few choices in two more-or-less orthogonal
 categories.
 
 A) What routable prefixes do we advertise to the guest?
 
   A.1) Always a default route (0.0.0.0/0 and ::/0)
 
 We tell the guest that every address is routable via the passt
 interface, regardless of routing setup on the host.  This essentially
 tells the guest to delegate all routing responsibility to passt.
 
 Advantages:
   * Simple
   * No need to update anything if routing configuration on the host
     changes
 Disadvantages:
   * If addresses are unroutable from the host, the guest will only
     know via ICMP/ICMPv6, rather than statically, which may be a worse
     UX on the guest side.  Plus we might need to actually implement
     those host unreachable ICMPs.
   * Might be messy if the guest has multiple interfacees - e.g. if we
     allow passt to be configured to attach to a specific host
     interface only, then we have multiple passts attached to a single
     guest: they'd all be advertising a default route.
 
   A.2) Copy routable prefixes from the host to the guest 
 
 I'm having a hard time figuring out the definition of this point. How
 would you define that? Strictly speaking, in the case at hand, nothing
 is routable: we have a /32 address. 
Right.. which means that if the host is working, it must have an
additional static route - also probably /32 - telling it how to get to
the gateway.  Indeed I can see it in the bug, initial comment:
    172.31.1.1 dev ens3 proto static scope link metric 100
With A.2 we'd copy that route to the guest - or at least one with the
same prefix (which is a single address in this case).

...
 
  We just
advertise those prefixes routable to the host to the guest
 (which might include an empty prefix == default route).
 
 Advantages:
   * Guest statically knows what addresses are routable via the passt
     interface
 Disadvantages:
   * What do we do with overlapping prefixes?  On the host we might
     have more specific routes pointing to a specific interface.  For
     the guest they all point to the passt interface, so what's the
     point?
   * Can we advertise an arbitrary set of static routes via all our
     mechanisms (--config-net, DHCP, NDP+DHCPv6)?  Even if we can it
     adds more complexity to that code
   * How do we update things if the host routing configuration changes?
   * What do we do if the host has source-based routing or other
     advanced stuff set up?
 
 B) What gateway, if any, do we advertise for each route?
 
   B.1) Copy it from the host
 
 Advantages:
  * Guest L3 configuration resembles that of the host 
 
 ...which is a fundamental design goal of passt: transparency, and
 pretending it doesn't exist. Otherwise we can have a route, a bridge,
 an interface, etc. 
Well... we want to be transparent for anything visible at L4.  For
things only visible at L3 - like routes, it's not possible for things
to look 100% identical, so I think we have some wiggle room in exactly
what we do.

...
  Now, while there are use cases that rely on different
aspects of this
 transparency (KubeVirt and service mesh integration) I understand this
 might sound a bit dogmatic, because you might say there are more
 important use cases (which I'm not aware of) or supposed benefits.
 
 What's far less dogmatic, though, is how many issues we happily and
 automatically avoid by relying on the sanity of the host networking
 configuration.
 
 By trying to copy it as close as possible, we avoid one very important
 source of issues, which is our interpretation or possible lack of
 knowledge about how applications we don't know about chose to interact
 with kernel and network setups. The main case fixed by this series
 shows exactly that: I think it's broken, but it works, and users
 expect it to work.
 
 And by trusting the host configuration we don't lose much: if that's
 broken, almost everything else is broken anyway. 
It's not a question of "trust" in the host configuration, it's the
fact that parts of the host configuration don't make sense in the
guest's context.  Most obviously the interface names from the host
routes can't be used in the guest.  We can and do use the same
addresses for the routers, but what does it really mean?  The guest
can't actually contact them as neighbours - when it tries they just
ARP to passt's fake MAC and the packets get routed by the host kernel
regardless of what router the guest was trying to send them to - in
fact neither passt nor the host kernel will even know what router the
guest thought it was using.

...
 
 
Disadvantages:
  * If the host route doesn't have a gateway we have to fall back on
    B.2 or B.3 anyway 
 
 Well, they are a particular case of B.1 then: what's the disadvantage? 
Two cases is more complex than one.

...
  This is consistent (especially with this series, and
especially if we
 start adapting the *default* behaviours in this sense).
 
   * Misleading: in fact everything is routed by
passt and the host
    before it reaches any gateway we're listing here 
 
 But passt isn't supposed to be a router...? Let's say we have multiple
 routes on the host, we configure or advertise multiple routes to the
 guest. Does that make passt a router? I don't think so: we're just
 associating them as closely as possible, without fancy interpretations.
 
 A router has its own routing table, passt's would simply be a copy.
 Right now it has essentially none. 
Sorry, by "passt" here I really meant the host kernel, which
absolutely will route the packets.  There's no guarantee they'll even
go next to the router the guest thought it was using, although it's
likely.

...
 
    B.2) Pick an
address to represent passt as gateway
 
 Advantages:
  * Accurately represents that everything is routed by passt 
 
 This is configurable, actually, but no, I insist that passt isn't
 *functionally* routing anything, or at least that we should get as
 close as possible to that. 
Again, the host kernel definitely will, and there's no avoiding that.

...
 
   * We can make
this the same as the NAT-to-host address, so we only
    have one "magic" address (per AF) 
 
 Not really, if it's configurable. 
I mean one per passt instance, not one globally.  As opposed to the
gateway address and the NAT-to-host address being potentially
different magic addresses in a single instance.

...
 
 
Disadvantages:
  * Have to allocate an address that's safe, which is tricky (but we
    usually want this for NAT-to-host anyway) 
 
 There's a difference between picking an address by default and letting
 the user configure one. Besides, at least for IPv4, I don't think such
 an address exists. 
There certainly isn't one we can use everywhere.  I think we have some
options for probing one that will be safe in a particular case.

...
 
   * Do we want
just one address, or one for each distinct gateway from
    the host?
  * If we can't pick something in the interfaces "natural" prefix, we
    will also need to advertise a static route to reach it.
 
   B.3) Don't advertise a gateway for any route
 
 passt essentially proxy ARPs for the entire internet.
 
 Advantages:
   * No need to allocate an address - in fact passt need not have any
     guest facing IP at all
   * Extends naturally if we ever have a guest<->passt transport that's
     point-to-point rather than pseudo-ethernet
 Disadvantages:
   * Guest ARP / neighbour tables could get real big 
 
 ...it would also break a number of applications that peek at netlink
 (or do ioctl()s) to check they are in fact online. 
Uh.. what exactly are they looking at?  We'd still have at least one
route, they just wouldn't have gateways attached to them.  But as you
pointed out above I don't think we can do this with DHCP, which pretty
much kills it anyway.

...
 
  The status quo
is, roughly, A.1+B.1, except that we also enforce that
 the host must have a default route, which sidesteps one of the
 complications of B.1.  IIUC, this series is implementing A.2+B.1.
 
 Thinking about it, I'm moderately convinced that B.1 is a bad idea.
 I'm leaning towards B.2 - combining it with the NAT-to-host cleanups
 to have a more concrete guest-visible address for passt itself - but
 I'm also open to B.3. 
 
 ...that, especially B.3, sounds like another tool, or at least like
 another mode, because it conflicts quite a bit with design goals. 
...
  They're different from design _choices_ in the
sense that that's what
 I've been "selling" to users and what I and others have been
 implementing in integrations so far. 
So the ways L4 transparency are valuable (including guest address) are
pretty clear to me.  Are there also cases where the (partial) L3
transparency matter?  They're certainly not obvious to me

...
 
  I'm not
sure about A.1 vs. A.2.  I was leaning towards A.2, but on
 further consideration, I feel like the fact that A.1 automatically
 works for routing changes on the host might outweigh the fact that he
 guest only gets limited information (ICMP) about what's routable. 
 
 I don't think A.2 is doable, 
?? AFAICT this series is doing A.2

...
  but even if it were, yes, I don't think
 it would be worth the effort. If needed (and I never saw a request in
 this sense), we could enrich ICMP/ICMPv6 handling guest- or
 container-side quite a bit. 
-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

On Wed, May 17, 2023 at 08:52:50AM +0200, Stefano Brivio wrote:
...
  On Wed, 17 May 2023 11:15:06 +1000
 David Gibson <david(a)gibson.dropbear.id.au> wrote:
 
  On Tue, May 16, 2023 at 11:42:09PM +0200, Stefano
Brivio wrote:
  On Tue, 16 May 2023 15:06:29 +1000
 David Gibson <david(a)gibson.dropbear.id.au> wrote:
   
  On Sun, May 14, 2023 at 08:14:05PM +0200, Stefano
Brivio wrote:  
 > This series, along with pseudo-related fixes, enables:
 > 
 > - optional copy of all routes from selected interface in outer
 >   namespace, to (hopefully!) fix the issue reported by Callum at:
 >     https://github.com/containers/podman/issues/18539
 > 
 > - optional copy of all addresses, mostly for consistency. It doesn't,
 >   however, enable assignment of multiple addresses in the sense
 >   requested at:
 >     https://bugs.passt.top/show_bug.cgi?id=47
 > 
 >   because the addresses still need to be present on the host, and
 >   the "outer" address isn't selected depending on the address used
 >   inside the container
 > 
 > - operation without a gateway address, to (again, hopefully) support
 >   usage of Wireguard endpoints established outside the container,
 >     https://bugs.passt.top/show_bug.cgi?id=49
 > 
 > I tested the single functionalities introduced here, but I didn't
 > try to reproduce the setups where the issues were reported, so some
 > help with testing is definitely fundamental here. Thanks.    
 
 I've sent reviews for some of the simpler patches in this series which
 make sense even without the context of the overall aim.  I think those
 can be applied immediately.   
 
 Those are actually the least important patches for users   
 
 Well, granted.
 
  -- and I can't
 apply 6/10 without breaking Podman's CI plus probably a number of
 deployments (that's why it comes after 5/10)... so, no, I would rather
 not apply the rest for the moment.   
 
 Uh.. true, 6/10 is problematic, but I think the other easy ones could
 be applied safely enough. 
 
 With two hands and (worryingly close to) just 24 hours in a day, I
 honestly can't picture even a quick rebase and retest for those being a
 priority, while happily keeping around the issue that 5/10 fixes. 
Yeah, ok, fair enough.  This sort of thing is, of course, why I'm
spending so much effort trying to improve the testing situation.

...
 
 
 
  For the rest
of the series, I want to address the generalities before
 doing detailed review of the implementation.
 
 I think the basic idea here is sound: we want to expose anything
 routable to the host as routable to the guest, even when the host has
 a more complex routing setup that just a netmask on the "main"
 interface and a default gateway within that prefix.   
 
 The intentions behind this series are actually slightly different:
 
 - we have a complete breakage in a seemingly common use case (I would
   even say cloud-init setups in general), and I'd like to fix that
   sooner rather than later   
 
 Well, sure, but we should at least think about where we're going with
 this longer term, so we don't box ourselves in. 
 
 I don't think this is going to "box us in" -- I'm just proposing to
 change this after about two years, and we can definitely change it
 again, as long as things keep working. If keeping things working is
 boxing us in, well, I can't see that as a bad thing.
 
 That is, I don't see people doing screen-scraping of 'ip route show'
 and writing applications around that, and surely not adapting
 applications to what pasta does. Not at the moment, and surely not for
 a long while. 
That's a good point - but also supports what I'm saying below about
having some wiggle room about what we do at L3.

...
 
 
  - this concerns only the direct configuration pasta
does, with
   --config-net. What we advertise is definitely related, but not the
   same topic... to the point that the issues fixed by this series don't
   even occur with a DHCP client:
     https://github.com/containers/podman/issues/18539#issuecomment-1545023424   
 
 Ah, interesting.  It looks like dhclient (or rather dhclient-script, I
 expect) is adding an explicit /32 route to the default gateway.  It
 seems to me the best quick fix for --config-net is to do the same
 thing.  Basically rather than expanding the netmask as we did in 6/10,
 if the gateway address is not in the interface's netmask add a /32 or
 /128 route to the gateway. 
 
 That's the first option I considered (of course!):
   https://github.com/containers/podman/issues/18539#issuecomment-1545260780
   https://github.com/containers/podman/issues/18539#issuecomment-1546967377
 
 and only as I started implementing it, I realised that we can have
 anyway chained dependencies which aren't that easy to handle, especially
 if we admit an arbitrary number of routes and we need to sort them. 
I'm not understanding where these dependencies are coming from.  I'm
not proposing copying anything from the host here, what I'm suggesting
is that --config-net do the equivalent of:

	if ($gw not in <ifi's built in prefix>)
		ip route add $gw/32 dev $ifname

...
  Plus it's going to be 1. more code 2. actually
"complicated" code. This
 is stupidly simple instead.
 
 I have some experience of fixing IPv6 FIB code in kernel with
 consequences on sorting/selection, and there are just so many hidden
 details involved in interpretation of Linux-style routes and ways of
 shadowing them. 
Right, this sounds like more reason to be nervous of blindly copying
routes from the host: it's hard to know how we'll have to tweak them
to make sense in the guest context.

...
 
 
    And, in general, we can't advertise everything
we can configure (say,
   a route without router over DHCP).   
 
 Ah, true.  The DHCP options for static routes are even more limited
 than I realized.  Ok, that nixes option B.3.
 
    I'd be much more careful about what we
advertise. We have direct
   control of what we configure via netlink, but for DHCP, NDP, DHCPv6,
   we need to think of possible interpretations and common half-bugs as
   well.
   
  But I think we want to think a bit more deeply
about exactly what we
 need/want to expose here.
 
 Even with the current code, the default gateway address we advertise
 to the guest is kind of meaningless: the guest cannot directly access
 that gateway, everything really goes through passt on the host.   
 
 In the simplest, probably most common network setups, that's actually
 the gateway that connects our guest to other nodes.   
 
 I don't understand what you mean by this.  Yes, we have the same IP
 for the gateway that the host sees, but the NAT to host means that we
 can't even talk to the gateway at L4. 
 
 It's disabled by default in Podman. It's the default behaviour in
 passt because this started from KubeVirt and that's what they expect,
 but that's about it. Once the address is configurable, this is not a
 valid point, in general.
 
 A gateway doesn't need to be a host, and it's very often, functionally,
 not a host. This is by design: RFC 791, 2.2:
 
   In a gateway the higher level protocols need not be implemented and
   the GGP functions are added to the IP module.
 
  Literally the only thing the guest kernel will do
with that gateway
 address is put it into ARP and neighbour discovery packets, which passt
 will resolve to its own MAC, like nearly every other IP. 
 
 No, the guest kernel might also have netfilter rules, specifying that
 gateway address, that were originally designed for the host, as if guest
 and passt didn't exist. Those might happily use the gateway address to
 represent the notion of gateway. 
Ok... trying to improve my understanding of this case.

If we are using NAT to host on the gateway address, that would
silently change the semantics of ip filter rules using that address,
which doesn't seem great.

If we're not using map-gw, and the rules are about filtering traffic
going to the gateway in it's (possible) capacity as just another host,
then those rules are still valid, regardless of whether that address
is still the gateway in the guest.

Are you considering a case where a guest-side script is building the
iptables rules based on the gateway address (which it retrieves from
ip route show, or dhcp or something)?  Or something else?


...
 
 
  For other cases, I think we should eventually
implement
 https://bugs.passt.top/show_bug.cgi?id=47 anyway, and it goes without
 saying that, then, we can't just use the same host route no matter what
 the container chooses. We'll need to match them.   
 
 Oh.. I'm wondering if I've been confusing by using "host route" in
two
 different ways: one being "a route taken from the passt host system"
 and the other meaning "a route to a single network host, that is /32
 or /128".
 
 I agree that we should move to allowing multiple IPs on the guest
 side, but I don't see how that conflicts with the routing issue here. 
 
 It's related because, once we allow them, different host routes should
 actually be used, so having them in the guest/container too,
 aiming at a 1:1 mapping, should simplify things rather than become
 misleading. 
But whatever routers we set in the guest, they'll all ARP back to the
same thing, so I'm not sure what difference it makes.

...
 
 
  I mean, I'm not saying that the behaviour from
this series is complete
 and self-consistent, just that it works around obvious, urgent issues
 and at the same time it looks like we'll probably need something
 similar to support further use cases.   
 
 Adding a /32 or /128 route to the gateway seems a simpler way to do
 that to me.  Plus it matches the behaviour that DHCP seems to be doing
 anyway. 
 
 For the reason why it's not really simple, see above. About what DHCP
 clients do: that's in general not the case for udhcpc, dhcpcd, pump, or
 NetworkManager. See also e1c94637ad50 ("dhcp: Send option 121 if the
 default gateway is not on the assigned subnet"). 
Oh, I'd missed that option 121 stuff.  I'm guessing dhclient(-script)
is actually setting the route in response to that, rather than it
being built in.  But now I'm even more confused.. what I'm suggesting
is basically the exact equivalent logic for config-net, so I'm not
sure why you're objecting.

[As an aside, option 121 explicitly allows advertising local routes,
which means B.3 is possible with DHCP after all - althouh maybe not
wise, since I gather option 121 isn't necessarily widely supported]

...
  As far as I know, that's just what the script
used in conjunction with
 ISC's dhclient does on _some_ distributions.
 
 
 
  This works because the gateway address (like
everything) will ARP/NDP
 to passt's host side MAC address and once the packets hit passt it
 doesn't matter what the guest thought the routing was going to be.
 
 I think we have a few choices in two more-or-less orthogonal
 categories.
 
 A) What routable prefixes do we advertise to the guest?
 
   A.1) Always a default route (0.0.0.0/0 and ::/0)
 
 We tell the guest that every address is routable via the passt
 interface, regardless of routing setup on the host.  This essentially
 tells the guest to delegate all routing responsibility to passt.
 
 Advantages:
   * Simple
   * No need to update anything if routing configuration on the host
     changes
 Disadvantages:
   * If addresses are unroutable from the host, the guest will only
     know via ICMP/ICMPv6, rather than statically, which may be a worse
     UX on the guest side.  Plus we might need to actually implement
     those host unreachable ICMPs.
   * Might be messy if the guest has multiple interfacees - e.g. if we
     allow passt to be configured to attach to a specific host
     interface only, then we have multiple passts attached to a single
     guest: they'd all be advertising a default route.
 
   A.2) Copy routable prefixes from the host to the guest   
 
 I'm having a hard time figuring out the definition of this point. How
 would you define that? Strictly speaking, in the case at hand, nothing
 is routable: we have a /32 address.   
 
 Right.. which means that if the host is working, it must have an
 additional static route - also probably /32 - telling it how to get to
 the gateway.  Indeed I can see it in the bug, initial comment:
     172.31.1.1 dev ens3 proto static scope link metric 100
 With A.2 we'd copy that route to the guest - or at least one with the
 same prefix (which is a single address in this case). 
 
 Blindly copying routes is one thing. Figuring out what subnets are
 routable and which ones aren't is a different matter, and _that_ is
 what I don't consider generally feasible. 
Oh, sorry, I was unclear.  By "routable" I meant just "has a matching
entry in the host's routes" not any deeper analysis of whether they're
actually reachable.  So blindly(ish) copying routes is exactly what
I'm suggesting for A.2

...
 
 
 
  We just
advertise those prefixes routable to the host to the guest
 (which might include an empty prefix == default route).
 
 Advantages:
   * Guest statically knows what addresses are routable via the passt
     interface
 Disadvantages:
   * What do we do with overlapping prefixes?  On the host we might
     have more specific routes pointing to a specific interface.  For
     the guest they all point to the passt interface, so what's the
     point?
   * Can we advertise an arbitrary set of static routes via all our
     mechanisms (--config-net, DHCP, NDP+DHCPv6)?  Even if we can it
     adds more complexity to that code
   * How do we update things if the host routing configuration changes?
   * What do we do if the host has source-based routing or other
     advanced stuff set up?
 
 B) What gateway, if any, do we advertise for each route?
 
   B.1) Copy it from the host
 
 Advantages:
  * Guest L3 configuration resembles that of the host   
 
 ...which is a fundamental design goal of passt: transparency, and
 pretending it doesn't exist. Otherwise we can have a route, a bridge,
 an interface, etc.   
 
 Well... we want to be transparent for anything visible at L4.  For
 things only visible at L3 - like routes, it's not possible for things
 to look 100% identical, so I think we have some wiggle room in exactly
 what we do. 
 
 With this series, in most cases, things will actually be 100% identical.
 
 But no, the transparency design goal applies especially to L3:
   https://passt.top/passt/about/#motivation
 
 let alone ALGs, which are probably less common nowadays -- but with
 service meshes (increasingly common), L3 transparency is very helpful
 to have. Otherwise libslirp would be absolutely enough from that
 perspective. 
I'm still not seeing what concrete examples need transparency to the
*routes* rather than just the addresses at L3.

...
 
 
  Now, while there are use cases that rely on different
aspects of this
 transparency (KubeVirt and service mesh integration) I understand this
 might sound a bit dogmatic, because you might say there are more
 important use cases (which I'm not aware of) or supposed benefits.
 
 What's far less dogmatic, though, is how many issues we happily and
 automatically avoid by relying on the sanity of the host networking
 configuration.
 
 By trying to copy it as close as possible, we avoid one very important
 source of issues, which is our interpretation or possible lack of
 knowledge about how applications we don't know about chose to interact
 with kernel and network setups. The main case fixed by this series
 shows exactly that: I think it's broken, but it works, and users
 expect it to work.
 
 And by trusting the host configuration we don't lose much: if that's
 broken, almost everything else is broken anyway.   
 
 It's not a question of "trust" in the host configuration, it's the
 fact that parts of the host configuration don't make sense in the
 guest's context.  Most obviously the interface names from the host
 routes can't be used in the guest. 
 
 By default, with containers, even interface names are copied. Indices,
 of course, aren't, but that's not something users or applications
 typically try to fiddle with. 
Huh?  If the interface name is anything other than the pasta template
interface, it won't even exist in the guest.  So how can you possibly
copy it?

...
 
  We can and do
use the same
 addresses for the routers, but what does it really mean?  The guest
 can't actually contact them as neighbours - when it tries they just
 ARP to passt's fake MAC and the packets get routed by the host kernel
 regardless of what router the guest was trying to send them to - in
 fact neither passt nor the host kernel will even know what router the
 guest thought it was using. 
 
 This is an L2 matter, which was never a problem for any project using
 this. 
It's really not, we'll absolutely hit the IP stack for routing, not
just the datalink layer in the host kernel.

...
 
 
 
 
Disadvantages:
  * If the host route doesn't have a gateway we have to fall back on
    B.2 or B.3 anyway   
 
 Well, they are a particular case of B.1 then: what's the disadvantage?   
 
 Two cases is more complex than one. 
 
 ...but, with this series, we don't implement two different cases...?
 
 
  This is
consistent (especially with this series, and especially if we
 start adapting the *default* behaviours in this sense).
   
   * Misleading: in fact everything is routed by
passt and the host
    before it reaches any gateway we're listing here   
 
 But passt isn't supposed to be a router...? Let's say we have multiple
 routes on the host, we configure or advertise multiple routes to the
 guest. Does that make passt a router? I don't think so: we're just
 associating them as closely as possible, without fancy interpretations.
 
 A router has its own routing table, passt's would simply be a copy.
 Right now it has essentially none.   
 
 Sorry, by "passt" here I really meant the host kernel, which
 absolutely will route the packets.  There's no guarantee they'll even
 go next to the router the guest thought it was using, although it's
 likely. 
 
 Right. I'd just try to make it as likely as possible. That doesn't come
 with this series, and, for instance, me(a)yawnt.com already checked and
 told me this series isn't enough for the "regular" Wireguard case (with
 the endpoint in the outer namespace), but this clearly appears to be
 getting closer to what we'll need to "naturally" support that.
 
 
 
    B.2) Pick an address to represent passt as gateway
 
 Advantages:
  * Accurately represents that everything is routed by passt   
 
 This is configurable, actually, but no, I insist that passt isn't
 *functionally* routing anything, or at least that we should get as
 close as possible to that.   
 
 Again, the host kernel definitely will, and there's no avoiding that.
 
 
   * We can
make this the same as the NAT-to-host address, so we only
    have one "magic" address (per AF)   
 
 Not really, if it's configurable.   
 
 I mean one per passt instance, not one globally.  As opposed to the
 gateway address and the NAT-to-host address being potentially
 different magic addresses in a single instance.
 
 
 
Disadvantages:
  * Have to allocate an address that's safe, which is tricky (but we
    usually want this for NAT-to-host anyway)   
 
 There's a difference between picking an address by default and letting
 the user configure one. Besides, at least for IPv4, I don't think such
 an address exists.   
 
 There certainly isn't one we can use everywhere.  I think we have some
 options for probing one that will be safe in a particular case. 
 
 Well... that doesn't sound great. 
Of course.  We're trying to allocate scarce IP space out of nowhere,
it's always going to be messy.

...
 
 
 
   * Do we want
just one address, or one for each distinct gateway from
    the host?
  * If we can't pick something in the interfaces "natural" prefix, we
    will also need to advertise a static route to reach it.
 
   B.3) Don't advertise a gateway for any route
 
 passt essentially proxy ARPs for the entire internet.
 
 Advantages:
   * No need to allocate an address - in fact passt need not have any
     guest facing IP at all
   * Extends naturally if we ever have a guest<->passt transport that's
     point-to-point rather than pseudo-ethernet
 Disadvantages:
   * Guest ARP / neighbour tables could get real big   
 
 ...it would also break a number of applications that peek at netlink
 (or do ioctl()s) to check they are in fact online.   
 
 Uh.. what exactly are they looking at?  We'd still have at least one
 route, they just wouldn't have gateways attached to them.  But as you
 pointed out above I don't think we can do this with DHCP, which pretty
 much kills it anyway.
 
 
  The
status quo is, roughly, A.1+B.1, except that we also enforce that
 the host must have a default route, which sidesteps one of the
 complications of B.1.  IIUC, this series is implementing A.2+B.1.
 
 Thinking about it, I'm moderately convinced that B.1 is a bad idea.
 I'm leaning towards B.2 - combining it with the NAT-to-host cleanups
 to have a more concrete guest-visible address for passt itself - but
 I'm also open to B.3.   
 
 ...that, especially B.3, sounds like another tool, or at least like
 another mode, because it conflicts quite a bit with design goals.   
 
  They're different from design _choices_ in
the sense that that's what
 I've been "selling" to users and what I and others have been
 implementing in integrations so far.   
 
 So the ways L4 transparency are valuable (including guest address) are
 pretty clear to me.  Are there also cases where the (partial) L3
 transparency matter?  They're certainly not obvious to me 
 
 Absolutely, and I thought I explained this a number of times, but...
 service meshes using netfilter. Applications being moved from "host" to
 containers, or from containers to VMs. That's where we want to pretend
 nothing changes, and that usually causes more L3 headaches rather than
 L4. 
I'm clear on how NAT breaks everything.  But I'm still not seeing how
the list of routes matters.

...
 
 
 
  I'm not
sure about A.1 vs. A.2.  I was leaning towards A.2, but on
 further consideration, I feel like the fact that A.1 automatically
 works for routing changes on the host might outweigh the fact that he
 guest only gets limited information (ICMP) about what's routable.   
 
 I don't think A.2 is doable,   
 
 ?? AFAICT this series is doing A.2 
 
 Not really, we're not figuring out routable prefixes, just blindly
 copying routing entries. It's closer to A.2 than A.1, but it's not
 that, either. 
Again, the only "figuring out of routable prefixes" I was suggesting
for A.2 is "is there an entry in the host routing table".
i.e. blindly copying route entries.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson