[PATCH 00/18] Test fixes, batch 5

[PATCH 00/18] Test fixes, batch 5

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 2245 bytes --]

Here's yet another batch of fixes to make the tests more robust
against different environments.  With this lot, I'm now able to run
the pasta, passt, passt_in_ns and two_guests tests on my Fedora
system.  I'm still hitting problems with the perf tests.

This series (specifically 11/18) updates the demo to use socat instead
of openbsd netcat.  This is needed, or the change to socat in the
mbuto image would break the demo.  However, I'm hitting unrelated
problems trying to run the demos, so the switch to socat is, alas,
untested.

David Gibson (18):
  tests: Remove no longer needed /usr/bin/bash link
  tests: Let Fedora find dhclient-script in /usr/sbin
  tests: Add rudimentary debugging to dhclient-script
  tests: Add some extra dhclient support directories to mbuto.img
  tests: More robust parsing of resolv.conf for DHCP tests
  tests: Handle the case of a nameserver on host localhost
  tests: Correctly handle domain search list in dhclient-script
  tests: Fix detection of empty 'hout' responses in passt{,_in_ns} tests
  tests: Fix creation of test file in udp passt tests
  valgrind needs futex
  tests: Use socat instead of netcat
  tests: Remove unnecessary ^D in passt_in_ns teardown
  tests: Remove unnecessary truncation of temporary files in udp tests
  tests: Use dhclient --no-pid for namespaces in two_guests tests
  tests: Clean up better after iperf tests
  tests: No need to retrieve host ifname in ndp/pasta
  tests: Correct determination of host interface name in tests
  demo: Use git protocol downloads

 Makefile              |   2 +-
 test/README.md        |   2 +-
 test/demo/passt       |  10 +--
 test/demo/pasta       |  14 ++---
 test/dhcp/passt       |  22 +++----
 test/lib/setup        |  24 ++++++--
 test/lib/test         |   2 +-
 test/ndp/pasta        |   5 +-
 test/passt.mbuto      |  18 +++---
 test/tcp/passt        |  36 +++++------
 test/tcp/passt_in_ns  | 138 +++++++++++++++++++++---------------------
 test/tcp/pasta        |  58 +++++++++---------
 test/two_guests/basic |  24 ++++----
 test/udp/passt        |  26 ++++----
 test/udp/passt_in_ns  |  82 +++++++++++--------------
 test/udp/pasta        |  34 +++++------
 16 files changed, 247 insertions(+), 250 deletions(-)

-- 
2.36.1

[PATCH 01/18] tests: Remove no longer needed /usr/bin/bash link

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 857 bytes --]

AFAICT the symlink we created in mbuto from /usr/bin/bash to /bin/sh was
for the benefit of a dhclient-script which used /usr/bin/bash as its
interpreter (e.g. in Fedora).  That was a bit risky if the script really
did require bash and we linked it to dash or another shell.

We now supply our own custom dhclient-script, so we don't need the
link any more.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/passt.mbuto | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/test/passt.mbuto b/test/passt.mbuto
index 968c1cd..2990020 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -19,8 +19,7 @@ KMODS="${KMODS:- virtio_net virtio_pci}"
 
 LINKS="${LINKS:-
 	 ash,dash,bash		/init
-	 ash,dash,bash		/bin/sh
-	 ash,dash,bash		/usr/bin/bash}"
+	 ash,dash,bash		/bin/sh}"
 
 DIRS="${DIRS} /tmp /sbin"
 
-- 
@@ -19,8 +19,7 @@ KMODS="${KMODS:- virtio_net virtio_pci}"
 
 LINKS="${LINKS:-
 	 ash,dash,bash		/init
-	 ash,dash,bash		/bin/sh
-	 ash,dash,bash		/usr/bin/bash}"
+	 ash,dash,bash		/bin/sh}"
 
 DIRS="${DIRS} /tmp /sbin"
 
-- 
2.36.1

[PATCH 02/18] tests: Let Fedora find dhclient-script in /usr/sbin

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 708 bytes --]

Modern Fedora (and RHEL) systems have /sbin as a symlink to /usr/sbin
(along with a number of similar links).  Along with that it expects to
find dhclient-script in /usr/sbin/dhclient-script rather than
/sbin/dhclient-script.

Link them together in our mbuto image so that the Fedora build of dhclient
can find it.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/passt.mbuto | 1 +
 1 file changed, 1 insertion(+)

diff --git a/test/passt.mbuto b/test/passt.mbuto
index 2990020..500c34f 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -41,6 +41,7 @@ FIXUP="${FIXUP}"'
 exit 0
 EOF
 	chmod 755 /sbin/dhclient-script
+	ln -s /sbin /usr/sbin
 	:> /etc/fstab
 	sh +m
 '
-- 
@@ -41,6 +41,7 @@ FIXUP="${FIXUP}"'
 exit 0
 EOF
 	chmod 755 /sbin/dhclient-script
+	ln -s /sbin /usr/sbin
 	:> /etc/fstab
 	sh +m
 '
-- 
2.36.1

[PATCH 03/18] tests: Add rudimentary debugging to dhclient-script

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 988 bytes --]

We now supply a minimal dhclient-script of our own in the mbuto boot image.
There are some problems with it, so add some basic logging to help debug
it.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/passt.mbuto | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/test/passt.mbuto b/test/passt.mbuto
index 500c34f..fb59cec 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -21,11 +21,15 @@ LINKS="${LINKS:-
 	 ash,dash,bash		/init
 	 ash,dash,bash		/bin/sh}"
 
-DIRS="${DIRS} /tmp /sbin"
+DIRS="${DIRS} /tmp /sbin /var/log"
 
 FIXUP="${FIXUP}"'
 	cat > /sbin/dhclient-script << EOF
 #!/bin/sh
+LOG=/var/log/dhclient-script.log
+echo \${reason} \${interface} >> \$LOG
+set >> \$LOG
+
 [ -n "\${new_interface_mtu}" ]       && ip link set dev \${interface} mtu \${new_interface_mtu}
 
 [ -n "\${new_ip_address}" ]          && ip addr add \${new_ip_address}/\${new_subnet_mask} dev \${interface}
-- 
@@ -21,11 +21,15 @@ LINKS="${LINKS:-
 	 ash,dash,bash		/init
 	 ash,dash,bash		/bin/sh}"
 
-DIRS="${DIRS} /tmp /sbin"
+DIRS="${DIRS} /tmp /sbin /var/log"
 
 FIXUP="${FIXUP}"'
 	cat > /sbin/dhclient-script << EOF
 #!/bin/sh
+LOG=/var/log/dhclient-script.log
+echo \${reason} \${interface} >> \$LOG
+set >> \$LOG
+
 [ -n "\${new_interface_mtu}" ]       && ip link set dev \${interface} mtu \${new_interface_mtu}
 
 [ -n "\${new_ip_address}" ]          && ip addr add \${new_ip_address}/\${new_subnet_mask} dev \${interface}
-- 
2.36.1

[PATCH 04/18] tests: Add some extra dhclient support directories to mbuto.img

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 712 bytes --]

Although it can operate without them, dhclient can issue errors if it
doesn't have /var/run to write a pid file and /var/lib to write a leases
file.  Create those in mbuto.img to stop it complaining.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/passt.mbuto | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/passt.mbuto b/test/passt.mbuto
index fb59cec..b33c0a8 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -21,7 +21,7 @@ LINKS="${LINKS:-
 	 ash,dash,bash		/init
 	 ash,dash,bash		/bin/sh}"
 
-DIRS="${DIRS} /tmp /sbin /var/log"
+DIRS="${DIRS} /tmp /sbin /var/log /var/run /var/lib"
 
 FIXUP="${FIXUP}"'
 	cat > /sbin/dhclient-script << EOF
-- 
@@ -21,7 +21,7 @@ LINKS="${LINKS:-
 	 ash,dash,bash		/init
 	 ash,dash,bash		/bin/sh}"
 
-DIRS="${DIRS} /tmp /sbin /var/log"
+DIRS="${DIRS} /tmp /sbin /var/log /var/run /var/lib"
 
 FIXUP="${FIXUP}"'
 	cat > /sbin/dhclient-script << EOF
-- 
2.36.1

[PATCH 05/18] tests: More robust parsing of resolv.conf for DHCP tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 3725 bytes --]

To check publishing of DNS information via DHCP, we need to extract a list
of nameservers and/or search domains from resolv.conf in the test script.

The current version (usually) leaves the result with a trailing ','.
That's usually ok because it happens on both guest and host sides.  However
it's kind of confusing, and might stop working if the host had a
resolv.conf without a trailing \n on the last line.  It also makes some
later changes we'll need more difficult.

So, normalize the output from resolv.conf a bit further, removing any
trailing ','.  It turns out we can do this with a slightly less complex
sed expression than the one we already have.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/dhcp/passt | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/test/dhcp/passt b/test/dhcp/passt
index 00047a9..a5491bc 100644
--- a/test/dhcp/passt
+++ b/test/dhcp/passt
@@ -35,15 +35,15 @@ gout	MTU ip -j link show | jq -rM '.[] | select(.ifname == "__IFNAME__").mtu'
 check	[ __MTU__ = 65520 ]
 
 test	DHCP: DNS
-gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__DNS__" = "__HOST_DNS__" ]
 
 # FQDNs should be terminated by dots, but the guest DHCP client might omit them:
 # strip them first
 test	DHCP: search list
-gout	SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__SEARCH__" = "__HOST_SEARCH__" ]
 
 test	DHCPv6: address
@@ -59,11 +59,11 @@ check	[ "__GW6__" = "__HOST_GW6__" ]
 
 # Strip interface specifier: interface names might differ between host and guest
 test	DHCPv6: DNS
-gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__DNS6__" = "__HOST_DNS6__" ]
 
 test	DHCPv6: search list
-gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__SEARCH6__" = "__HOST_SEARCH6__" ]
-- 
@@ -35,15 +35,15 @@ gout	MTU ip -j link show | jq -rM '.[] | select(.ifname == "__IFNAME__").mtu'
 check	[ __MTU__ = 65520 ]
 
 test	DHCP: DNS
-gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__DNS__" = "__HOST_DNS__" ]
 
 # FQDNs should be terminated by dots, but the guest DHCP client might omit them:
 # strip them first
 test	DHCP: search list
-gout	SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_SEARCH sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__SEARCH__" = "__HOST_SEARCH__" ]
 
 test	DHCPv6: address
@@ -59,11 +59,11 @@ check	[ "__GW6__" = "__HOST_GW6__" ]
 
 # Strip interface specifier: interface names might differ between host and guest
 test	DHCPv6: DNS
-gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__DNS6__" = "__HOST_DNS6__" ]
 
 test	DHCPv6: search list
-gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
-hout	HOST_SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/\(.*\)/\1\n/g'
+gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
+hout	HOST_SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
 check	[ "__SEARCH6__" = "__HOST_SEARCH6__" ]
-- 
2.36.1

[PATCH 06/18] tests: Handle the case of a nameserver on host localhost

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1856 bytes --]

We previously introduced a change to passt to handle the case where the
host machine is its own nameserver - so resolv.conf points to 127.0.0.1.
In this case we advertize the gateway as the DNS server for the guest,
which in turn will be redirected back to the host by existing passt logic.

The dhcp/passt doesn't handle this case correctly, so add some logic to
account for it.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/dhcp/passt | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/dhcp/passt b/test/dhcp/passt
index a5491bc..3d2e939 100644
--- a/test/dhcp/passt
+++ b/test/dhcp/passt
@@ -37,7 +37,7 @@ check	[ __MTU__ = 65520 ]
 test	DHCP: DNS
 gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/,$//;s/$/\n/'
-check	[ "__DNS__" = "__HOST_DNS__" ]
+check	[ "__DNS__" = "__HOST_DNS__" ] || [ "__DNS__" = "__HOST_GW__" -a "__HOST_DNS__" = "127.0.0.1" ]
 
 # FQDNs should be terminated by dots, but the guest DHCP client might omit them:
 # strip them first
@@ -61,7 +61,7 @@ check	[ "__GW6__" = "__HOST_GW6__" ]
 test	DHCPv6: DNS
 gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
-check	[ "__DNS6__" = "__HOST_DNS6__" ]
+check	[ "__DNS6__" = "__HOST_DNS6__" ] || [ "__DNS6__" = "__HOST_GW6__" -a "__HOST_DNS6__" = "::1" ]
 
 test	DHCPv6: search list
 gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
-- 
@@ -37,7 +37,7 @@ check	[ __MTU__ = 65520 ]
 test	DHCP: DNS
 gout	DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 hout	HOST_DNS sed -n 's/^nameserver \([0-9]*\.\)\(.*\)/\1\2/p' /etc/resolv.conf | head -n3 | tr '\n' ',' | sed 's/,$//;s/$/\n/'
-check	[ "__DNS__" = "__HOST_DNS__" ]
+check	[ "__DNS__" = "__HOST_DNS__" ] || [ "__DNS__" = "__HOST_GW__" -a "__HOST_DNS__" = "127.0.0.1" ]
 
 # FQDNs should be terminated by dots, but the guest DHCP client might omit them:
 # strip them first
@@ -61,7 +61,7 @@ check	[ "__GW6__" = "__HOST_GW6__" ]
 test	DHCPv6: DNS
 gout	DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
 hout	HOST_DNS6 sed -n 's/^nameserver \([^:]*:\)\([^%]*\).*/\1\2/p' /etc/resolv.conf | tr '\n' ',' | sed 's/,$//;s/$/\n/'
-check	[ "__DNS6__" = "__HOST_DNS6__" ]
+check	[ "__DNS6__" = "__HOST_DNS6__" ] || [ "__DNS6__" = "__HOST_GW6__" -a "__HOST_DNS6__" = "::1" ]
 
 test	DHCPv6: search list
 gout	SEARCH6 sed 's/\. / /g' /etc/resolv.conf | sed 's/\.$//g' | sed -n 's/^search \(.*\)/\1/p' | tr ' \n' ',' | sed 's/,$//;s/$/\n/'
-- 
2.36.1

[PATCH 07/18] tests: Correctly handle domain search list in dhclient-script

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1992 bytes --]

Currently our small custom dhclient-script only handles the 'domain-name'
option, which can just list a single domain, not the 'domain-search'
option, which can handle several.  Correct it to handle both.

We also weren't emptying the resolv.conf file before we began, which
could lead to surprising contents after multiple DHCP transactions.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/passt.mbuto | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/test/passt.mbuto b/test/passt.mbuto
index b33c0a8..c515a0c 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -34,13 +34,13 @@ set >> \$LOG
 
 [ -n "\${new_ip_address}" ]          && ip addr add \${new_ip_address}/\${new_subnet_mask} dev \${interface}
 [ -n "\${new_routers}" ]             && for r in \${new_routers}; do ip route add default via \${r} dev \${interface}; done
+:> /etc/resolv.conf
 [ -n "\${new_domain_name_servers}" ] && for d in \${new_domain_name_servers}; do echo "nameserver \${d}" >> /etc/resolv.conf; done
-[ -n "\${new_domain_name}" ]         && (printf "search"; for d in \${new_domain_name}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
-
+[ -n "\${new_domain_name}" ]         && echo "search \${new_domain_name}" >> /etc/resolf.conf
+[ -n "\${new_domain_search}" ]       && (printf "search"; for d in \${new_domain_search}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
 [ -n "\${new_ip6_address}" ]         && ip addr add \${new_ip6_address}/\${new_ip6_prefixlen} dev \${interface}
 [ -n "\${new_dhcp6_name_servers}" ]  && for d in \${new_dhcp6_name_servers}; do echo "nameserver \${d}%\${interface}" >> /etc/resolv.conf; done
 [ -n "\${new_dhcp6_domain_search}" ] && (printf "search"; for d in \${new_dhcp6_domain_search}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
-
 [ -n "\${new_host_name}" ]           && hostname "\${new_host_name}"
 exit 0
 EOF
-- 
@@ -34,13 +34,13 @@ set >> \$LOG
 
 [ -n "\${new_ip_address}" ]          && ip addr add \${new_ip_address}/\${new_subnet_mask} dev \${interface}
 [ -n "\${new_routers}" ]             && for r in \${new_routers}; do ip route add default via \${r} dev \${interface}; done
+:> /etc/resolv.conf
 [ -n "\${new_domain_name_servers}" ] && for d in \${new_domain_name_servers}; do echo "nameserver \${d}" >> /etc/resolv.conf; done
-[ -n "\${new_domain_name}" ]         && (printf "search"; for d in \${new_domain_name}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
-
+[ -n "\${new_domain_name}" ]         && echo "search \${new_domain_name}" >> /etc/resolf.conf
+[ -n "\${new_domain_search}" ]       && (printf "search"; for d in \${new_domain_search}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
 [ -n "\${new_ip6_address}" ]         && ip addr add \${new_ip6_address}/\${new_ip6_prefixlen} dev \${interface}
 [ -n "\${new_dhcp6_name_servers}" ]  && for d in \${new_dhcp6_name_servers}; do echo "nameserver \${d}%\${interface}" >> /etc/resolv.conf; done
 [ -n "\${new_dhcp6_domain_search}" ] && (printf "search"; for d in \${new_dhcp6_domain_search}; do printf " %s" "\${d}"; done; printf "\n") >> /etc/resolv.conf
-
 [ -n "\${new_host_name}" ]           && hostname "\${new_host_name}"
 exit 0
 EOF
-- 
2.36.1

[PATCH 08/18] tests: Fix detection of empty 'hout' responses in passt{,_in_ns} tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1849 bytes --]

The dhcp/passt and dhcp/passt_in_ns tests at least, and maybe others
use 'hout' commands that need to be able to detect empty output.
However, we don't set PS1, which means the screen-scraping logic which
detects this may not be reliable.  In addition, if the host is using a
recent bash, it will have bracketed paste mode enabled which will also
add escape codes which will mess up the empty output detection.

Set the prompt and disable bracketed paste mode from the passt and
passt_in_ns setups to avoid these problems.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/lib/setup | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/test/lib/setup b/test/lib/setup
index 0d0f3cf..adbfa9e 100755
--- a/test/lib/setup
+++ b/test/lib/setup
@@ -57,6 +57,13 @@ setup_passt() {
 		" -device virtio-net-pci,netdev=hostnet0,x-txburst=16384"  \
 		" -netdev socket,fd=5,id=hostnet0"
 	pane_status GUEST
+
+	# Set things up to reduce problems parsing host command output
+	pane_run HOST "PS1='$ '"
+	pane_wait HOST
+	# Non-bash shells will fail this, but also don't need it, so ignore errors
+	pane_run HOST "bind 'set enable-bracketed-paste off'"
+	pane_wait HOST
 }
 
 # setup_pasta() - Create a network and user namespace, connect pasta to it
@@ -168,6 +175,13 @@ setup_passt_in_ns() {
 		" -device virtio-net-pci,netdev=hostnet0,x-txburst=524288" \
 		" -netdev socket,fd=5,id=hostnet0"
 	pane_status GUEST
+
+	# Set things up to reduce problems parsing host command output
+	pane_run HOST "PS1='$ '"
+	pane_wait HOST
+	# Non-bash shells will fail this, but also don't need it, so ignore errors
+	pane_run HOST "bind 'set enable-bracketed-paste off'"
+	pane_wait HOST
 }
 
 # setup_two_guests() - Set up two namespace, run qemu and passt in both of them
-- 
@@ -57,6 +57,13 @@ setup_passt() {
 		" -device virtio-net-pci,netdev=hostnet0,x-txburst=16384"  \
 		" -netdev socket,fd=5,id=hostnet0"
 	pane_status GUEST
+
+	# Set things up to reduce problems parsing host command output
+	pane_run HOST "PS1='$ '"
+	pane_wait HOST
+	# Non-bash shells will fail this, but also don't need it, so ignore errors
+	pane_run HOST "bind 'set enable-bracketed-paste off'"
+	pane_wait HOST
 }
 
 # setup_pasta() - Create a network and user namespace, connect pasta to it
@@ -168,6 +175,13 @@ setup_passt_in_ns() {
 		" -device virtio-net-pci,netdev=hostnet0,x-txburst=524288" \
 		" -netdev socket,fd=5,id=hostnet0"
 	pane_status GUEST
+
+	# Set things up to reduce problems parsing host command output
+	pane_run HOST "PS1='$ '"
+	pane_wait HOST
+	# Non-bash shells will fail this, but also don't need it, so ignore errors
+	pane_run HOST "bind 'set enable-bracketed-paste off'"
+	pane_wait HOST
 }
 
 # setup_two_guests() - Set up two namespace, run qemu and passt in both of them
-- 
2.36.1

[PATCH 09/18] tests: Fix creation of test file in udp passt tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1113 bytes --]

In what looks like a copy-and-paste error from the TCP script, the
udp/passt test script creates a test file called '__TEMP_BIG__', while
the commands it use the variable __TEMP__.  Correct this so that a) we
actually transfer the data we created for the purpose and b) we don't
leave a stale __TEMP_BIG__ file in the current directory.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/udp/passt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/udp/passt b/test/udp/passt
index 8a848bd..9598d0e 100644
--- a/test/udp/passt
+++ b/test/udp/passt
@@ -18,7 +18,7 @@ test	UDP/IPv4: host to guest
 temp	TEMP
 temp	NC_PID
 guestb	(nc -u -q1 -4 -l 10001 & echo $! > __NC_PID__) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	dd if=/dev/urandom bs=1k count=5 > __TEMP_BIG__ && printf "\nEND_OF_TEST\n" >> __TEMP__
+host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 host	cat __TEMP__ | nc -u -q1 -N 127.0.0.1 10001
 guestw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
-- 
@@ -18,7 +18,7 @@ test	UDP/IPv4: host to guest
 temp	TEMP
 temp	NC_PID
 guestb	(nc -u -q1 -4 -l 10001 & echo $! > __NC_PID__) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	dd if=/dev/urandom bs=1k count=5 > __TEMP_BIG__ && printf "\nEND_OF_TEST\n" >> __TEMP__
+host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 host	cat __TEMP__ | nc -u -q1 -N 127.0.0.1 10001
 guestw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
-- 
2.36.1

[PATCH 10/18] valgrind needs futex

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 817 bytes --]

Some versions of valgrind (such as the version on my Fedora laptop -
valgrind-3.19.0-3.fc36.x86_64) use futexes.  But futex is currently not
allowed in the seccomp filter, even with the extra calls added for
valgrind builds.  Add it, to avoid spurious valgrind failures.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Makefile b/Makefile
index 6f7c971..0de872e 100644
--- a/Makefile
+++ b/Makefile
@@ -115,7 +115,7 @@ qrap: $(QRAP_SRCS) passt.h
 
 valgrind: EXTRA_SYSCALLS="rt_sigprocmask rt_sigtimedwait rt_sigaction \
 			  getpid gettid kill clock_gettime mmap munmap open \
-			  unlink gettimeofday"
+			  unlink gettimeofday futex"
 valgrind: CFLAGS:=-g -O0 $(filter-out -O%,$(CFLAGS))
 valgrind: all
 
-- 
@@ -115,7 +115,7 @@ qrap: $(QRAP_SRCS) passt.h
 
 valgrind: EXTRA_SYSCALLS="rt_sigprocmask rt_sigtimedwait rt_sigaction \
 			  getpid gettid kill clock_gettime mmap munmap open \
-			  unlink gettimeofday"
+			  unlink gettimeofday futex"
 valgrind: CFLAGS:=-g -O0 $(filter-out -O%,$(CFLAGS))
 valgrind: all
 
-- 
2.36.1

[PATCH 11/18] tests: Use socat instead of netcat

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 47597 bytes --]

Commit 41c02e10 ("tests: Use nmap-ncat instead of openbsd netcat for pasta
tests") updated the pasta tests to use the nmap version of ncat instead of
the openbsd version, for greater portability.

For some upcoming changes, however, we'll be wanting to use socat.
"socat" can do everything "ncat" can and more, so let's move all the
tests using host tools (either directly on the host or via mbuto
generated images) to using socat instead.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/README.md        |   2 +-
 test/demo/passt       |   8 +--
 test/demo/pasta       |  12 ++--
 test/passt.mbuto      |   2 +-
 test/tcp/passt        |  36 +++++------
 test/tcp/passt_in_ns  | 138 +++++++++++++++++++++---------------------
 test/tcp/pasta        |  58 +++++++++---------
 test/two_guests/basic |  22 +++----
 test/udp/passt        |  22 +++----
 test/udp/passt_in_ns  |  72 +++++++++++-----------
 test/udp/pasta        |  28 ++++-----
 11 files changed, 200 insertions(+), 200 deletions(-)

diff --git a/test/README.md b/test/README.md
index 201332b..581949a 100644
--- a/test/README.md
+++ b/test/README.md
@@ -29,7 +29,7 @@ on a system, i.e. common utilities such as a shell are not included here.
 Example for Debian, and possibly most Debian-based distributions:
 
     build-essential git jq strace iperf3 qemu-system-x86 tmux sipcalc bc
-    clang-tidy cppcheck isc-dhcp-common psmisc linux-cpupower ncat
+    clang-tidy cppcheck isc-dhcp-common psmisc linux-cpupower socat
     netcat-openbsd fakeroot lz4 lm-sensors qemu-system-arm qemu-system-ppc
     qemu-system-misc qemu-system-x86 valgrind
 
diff --git a/test/demo/passt b/test/demo/passt
index bcad718..dac8ac7 100644
--- a/test/demo/passt
+++ b/test/demo/passt
@@ -130,9 +130,9 @@ nl
 nl
 say	Let's try to communicate between host and guest.
 sleep	2
-guestb	nc -6 -l -p 5201
+guestb	socat TCP6-LISTEN:5201 STDIO
 sleep	2
-host	echo "Hello from the host" | nc -N ::1 5201
+host	echo "Hello from the host" | socat -u STDIN TCP6:[::1]:5201
 sleep	5
 
 nl
@@ -143,9 +143,9 @@ say	   the address of the default gateway.
 sleep	2
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 sleep	5
-hostb	nc -l -p 31337
+hostb	socat TCP4-LISTEN:31337 STDIO
 sleep	2
-guest	echo "Hello from the guest" | nc -N __GW__ 31337
+guest	echo "Hello from the guest" | socat -u STDIN TCP4:__GW__:x31337
 sleep	3
 
 nl
diff --git a/test/demo/pasta b/test/demo/pasta
index f4b7da2..d84d856 100644
--- a/test/demo/pasta
+++ b/test/demo/pasta
@@ -99,9 +99,9 @@ nl
 say	  pasta detects bound ports and forwards them.
 sleep	3
 
-nsb	ncat -6 -l ::1 31337
+nsb	socat TCP6-LISTEN:31337,bind=[::1] STDOUT
 sleep	2
-host	echo "Hello from the host" | ncat ::1 31337
+host	echo "Hello from the host" | socat -u STDIN TCP6:[::1]:31337
 sleep	5
 
 nl
@@ -110,9 +110,9 @@ say	Now the other way around...
 nl
 say	  we can use a loopback address
 sleep	2
-hostb	ncat -l ::1 31337
+hostb	socat TCP6-LISTEN:31337,bind=[::1] STDIO
 sleep	2
-ns	echo "Hello from the namespace" | ncat ::1 31337
+ns	echo "Hello from the namespace" | socat -u STDIN TCP6:[::1]:31337
 sleep	5
 
 nl
@@ -120,9 +120,9 @@ say	  or the address of the default gateway.
 sleep	2
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 sleep	5
-hostb	ncat -l 0.0.0.0 31337
+hostb	socat TCP4-LISTEN:31337 STDIO
 sleep	2
-ns	echo "Hello from the namespace" | ncat __GW__ 31337
+ns	echo "Hello from the namespace" | socat -u STDIN TCP4:__GW__:31337
 sleep	3
 
 nl
diff --git a/test/passt.mbuto b/test/passt.mbuto
index c515a0c..8b41674 100755
--- a/test/passt.mbuto
+++ b/test/passt.mbuto
@@ -12,7 +12,7 @@
 
 PROGS="${PROGS:-ash,dash,bash ip mount ls insmod mkdir ln cat chmod lsmod
        modprobe find grep mknod mv rm umount jq iperf3 dhclient hostname
-       sed tr chown sipcalc cut md5sum nc dd strace ping tail killall sleep
+       sed tr chown sipcalc cut md5sum socat dd strace ping tail killall sleep
        sysctl nproc tcp_rr tcp_crr udp_rr which tee seq bc}"
 
 KMODS="${KMODS:- virtio_net virtio_pci}"
diff --git a/test/tcp/passt b/test/tcp/passt
index 4baa610..6e65137 100644
--- a/test/tcp/passt
+++ b/test/tcp/passt
@@ -11,75 +11,75 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-gtools	nc ip jq cat md5sum cut
-htools	dd nc ip jq cat md5sum cut
+gtools	socat ip jq md5sum cut
+htools	dd socat ip jq md5sum cut
 
 test	TCP/IPv4: host to guest: big transfer
 temp	TEMP_BIG
-guestb	nc -4 -l 10001 > test_big.bin
+guestb	socat -u TCP4-LISTEN:10001,reuseaddr OPEN:test_big.bin,create,trunc
 host	dd if=/dev/urandom bs=1M count=10 > __TEMP_BIG__
-host	cat __TEMP_BIG__ | nc -N 127.0.0.1 10001
+host	socat -u OPEN:__TEMP_BIG__ TCP4:127.0.0.1:10001
 guestw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__HOST_MD5_BIG__" ]
 
 test	TCP/IPv4: guest to host: big transfer
-hostb	nc -4 -l 10003 > __TEMP_BIG__
+hostb	socat -u TCP4-LISTEN:10003,bind=127.0.0.1,reuseaddr OPEN:__TEMP_BIG__,create,trunc
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest	cat test_big.bin | nc -N __GW__ 10003
+guest	socat -u OPEN:test_big.bin TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__HOST_MD5_BIG__" ]
 
 test	TCP/IPv4: host to guest: small transfer
 temp	TEMP_SMALL
-guestb	nc -4 -l 10001 > test_small.bin
+guestb	socat -u TCP4-LISTEN:10001,reuseaddr OPEN:test_small.bin,create,trunc
 host	dd if=/dev/urandom bs=2k count=1 > __TEMP_SMALL__
-host	cat __TEMP_SMALL__ | nc -N 127.0.0.1 10001
+host	socat -u OPEN:__TEMP_SMALL__ TCP4:127.0.0.1:10001
 guestw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__HOST_MD5_SMALL__" ]
 
 test	TCP/IPv4: guest to host: small transfer
-hostb	nc -4 -l 10003 > __TEMP_SMALL__
+hostb	socat -u TCP4-LISTEN:10003,bind=127.0.0.1,reuseaddr OPEN:__TEMP_SMALL__,create,trunc
 sleep	1
-guest	cat test_small.bin | nc -N __GW__ 10003
+guest	socat -u OPEN:test_small.bin TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__HOST_MD5_SMALL__" ]
 
 
 test	TCP/IPv6: host to guest: big transfer
-guestb	nc -6 -l 10001 > test_big.bin
+guestb	socat -u TCP6-LISTEN:10001,reuseaddr OPEN:test_big.bin,create,trunc
 sleep	1
-host	cat __TEMP_BIG__ | nc -N ::1 10001
+host	socat -u OPEN:__TEMP_BIG__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__HOST_MD5_BIG__" ]
 
 test	TCP/IPv6: guest to host: big transfer
-hostb	nc -6 -l 10003 > __TEMP_BIG__
+hostb	socat -u TCP6-LISTEN:10003,bind=[::1],reuseaddr OPEN:__TEMP_BIG__,create,trunc
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-guest	cat test_big.bin | nc -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test_big.bin TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__HOST_MD5_BIG__" ]
 
 test	TCP/IPv6: host to guest: small transfer
-guestb	nc -6 -l 10001 > test_small.bin
+guestb	socat -u TCP6-LISTEN:10001,reuseaddr OPEN:test_small.bin,create,trunc
 sleep	1
-host	cat __TEMP_SMALL__ | nc -N ::1 10001
+host	socat -u OPEN:__TEMP_SMALL__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__HOST_MD5_SMALL__" ]
 
 test	TCP/IPv6: guest to host: small transfer
-hostb	nc -6 -l 10003 > __TEMP_SMALL__
+hostb	socat -u TCP6-LISTEN:10003,bind=[::1],reuseaddr OPEN:__TEMP_SMALL__,create,trunc
 sleep	1
-guest	cat test_small.bin | nc -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test_small.bin TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__HOST_MD5_SMALL__" ]
diff --git a/test/tcp/passt_in_ns b/test/tcp/passt_in_ns
index 2a11e0b..ab22485 100644
--- a/test/tcp/passt_in_ns
+++ b/test/tcp/passt_in_ns
@@ -11,15 +11,15 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-gtools	nc ip jq cat md5sum cut
-htools	dd nc ip jq cat md5sum cut
-nstools	nc ip jq cat md5sum cut
+gtools	socat ip jq md5sum cut
+htools	dd socat ip jq md5sum cut
+nstools	socat ip jq md5sum cut
 
 test	TCP/IPv4: host to guest: big transfer
 temp	TEMP_BIG
-guestb	nc -4 -l 10001 > test_big.bin
-host	dd if=/dev/urandom bs=1M count=10 > __TEMP_BIG__
-host	cat __TEMP_BIG__ | nc -N 127.0.0.1 10001
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_big.bin,create,trunc
+host	dd if=/dev/urandom bs=1M count=10 of=__TEMP_BIG__
+host	socat -u OPEN:__TEMP_BIG__ TCP4:127.0.0.1:10001
 guestw
 hout	MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
@@ -27,63 +27,63 @@ check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: host to ns: big transfer
 temp	TEMP_NS_BIG
-nsb	nc -4 -l 10002 > __TEMP_NS_BIG__
-host	cat __TEMP_BIG__ | nc -N 127.0.0.1 10002
+nsb	socat -u TCP4-LISTEN:10002 OPEN:__TEMP_NS_BIG__,create,trunc
+host	socat -u OPEN:__TEMP_BIG__ TCP4:127.0.0.1:10002
 nsw
 nsout	NS_MD5_BIG md5sum __TEMP_NS_BIG__ | cut -d' ' -f1
 check	[ "__NS_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: guest to host: big transfer
-hostb	nc -4 -l 10003 > __TEMP_BIG__
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest	cat test_big.bin | nc -N __GW__ 10003
+guest	socat -u OPEN:test_big.bin TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: guest to ns: big transfer
-nsb	nc -4 -l 10002 > __TEMP_BIG__
-guest	cat test_big.bin | nc -N __GW__ 10002
+nsb	socat -u TCP4-LISTEN:10002 OPEN:__TEMP_BIG__,create,trunc
+guest	socat -u OPEN:test_big.bin TCP4:__GW__:10002
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to host (spliced): big transfer
 sleep	1
-hostb	nc -4 -l 10003 > __TEMP_BIG__
-ns	cat __TEMP_NS_BIG__ | nc -N 127.0.0.1 10003
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to host (via tap): big transfer
-hostb	nc -4 -l 10003 > __TEMP_BIG__
-ns	cat __TEMP_NS_BIG__ | nc -N __GW__ 10003
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to guest (using loopback address): big transfer
-guestb	nc -4 -l 10001 > test_big.bin
-ns	cat __TEMP_NS_BIG__ | nc -N 127.0.0.1 10001
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_big.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:127.0.0.1:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to guest (using namespace address): big transfer
-guestb	nc -4 -l 10001 > test_big.bin
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_big.bin,create,trunc
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
 nsout	ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[0].local'
-ns	cat __TEMP_NS_BIG__ | nc -N __ADDR__ 10001
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:__ADDR__:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: host to guest: small transfer
 temp	TEMP_SMALL
-guestb	nc -4 -l 10001 > test_small.bin
-host	dd if=/dev/urandom bs=2k count=100 > __TEMP_SMALL__
-host	cat __TEMP_SMALL__ | nc -N 127.0.0.1 10001
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_small.bin,create,trunc
+host	dd if=/dev/urandom bs=2k count=100 of=__TEMP_SMALL__
+host	socat -u OPEN:__TEMP_SMALL__ TCP4:127.0.0.1:10001
 guestw
 hout	MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
@@ -91,171 +91,171 @@ check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: host to ns: small transfer
 temp	TEMP_NS_SMALL
-nsb	nc -4 -l 10002 > __TEMP_NS_SMALL__
-host	cat __TEMP_SMALL__ | nc -N 127.0.0.1 10002
+nsb	socat -u TCP4-LISTEN:10002 OPEN:__TEMP_NS_SMALL__,create,trunc
+host	socat -u OPEN:__TEMP_SMALL__ TCP4:127.0.0.1:10002
 nsw
 nsout	NS_MD5_SMALL md5sum __TEMP_NS_SMALL__ | cut -d' ' -f1
 check	[ "__NS_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: guest to host: small transfer
-hostb	nc -4 -l 10003 > __TEMP_SMALL__
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest	cat test_small.bin | nc -N __GW__ 10003
+guest	socat -u OPEN:test_small.bin TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: guest to ns: small transfer
-nsb	nc -4 -l 10002 > __TEMP_SMALL__
-guest	cat test_small.bin | nc -N __GW__ 10002
+nsb	socat -u TCP4-LISTEN:10002 OPEN:__TEMP_SMALL__,create,trunc
+guest	socat -u OPEN:test_small.bin TCP4:__GW__:10002
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to host (spliced): small transfer
 sleep	1
-hostb	nc -4 -l 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | nc -N 127.0.0.1 10003
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to host (via tap): small transfer
-hostb	nc -4 -l 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | nc -N __GW__ 10003
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to guest (using loopback address): small transfer
-guestb	nc -4 -l 10001 > test_small.bin
-ns	cat __TEMP_NS_SMALL__ | nc -N 127.0.0.1 10001
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_small.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP4:127.0.0.1:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to guest (using namespace address): small transfer
-guestb	nc -4 -l 10001 > test_small.bin
-ns	cat __TEMP_NS_SMALL__ | nc -N __ADDR__ 10001
+guestb	socat -u TCP4-LISTEN:10001 OPEN:test_small.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP4:__ADDR__:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: host to guest: big transfer
-guestb	nc -6 -l 10001 > test_big.bin
-host	cat __TEMP_BIG__ | nc -N ::1 10001
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_big.bin,create,trunc
+host	socat -u OPEN:__TEMP_BIG__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: host to ns: big transfer
-nsb	nc -6 -l 10002 > __TEMP_NS_BIG__
-host	cat __TEMP_BIG__ | nc -N ::1 10002
+nsb	socat -u TCP6-LISTEN:10002 OPEN:__TEMP_NS_BIG__,create,trunc
+host	socat -u OPEN:__TEMP_BIG__ TCP6:[::1]:10002
 nsw
 nsout	NS_MD5_BIG md5sum __TEMP_NS_BIG__ | cut -d' ' -f1
 check	[ "__NS_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: guest to host: big transfer
-hostb	nc -6 -l 10003 > __TEMP_BIG__
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-guest	cat test_big.bin | nc -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test_big.bin TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: guest to ns: big transfer
-nsb	nc -6 -l 10002 > __TEMP_BIG__
-guest	cat test_big.bin | nc -N __GW6__%__IFNAME__ 10002
+nsb	socat -u TCP6-LISTEN:10002 OPEN:__TEMP_BIG__,create,trunc
+guest	socat -u OPEN:test_big.bin TCP6:[__GW6__%__IFNAME__]:10002
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: ns to host (spliced): big transfer
-hostb	nc -6 -l 10003 > __TEMP_BIG__
-ns	cat __TEMP_NS_BIG__ | nc -N ::1 10003
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[::1]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: ns to host (via tap): big transfer
-hostb	nc -6 -l 10003 > __TEMP_BIG__
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS_BIG__ | nc -N __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: ns to guest (using loopback address): big transfer
-guestb	nc -6 -l 10001 > test_big.bin
-ns	cat __TEMP_NS_BIG__ | nc -N ::1 10001
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_big.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: ns to guest (using namespace address): big transfer
-guestb	nc -6 -l 10001 > test_big.bin
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_big.bin,create,trunc
 nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[0].local'
-ns	cat __TEMP_NS_BIG__ | nc -N __ADDR6__ 10001
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[__ADDR6__]:10001
 guestw
 gout	GUEST_MD5_BIG md5sum test_big.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: host to guest: small transfer
-guestb	nc -6 -l 10001 > test_small.bin
-host	cat __TEMP_SMALL__ | nc -N ::1 10001
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_small.bin,create,trunc
+host	socat -u OPEN:__TEMP_SMALL__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: host to ns: small transfer
-nsb	nc -6 -l 10002 > __TEMP_NS_SMALL__
-host	cat __TEMP_SMALL__ | nc -N ::1 10002
+nsb	socat -u TCP6-LISTEN:10002 OPEN:__TEMP_NS_SMALL__,create,trunc
+host	socat -u OPEN:__TEMP_SMALL__ TCP6:[::1]:10002
 nsw
 nsout	NS_MD5_SMALL md5sum __TEMP_NS_SMALL__ | cut -d' ' -f1
 check	[ "__NS_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: guest to host: small transfer
-hostb	nc -6 -l 10003 > __TEMP_SMALL__
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-guest	cat test_small.bin | nc -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test_small.bin TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: guest to ns: small transfer
-nsb	nc -6 -l 10002 > __TEMP_SMALL__
-guest	cat test_small.bin | nc -N __GW6__%__IFNAME__ 10002
+nsb	socat -u TCP6-LISTEN:10002 OPEN:__TEMP_SMALL__
+guest	socat -u OPEN:test_small.bin TCP6:[__GW6__%__IFNAME__]:10002
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to host (spliced): small transfer
-hostb	nc -6 -l 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | nc -N ::1 10003
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[::1]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to host (via tap): small transfer
-hostb	nc -6 -l 10003 > __TEMP_SMALL__
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS_SMALL__ | nc -N __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to guest (using loopback address): small transfer
-guestb	nc -6 -l 10001 > test_small.bin
-ns	cat __TEMP_NS_SMALL__ | nc -N ::1 10001
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_small.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[::1]:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to guest (using namespace address): small transfer
-guestb	nc -6 -l 10001 > test_small.bin
-ns	cat __TEMP_NS_SMALL__ | nc -N __ADDR6__ 10001
+guestb	socat -u TCP6-LISTEN:10001 OPEN:test_small.bin,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[__ADDR6__]:10001
 guestw
 gout	GUEST_MD5_SMALL md5sum test_small.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5_SMALL__" = "__MD5_SMALL__" ]
diff --git a/test/tcp/pasta b/test/tcp/pasta
index 068393a..0d6c729 100644
--- a/test/tcp/pasta
+++ b/test/tcp/pasta
@@ -11,31 +11,31 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-htools	dd ncat ip jq cat md5sum cut
-nstools	ncat ip jq cat md5sum cut
+htools	dd socat ip jq md5sum cut
+nstools	socat ip jq md5sum cut
 
 test	TCP/IPv4: host to ns: big transfer
 temp	TEMP_BIG
 temp	TEMP_NS_BIG
-nsb	ncat -4 -l 127.0.0.1 10002 > __TEMP_NS_BIG__
-host	dd if=/dev/urandom bs=1M count=10 > __TEMP_BIG__
-host	cat __TEMP_BIG__ | ncat 127.0.0.1 10002
+nsb	socat -u TCP4-LISTEN:10002,bind=127.0.0.1 OPEN:__TEMP_NS_BIG__,create,trunc
+host	dd if=/dev/urandom bs=1M count=10 of=__TEMP_BIG__
+host	socat -u OPEN:__TEMP_BIG__ TCP4:127.0.0.1:10002
 nsw
 hout	MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 nsout	NS_MD5_BIG md5sum __TEMP_NS_BIG__ | cut -d' ' -f1
 check	[ "__NS_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to host (spliced): big transfer
-hostb	ncat -4 -l 127.0.0.1 10003 > __TEMP_BIG__
-ns	cat __TEMP_NS_BIG__ | ncat 127.0.0.1 10003
+hostb	socat -u TCP4-LISTEN:10003,bind=127.0.0.1 OPEN:__TEMP_BIG__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv4: ns to host (via tap): big transfer
-hostb	ncat -4 -l 0.0.0.0 10003 > __TEMP_BIG__
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-ns	cat __TEMP_NS_BIG__ | ncat __GW__ 10003
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
@@ -43,50 +43,50 @@ check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 test	TCP/IPv4: host to ns: small transfer
 temp	TEMP_SMALL
 temp	TEMP_NS_SMALL
-nsb	ncat -4 -l 127.0.0.1 10002 > __TEMP_NS_SMALL__
-host	dd if=/dev/urandom bs=2k count=1 > __TEMP_SMALL__
-host	cat __TEMP_SMALL__ | ncat 127.0.0.1 10002
+nsb	socat -u TCP4-LISTEN:10002,bind=127.0.0.1 OPEN:__TEMP_NS_SMALL__,create,trunc
+host	dd if=/dev/urandom bs=2k count=1 of=__TEMP_SMALL__
+host	socat OPEN:__TEMP_SMALL__ TCP4:127.0.0.1:10002
 nsw
 hout	MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 nsout	NS_MD5_SMALL md5sum __TEMP_NS_SMALL__ | cut -d' ' -f1
 check	[ "__NS_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to host (spliced): small transfer
-hostb	ncat -4 -l 127.0.0.1 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | ncat 127.0.0.1 10003
+hostb	socat -u TCP4-LISTEN:10003,bind=127.0.0.1 OPEN:__TEMP_SMALL__,create,trunc
+ns	socat OPEN:__TEMP_NS_SMALL__ TCP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv4: ns to host (via tap): small transfer
-hostb	ncat -4 -l 0.0.0.0 10003 > __TEMP_SMALL__
+hostb	socat -u TCP4-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-ns	cat __TEMP_NS_SMALL__ | ncat __GW__ 10003
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP4:__GW__:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: host to ns: big transfer
-nsb	ncat -6 -l ::1 10002 > __TEMP_NS_BIG__
-host	cat __TEMP_BIG__ | ncat ::1 10002
+nsb	socat -u TCP6-LISTEN:10002,bind=[::1] OPEN:__TEMP_NS_BIG__,create,trunc
+host	socat -u OPEN:__TEMP_BIG__ TCP6:[::1]:10002
 nsw
 hout	MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 nsout	NS_MD5_BIG md5sum __TEMP_NS_BIG__ | cut -d' ' -f1
 check	[ "__NS_MD5_BIG__" = "__MD5_BIG__" ]
 
 test	TCP/IPv6: ns to host (spliced): big transfer
-hostb	ncat -6 -l ::1 10003 > __TEMP_BIG__
-ns	cat __TEMP_NS_BIG__ | ncat ::1 10003
+hostb	socat -u TCP6-LISTEN:10003,bind=[::1] OPEN:__TEMP_BIG__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[::1]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 nsw
 
 test	TCP/IPv6: ns to host (via tap): big transfer
-hostb	ncat -6 -l :: 10003 > __TEMP_BIG__
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_BIG__,create,trunc
 nsout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS_BIG__ | ncat __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS_BIG__ TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_BIG md5sum __TEMP_BIG__ | cut -d' ' -f1
 check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
@@ -94,24 +94,24 @@ check	[ "__HOST_MD5_BIG__" = "__MD5_BIG__" ]
 test	TCP/IPv6: host to ns: small transfer
 temp	TEMP_SMALL
 temp	TEMP_NS_SMALL
-nsb	ncat -6 -l ::1 10002 > __TEMP_NS_SMALL__
-host	dd if=/dev/urandom bs=2k count=1 > __TEMP_SMALL__
-host	cat __TEMP_SMALL__ | ncat ::1 10002
+nsb	socat -u TCP6-LISTEN:10002,bind=[::1] OPEN:__TEMP_NS_SMALL__,create,trunc
+host	dd if=/dev/urandom bs=2k count=1 of=__TEMP_SMALL__
+host	socat -u OPEN:__TEMP_SMALL__ TCP6:[::1]:10002
 nsw
 hout	MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 nsout	NS_MD5_SMALL md5sum __TEMP_NS_SMALL__ | cut -d' ' -f1
 check	[ "__NS_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to host (spliced): small transfer
-hostb	ncat -6 -l ::1 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | ncat ::1 10003
+hostb	socat -u TCP6-LISTEN:10003,bind=[::1] OPEN:__TEMP_SMALL__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[::1]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
 
 test	TCP/IPv6: ns to host (via tap): small transfer
-hostb	ncat -6 -l :: 10003 > __TEMP_SMALL__
-ns	cat __TEMP_NS_SMALL__ | ncat __GW6__%__IFNAME__ 10003
+hostb	socat -u TCP6-LISTEN:10003 OPEN:__TEMP_SMALL__,create,trunc
+ns	socat -u OPEN:__TEMP_NS_SMALL__ TCP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5_SMALL md5sum __TEMP_SMALL__ | cut -d' ' -f1
 check	[ "__HOST_MD5_SMALL__" = "__MD5_SMALL__" ]
diff --git a/test/two_guests/basic b/test/two_guests/basic
index 850a27b..cf0b975 100644
--- a/test/two_guests/basic
+++ b/test/two_guests/basic
@@ -11,9 +11,9 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-g1tools	ip jq dhclient nc cat
-g2tools	ip jq dhclient nc cat
-htools	ip jq cat md5sum cut
+g1tools	ip jq dhclient socat cat
+g2tools	ip jq dhclient socat cat
+htools	ip jq md5sum cut
 
 test	Interface names
 g1out	IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
@@ -46,8 +46,8 @@ check	[ "__ADDR2_6__" = "__HOST_ADDR6__" ]
 
 test	TCP/IPv4: guest 1 > guest 2
 g1out	GW1 ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest2b	nc -4 -l 10004 > msg
-guest1	echo "Hello_from_guest_1" | nc -N __GW1__ 10004
+guest2b	socat -u TCP4-LISTEN:10004 OPEN:msg,create,trunc
+guest1	echo "Hello_from_guest_1" | socat -u STDIN TCP4:__GW1__:10004
 guest2w
 sleep	1
 g2out	MSG2 cat msg
@@ -55,24 +55,24 @@ check	[ "__MSG2__" = "Hello_from_guest_1" ]
 
 test	TCP/IPv6: guest 2 > guest 1
 g2out	GW2_6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest1b	nc -6 -l 10001 > msg
-guest2	echo "Hello_from_guest_2" | nc -N __GW2_6__%__IFNAME2__ 10001
+guest1b	socat -u TCP6-LISTEN:10001 OPEN:msg,create,trunc
+guest2	echo "Hello_from_guest_2" | socat -u STDIN TCP6:[__GW2_6__%__IFNAME2__]:10001
 guest1w
 sleep	1
 g1out	MSG1 cat msg
 check	[ "__MSG1__" = "Hello_from_guest_2" ]
 
 test	UDP/IPv4: guest 1 > guest 2
-guest2b	nc -u -W1 -4 -l 10004 > msg
-guest1	echo "Hello_from_guest_1" | nc -u -q1 __GW1__ 10004
+guest2b	socat -u TCP4-LISTEN:10004 OPEN:msg,create,trunc
+guest1	echo "Hello_from_guest_1" | socat -u STDIN TCP4:__GW1__:10004
 guest2w
 sleep	1
 g2out	MSG2 cat msg
 check	[ "__MSG2__" = "Hello_from_guest_1" ]
 
 test	UDP/IPv6: guest 2 > guest 1
-guest1b	nc -u -W1 -6 -l 10001 > msg
-guest2	echo "Hello_from_guest_2" | nc -u -q1 -N __GW2_6__%__IFNAME2__ 10001
+guest1b	socat -u TCP6-LISTEN:10001 OPEN:msg,create,trunc
+guest2	echo "Hello_from_guest_2" | socat -u STDIN TCP6:[__GW2_6__%__IFNAME2__]:10001
 guest1w
 sleep	1
 g1out	MSG1 cat msg
diff --git a/test/udp/passt b/test/udp/passt
index 9598d0e..ee575e0 100644
--- a/test/udp/passt
+++ b/test/udp/passt
@@ -11,15 +11,15 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-gtools	nc tee grep cat ip jq md5sum cut
-htools	printf dd nc tee grep cat ip jq md5sum cut
+gtools	socat tee grep cat ip jq md5sum cut
+htools	printf dd socat tee grep cat ip jq md5sum cut
 
 test	UDP/IPv4: host to guest
 temp	TEMP
-temp	NC_PID
-guestb	(nc -u -q1 -4 -l 10001 & echo $! > __NC_PID__) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+temp	SC_PID
+guestb	(socat -u UDP4-LISTEN:10001 STDOUT & echo $! > __SC_PID__) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
-host	cat __TEMP__ | nc -u -q1 -N 127.0.0.1 10001
+host	socat -u OPEN:__TEMP__ UDP4:127.0.0.1:10001
 guestw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
@@ -27,27 +27,27 @@ check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
 
 test	UDP/IPv4: guest to host
 host	:> __TEMP__
-hostb	(nc -u -q1 -4 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest	cat test.bin | nc -u -q1 -N __GW__ 10003
+guest	socat -u OPEN:test.bin UDP4:__GW__:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
 
 test	UDP/IPv6: host to guest
-guestb	(nc -u -q1 -6 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
+guestb	(socat -u UDP6-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
 sleep	1
-host	cat __TEMP__ | nc -u -q1 -N ::1 10001
+host	socat -u OPEN:__TEMP__ UDP6:[::1]:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
 
 test	UDP/IPv6: guest to host
 host	:> __TEMP__
-hostb	(nc -u -q1 -6 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-guest	cat test.bin | nc -u -q1 -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test.bin UDP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
diff --git a/test/udp/passt_in_ns b/test/udp/passt_in_ns
index 3607251..3536298 100644
--- a/test/udp/passt_in_ns
+++ b/test/udp/passt_in_ns
@@ -11,16 +11,16 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-gtools	nc tee grep cat ip jq md5sum cut
-nstools	nc tee grep cat ip jq md5sum cut
-htools	printf dd nc tee grep cat ip jq md5sum cut
+gtools	socat tee grep cat ip jq md5sum cut
+nstools	socat tee grep cat ip jq md5sum cut
+htools	printf dd socat tee grep cat ip jq md5sum cut
 
 test	UDP/IPv4: host to guest
 temp	TEMP
-temp	NC_PID
-guestb	(nc -u -q1 -4 -l 10001 & echo $! > nc.pic) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pic))
+temp	SC_PID
+guestb	(socat -u UDP4-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
-host	cat __TEMP__ | nc -u -q1 -N 127.0.0.1 10001
+host	socat -u OPEN:__TEMP__ UDP4:127.0.0.1:10001
 guestw
 hout	MD5 md5sum __TEMP__ | cut -d' ' -f1
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
@@ -29,25 +29,25 @@ check	[ "__GUEST_MD5__" = "__MD5__" ]
 test	UDP/IPv4: host to ns
 temp	TEMP_NS
 ns	:> __TEMP_NS__
-nsb	(nc -u -q1 -4 -l 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	cat __TEMP__ | nc -u -q1 -N 127.0.0.1 10002
+nsb	(socat -u UDP4-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+host	socat -u OPEN:__TEMP__ UDP4:127.0.0.1:10002
 nsw
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: guest to host
 host	:> __TEMP__
-hostb	(nc -u -q1 -4 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-guest	cat test.bin | nc -u -q1 -N __GW__ 10003
+guest	socat -u OPEN:test.bin UDP4:__GW__:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: guest to ns
 ns	:> __TEMP_NS__
-nsb	(nc -u -q1 -4 -l 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-guest	cat test.bin | nc -u -q1 -N __GW__ 10002
+nsb	(socat -u UDP4-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+guest	socat -u OPEN:test.bin UDP4:__GW__:10002
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
@@ -55,97 +55,97 @@ check	[ "__HOST_MD5__" = "__MD5__" ]
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
 sleep	1
 host	:> __TEMP__
-hostb	(nc -u -q1 -4 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-ns	cat __TEMP_NS__ | nc -u -q1 -N 127.0.0.1 10003
+hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+ns	socat -u OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
 host	:> __TEMP__
-hostb	(nc -u -q1 -4 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-ns	cat __TEMP_NS__ | nc -u -q1 -N __GW__ 10003
+hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to guest (using loopback address)
-guestb	(nc -u -q1 -4 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
-ns	cat __TEMP_NS__ | nc -u -q1 -N 127.0.0.1 10001
+guestb	(socat -u UDP4-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
+ns	socat -u OPEN:__TEMP_NS__ UDP4:127.0.0.1:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to guest (using namespace address)
-guestb	(nc -u -q1 -4 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
+guestb	(socat -u UDP4-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
 nsout	ADDR ip -j -4 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[0].local'
-ns	cat __TEMP_NS__ | nc -u -q1 -N __ADDR__ 10001
+ns	socat -u OPEN:__TEMP_NS__ UDP4:__ADDR__:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to guest
-guestb	(nc -u -q1 -6 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
-host	cat __TEMP__ | nc -u -q1 -N ::1 10001
+guestb	(socat -u UDP6-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
+host	socat -u OPEN:__TEMP__ UDP6:[::1]:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
 ns	:> __TEMP_NS__
-nsb	(nc -u -q1 -6 -l 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	cat __TEMP__ | nc -u -q1 -N ::1 10002
+nsb	(socat -u UDP6-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: guest to host
 host	:> __TEMP__
-hostb	(nc -u -q1 -6 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-guest	cat test.bin | nc -u -q1 -N __GW6__%__IFNAME__ 10003
+guest	socat -u OPEN:test.bin UDP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: guest to ns
 ns	:> __TEMP_NS__
-nsb	(nc -u -q1 -6 -l 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-guest	cat test.bin | nc -u -q1 -N __GW6__%__IFNAME__ 10002
+nsb	(socat -u UDP6-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+guest	socat -u OPEN:test.bin UDP6:[__GW6__%__IFNAME__]:10002
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
 host	:> __TEMP__
-hostb	(nc -u -q1 -6 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-ns	cat __TEMP_NS__ | nc -u -q1 -N ::1 10003
+hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (via tap)
 host	:> __TEMP__
-hostb	(nc -u -q1 -6 -l 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS__ | nc -u -q1 -N __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to guest (using loopback address)
-guestb	(nc -u -q1 -6 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
-ns	cat __TEMP_NS__ | nc -u -q1 -N ::1 10001
+guestb	(socat -u UDP6-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to guest (using namespace address)
-guestb	(nc -u -q1 -6 -l 10001 & echo $! > nc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat nc.pid))
+guestb	(socat -u UDP6-LISTEN:10001 STDOUT & echo $! > sc.pid) | tee test.bin | (grep -qm1 "END_OF_TEST" && kill $(cat sc.pid))
 nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[0].local'
-ns	cat __TEMP_NS__ | nc -u -q1 -N __ADDR6__ 10001
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[__ADDR6__]:10001
 guestw
 gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
diff --git a/test/udp/pasta b/test/udp/pasta
index a7e5015..43e0513 100644
--- a/test/udp/pasta
+++ b/test/udp/pasta
@@ -11,18 +11,18 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-nstools	ncat tee grep cat ip jq md5sum cut
-htools	printf dd ncat tee grep cat ip jq md5sum cut
+nstools	socat tee grep cat ip jq md5sum cut
+htools	printf dd socat tee grep cat ip jq md5sum cut
 
 test	UDP/IPv4: host to ns
 temp	TEMP
 temp	TEMP_NS
 temp	NC_PID
 ns	:> __TEMP_NS__
-nsb	(ncat -u -4 -l 127.0.0.1 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+nsb	(socat -u UDP4-LISTEN:10002,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 
-host	cat __TEMP__ | ncat -u 127.0.0.1 10002
+host	socat OPEN:__TEMP__ UDP4:127.0.0.1:10002
 nsw
 hout	MD5 md5sum __TEMP__ | cut -d' ' -f1
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
@@ -30,26 +30,26 @@ check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
 host	:> __TEMP__
-hostb	(ncat -u -4 -l 127.0.0.1 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
-ns	cat __TEMP_NS__ | ncat -u 127.0.0.1 10003
+ns	socat OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
 host	:> __TEMP__
-hostb	(ncat -u -4 -l 0.0.0.0 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-ns	cat __TEMP_NS__ | ncat -u __GW__ 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
 ns	:> __TEMP_NS__
-nsb	(ncat -u -6 -l ::1 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	cat __TEMP__ | ncat -u ::1 10002
+nsb	(socat -u UDP6-LISTEN:10002,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
 hout	MD5 md5sum __TEMP__ | cut -d' ' -f1
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
@@ -57,9 +57,9 @@ check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
 host	:> __TEMP__
-hostb	(ncat -u -6 -l ::1 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
-ns	cat __TEMP_NS__ | ncat -u ::1 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
@@ -67,10 +67,10 @@ nsw
 
 test	UDP/IPv6: ns to host (via tap)
 host	:> __TEMP__
-hostb	(ncat -u -6 -l :: 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS__ | ncat -u __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
-- 
@@ -11,18 +11,18 @@
 # Copyright (c) 2021 Red Hat GmbH
 # Author: Stefano Brivio <sbrivio(a)redhat.com>
 
-nstools	ncat tee grep cat ip jq md5sum cut
-htools	printf dd ncat tee grep cat ip jq md5sum cut
+nstools	socat tee grep cat ip jq md5sum cut
+htools	printf dd socat tee grep cat ip jq md5sum cut
 
 test	UDP/IPv4: host to ns
 temp	TEMP
 temp	TEMP_NS
 temp	NC_PID
 ns	:> __TEMP_NS__
-nsb	(ncat -u -4 -l 127.0.0.1 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+nsb	(socat -u UDP4-LISTEN:10002,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 
-host	cat __TEMP__ | ncat -u 127.0.0.1 10002
+host	socat OPEN:__TEMP__ UDP4:127.0.0.1:10002
 nsw
 hout	MD5 md5sum __TEMP__ | cut -d' ' -f1
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
@@ -30,26 +30,26 @@ check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
 host	:> __TEMP__
-hostb	(ncat -u -4 -l 127.0.0.1 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
-ns	cat __TEMP_NS__ | ncat -u 127.0.0.1 10003
+ns	socat OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
 host	:> __TEMP__
-hostb	(ncat -u -4 -l 0.0.0.0 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
-ns	cat __TEMP_NS__ | ncat -u __GW__ 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
 ns	:> __TEMP_NS__
-nsb	(ncat -u -6 -l ::1 10002 & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
-host	cat __TEMP__ | ncat -u ::1 10002
+nsb	(socat -u UDP6-LISTEN:10002,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
 hout	MD5 md5sum __TEMP__ | cut -d' ' -f1
 nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
@@ -57,9 +57,9 @@ check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
 host	:> __TEMP__
-hostb	(ncat -u -6 -l ::1 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
-ns	cat __TEMP_NS__ | ncat -u ::1 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
@@ -67,10 +67,10 @@ nsw
 
 test	UDP/IPv6: ns to host (via tap)
 host	:> __TEMP__
-hostb	(ncat -u -6 -l :: 10003 & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
+hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-ns	cat __TEMP_NS__ | ncat -u __GW6__%__IFNAME__ 10003
+ns	socat -u OPEN:__TEMP_NS__ UDP6:[__GW6__%__IFNAME__]:10003
 hostw
 hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
-- 
2.36.1

[PATCH 12/18] tests: Remove unnecessary ^D in passt_in_ns teardown

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1197 bytes --]

teardown_passt_in_ns() sends a ^D to the NS pane, which appears to be
intended to terminate the nsenter running there, leaving the namespace.
However, we've also sent a ^D to the PASST pane which will exit the pasta
instance which created the namespace.  With the namespace destroyed the
nsenter in the NS pane will be killed, so it does not need to be exited
explicitly.

In fact sending the extra ^D can be harmful, since it will exit the shell
in which the nsenter was run, causing the whole pane to be closed.  That
can then mean that the "pane_wait NS" hangs indefinitely.  I believe this
will sometimes work, because there's a race between the various options
here, but it should be more reliable without the extra ^D.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/lib/setup | 2 --
 1 file changed, 2 deletions(-)

diff --git a/test/lib/setup b/test/lib/setup
index adbfa9e..752469c 100755
--- a/test/lib/setup
+++ b/test/lib/setup
@@ -306,8 +306,6 @@ teardown_passt_in_ns() {
 	[ ${VALGRIND} -eq 0 ] && pane_status GUEST
 	tmux send-keys -t ${PANE_PASST} "C-d"
 
-	tmux send-keys -t ${PANE_NS} "C-d"
-
 	pane_wait GUEST
 	pane_wait NS
 	pane_wait PASST
-- 
@@ -306,8 +306,6 @@ teardown_passt_in_ns() {
 	[ ${VALGRIND} -eq 0 ] && pane_status GUEST
 	tmux send-keys -t ${PANE_PASST} "C-d"
 
-	tmux send-keys -t ${PANE_NS} "C-d"
-
 	pane_wait GUEST
 	pane_wait NS
 	pane_wait PASST
-- 
2.36.1

[PATCH 13/18] tests: Remove unnecessary truncation of temporary files in udp tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 7893 bytes --]

All the UDP tests use :> to truncate some temporary data files.  This
appears to be so that they're empty before writing data to them with tee.
However tee, by default, truncates its output file anyway (you need tee -a
to append).  So drop the unnecessary truncations.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/udp/passt       |  2 --
 test/udp/passt_in_ns | 10 ----------
 test/udp/pasta       |  6 ------
 3 files changed, 18 deletions(-)

diff --git a/test/udp/passt b/test/udp/passt
index ee575e0..7c58648 100644
--- a/test/udp/passt
+++ b/test/udp/passt
@@ -26,7 +26,6 @@ gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
 
 test	UDP/IPv4: guest to host
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 guest	socat -u OPEN:test.bin UDP4:__GW__:10003
@@ -43,7 +42,6 @@ gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__HOST_MD5__" ]
 
 test	UDP/IPv6: guest to host
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
diff --git a/test/udp/passt_in_ns b/test/udp/passt_in_ns
index 3536298..8f8090b 100644
--- a/test/udp/passt_in_ns
+++ b/test/udp/passt_in_ns
@@ -28,7 +28,6 @@ check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: host to ns
 temp	TEMP_NS
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP4-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 host	socat -u OPEN:__TEMP__ UDP4:127.0.0.1:10002
 nsw
@@ -36,7 +35,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: guest to host
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 guest	socat -u OPEN:test.bin UDP4:__GW__:10003
@@ -45,7 +43,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: guest to ns
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP4-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 guest	socat -u OPEN:test.bin UDP4:__GW__:10002
 hostw
@@ -54,7 +51,6 @@ check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
 sleep	1
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 ns	socat -u OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
 hostw
@@ -62,7 +58,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
 hostw
@@ -93,7 +88,6 @@ gout	GUEST_MD5 md5sum test.bin | cut -d' ' -f1
 check	[ "__GUEST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP6-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
@@ -101,7 +95,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: guest to host
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 gout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
@@ -111,7 +104,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: guest to ns
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP6-LISTEN:10002 STDOUT & echo $! > __SC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 guest	socat -u OPEN:test.bin UDP6:[__GW6__%__IFNAME__]:10002
 hostw
@@ -119,7 +111,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
 hostw
@@ -127,7 +118,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __SC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __SC_PID__))
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
 ns	socat -u OPEN:__TEMP_NS__ UDP6:[__GW6__%__IFNAME__]:10003
diff --git a/test/udp/pasta b/test/udp/pasta
index 43e0513..25063ea 100644
--- a/test/udp/pasta
+++ b/test/udp/pasta
@@ -18,7 +18,6 @@ test	UDP/IPv4: host to ns
 temp	TEMP
 temp	TEMP_NS
 temp	NC_PID
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP4-LISTEN:10002,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 
@@ -29,7 +28,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
 ns	socat OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
@@ -38,7 +36,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
@@ -47,7 +44,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP6-LISTEN:10002,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
@@ -56,7 +52,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
 ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
@@ -66,7 +61,6 @@ check	[ "__HOST_MD5__" = "__MD5__" ]
 nsw
 
 test	UDP/IPv6: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-- 
@@ -18,7 +18,6 @@ test	UDP/IPv4: host to ns
 temp	TEMP
 temp	TEMP_NS
 temp	NC_PID
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP4-LISTEN:10002,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	dd if=/dev/urandom bs=1k count=5 > __TEMP__ && printf "\nEND_OF_TEST\n" >> __TEMP__
 
@@ -29,7 +28,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (recvmmsg/sendmmsg)
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003,bind=127.0.0.1 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
 ns	socat OPEN:__TEMP_NS__ UDP4:127.0.0.1:10003
@@ -38,7 +36,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv4: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP4-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW ip -j -4 route show|jq -rM '.[] | select(.dst == "default").gateway'
 ns	socat -u OPEN:__TEMP_NS__ UDP4:__GW__:10003
@@ -47,7 +44,6 @@ hout	HOST_MD5 md5sum __TEMP__ | cut -d' ' -f1
 check	[ "__HOST_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: host to ns
-ns	:> __TEMP_NS__
 nsb	(socat -u UDP6-LISTEN:10002,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP_NS__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 host	socat -u OPEN:__TEMP__ UDP6:[::1]:10002
 nsw
@@ -56,7 +52,6 @@ nsout	NS_MD5 md5sum __TEMP_NS__ | cut -d' ' -f1
 check	[ "__NS_MD5__" = "__MD5__" ]
 
 test	UDP/IPv6: ns to host (recvmmsg/sendmmsg)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003,bind=[::1] STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 sleep	1
 ns	socat -u OPEN:__TEMP_NS__ UDP6:[::1]:10003
@@ -66,7 +61,6 @@ check	[ "__HOST_MD5__" = "__MD5__" ]
 nsw
 
 test	UDP/IPv6: ns to host (via tap)
-host	:> __TEMP__
 hostb	(socat -u UDP6-LISTEN:10003 STDOUT & echo $! > __NC_PID__) | tee __TEMP__ | (grep -qm1 "END_OF_TEST" && kill $(cat __NC_PID__))
 nsout	GW6 ip -j -6 route show|jq -rM '.[] | select(.dst == "default").gateway'
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-- 
2.36.1

[PATCH 14/18] tests: Use dhclient --no-pid for namespaces in two_guests tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1241 bytes --]

Before starting the guests, these tests configure addresses in a pasta
namespace using dhclient.  However, because it's a user namespace, it's
not running as "real" root and can't write to the dhclient pid file.
This doesn't stop it working, but causes an ugly error message which we
can avoid by using the --no-pid option.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/lib/setup | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/test/lib/setup b/test/lib/setup
index 752469c..5c7b08f 100755
--- a/test/lib/setup
+++ b/test/lib/setup
@@ -233,13 +233,13 @@ setup_two_guests() {
 	pane_wait PASST_1
 	__ifname="$(pane_parse PASST_1)"
 
-	pane_run GUEST_1 "/sbin/dhclient -4 ${__ifname}"
-	pane_run GUEST_2 "/sbin/dhclient -4 ${__ifname}"
+	pane_run GUEST_1 "/sbin/dhclient -4 --no-pid ${__ifname}"
+	pane_run GUEST_2 "/sbin/dhclient -4 --no-pid ${__ifname}"
 	pane_status GUEST_1
 	pane_status GUEST_2
 	sleep 2
-	pane_run GUEST_1 "/sbin/dhclient -6 ${__ifname}"
-	pane_run GUEST_2 "/sbin/dhclient -6 ${__ifname}"
+	pane_run GUEST_1 "/sbin/dhclient -6 --no-pid ${__ifname}"
+	pane_run GUEST_2 "/sbin/dhclient -6 --no-pid ${__ifname}"
 	pane_status GUEST_1
 	pane_status GUEST_2
 
-- 
@@ -233,13 +233,13 @@ setup_two_guests() {
 	pane_wait PASST_1
 	__ifname="$(pane_parse PASST_1)"
 
-	pane_run GUEST_1 "/sbin/dhclient -4 ${__ifname}"
-	pane_run GUEST_2 "/sbin/dhclient -4 ${__ifname}"
+	pane_run GUEST_1 "/sbin/dhclient -4 --no-pid ${__ifname}"
+	pane_run GUEST_2 "/sbin/dhclient -4 --no-pid ${__ifname}"
 	pane_status GUEST_1
 	pane_status GUEST_2
 	sleep 2
-	pane_run GUEST_1 "/sbin/dhclient -6 ${__ifname}"
-	pane_run GUEST_2 "/sbin/dhclient -6 ${__ifname}"
+	pane_run GUEST_1 "/sbin/dhclient -6 --no-pid ${__ifname}"
+	pane_run GUEST_2 "/sbin/dhclient -6 --no-pid ${__ifname}"
 	pane_status GUEST_1
 	pane_status GUEST_2
 
-- 
2.36.1

[PATCH 15/18] tests: Clean up better after iperf tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 941 bytes --]

The iperf based test commands create a bunch of .bw and .pid files for
each iperf client and server.  The server side .bw files are cleaned
up afterwards, but the pid files are not, and none of the client side
files are cleaned up.  The latter doesn't really matter when the
client is run on ephemeral guests, but sometimes we run it in a
namespace that shares the filesystem with the host.

Clean up all of these files after the tests.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/lib/test | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/lib/test b/test/lib/test
index cb89e0b..e5a8632 100755
--- a/test/lib/test
+++ b/test/lib/test
@@ -79,7 +79,7 @@ test_iperf3() {
 	pane_wait "${__pane}"
 	pane_parse "${__pane}"
 	pane_run "${__pane}" 'for i in $(seq 0 '${__procs}'); do'	\
-		'rm -f s${i}.bw; done'
+		'rm -f [cs]${i}.bw [cs]${i}.pid; done'
 	pane_status "${__pane}"
 }
 
-- 
@@ -79,7 +79,7 @@ test_iperf3() {
 	pane_wait "${__pane}"
 	pane_parse "${__pane}"
 	pane_run "${__pane}" 'for i in $(seq 0 '${__procs}'); do'	\
-		'rm -f s${i}.bw; done'
+		'rm -f [cs]${i}.bw [cs]${i}.pid; done'
 	pane_status "${__pane}"
 }
 
-- 
2.36.1

[PATCH 16/18] tests: No need to retrieve host ifname in ndp/pasta

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1517 bytes --]

With pasta, the namespace interface name is generally the same as the host
interface name.  We already rely on this in the dhcp/pasta tests, but for
no clear reason ndp/pasta separately determines the host interface name.
Remove this unnecessary step.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/ndp/pasta | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/test/ndp/pasta b/test/ndp/pasta
index 2a2dd1a..28fdd72 100644
--- a/test/ndp/pasta
+++ b/test/ndp/pasta
@@ -16,15 +16,14 @@ htools	ip jq sipcalc grep cut
 
 test	Interface name
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
+check	[ -n "__IFNAME__" ]
 ns	ip link set dev __IFNAME__ up
 sleep	2
-hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
-check	[ -n "__IFNAME__" ]
 
 test	SLAAC: prefix
 nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global" and .prefixlen == 64).local'
 nsout	PREFIX6 sipcalc __ADDR6__/64 | grep prefix | cut -d' ' -f4
-hout	HOST_ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[] | select(.scope == "global").local'
+hout	HOST_ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global").local'
 hout	HOST_PREFIX6 sipcalc __HOST_ADDR6__/64 | grep prefix | cut -d' ' -f4
 check	[ "__PREFIX6__" = "__HOST_PREFIX6__" ]
 
-- 
@@ -16,15 +16,14 @@ htools	ip jq sipcalc grep cut
 
 test	Interface name
 nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
+check	[ -n "__IFNAME__" ]
 ns	ip link set dev __IFNAME__ up
 sleep	2
-hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
-check	[ -n "__IFNAME__" ]
 
 test	SLAAC: prefix
 nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global" and .prefixlen == 64).local'
 nsout	PREFIX6 sipcalc __ADDR6__/64 | grep prefix | cut -d' ' -f4
-hout	HOST_ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__HOST_IFNAME__").addr_info[] | select(.scope == "global").local'
+hout	HOST_ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global").local'
 hout	HOST_PREFIX6 sipcalc __HOST_ADDR6__/64 | grep prefix | cut -d' ' -f4
 check	[ "__PREFIX6__" = "__HOST_PREFIX6__" ]
 
-- 
2.36.1

[PATCH 17/18] tests: Correct determination of host interface name in tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1838 bytes --]

By default, passt itself attaches to the first host interface with a
default route.  However, when determining the host interface name the tests
implicitly select the *last* host interface: they use a jq expression which
will list all interfaces with default routes, but the way output detection
works in the scripts, it will only pick up the last line.

If there are multiple interfaces with default routes on the host, and they
each have a different address, this can cause spurious test failures.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/dhcp/passt       | 2 +-
 test/two_guests/basic | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/dhcp/passt b/test/dhcp/passt
index 3d2e939..f45227a 100644
--- a/test/dhcp/passt
+++ b/test/dhcp/passt
@@ -16,7 +16,7 @@ htools	ip jq sed tr head
 
 test	Interface name
 gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
+hout	HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]'
 check	[ -n "__IFNAME__" ]
 
 test	DHCP: address
diff --git a/test/two_guests/basic b/test/two_guests/basic
index cf0b975..f7c016d 100644
--- a/test/two_guests/basic
+++ b/test/two_guests/basic
@@ -18,7 +18,7 @@ htools	ip jq md5sum cut
 test	Interface names
 g1out	IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
 g2out	IFNAME2 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
+hout	HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]'
 check	[ -n "__IFNAME1__" ]
 check	[ -n "__IFNAME2__" ]
 
-- 
@@ -18,7 +18,7 @@ htools	ip jq md5sum cut
 test	Interface names
 g1out	IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
 g2out	IFNAME2 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
-hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
+hout	HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]'
 check	[ -n "__IFNAME1__" ]
 check	[ -n "__IFNAME2__" ]
 
-- 
2.36.1

[PATCH 18/18] demo: Use git protocol downloads

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1014 bytes --]

For some reason using https to clone from the passt git repo is very slow,
at least from network-distant places.  Use git protocol in the demo instead
to avoid a tedious wait to get the source.

Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
 test/demo/passt | 2 +-
 test/demo/pasta | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/demo/passt b/test/demo/passt
index dac8ac7..72744b1 100644
--- a/test/demo/passt
+++ b/test/demo/passt
@@ -22,7 +22,7 @@ say	Let's fetch the source
 sleep	1
 tempdir	TEMPDIR
 host	cd __TEMPDIR__
-host	git clone https://passt.top/passt
+host	git clone git://passt.top/passt
 sleep	1
 
 say	 and build it.
diff --git a/test/demo/pasta b/test/demo/pasta
index d84d856..9a5c381 100644
--- a/test/demo/pasta
+++ b/test/demo/pasta
@@ -22,7 +22,7 @@ say	Let's fetch the source
 sleep	1
 tempdir	TEMPDIR
 host	cd __TEMPDIR__
-host	git clone https://passt.top/passt
+host	git clone git://passt.top/passt
 sleep	1
 
 say	 and build it.
-- 
@@ -22,7 +22,7 @@ say	Let's fetch the source
 sleep	1
 tempdir	TEMPDIR
 host	cd __TEMPDIR__
-host	git clone https://passt.top/passt
+host	git clone git://passt.top/passt
 sleep	1
 
 say	 and build it.
-- 
2.36.1

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 2557 bytes --]

On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:
> By default, passt itself attaches to the first host interface with a
> default route.  However, when determining the host interface name the tests
> implicitly select the *last* host interface: they use a jq expression which
> will list all interfaces with default routes, but the way output detection
> works in the scripts, it will only pick up the last line.
> 
> If there are multiple interfaces with default routes on the host, and they
> each have a different address, this can cause spurious test
> failures.

It seems this change is not enough to always fix the tests when there
are multiple default routes.  I'm still sometimes getting failures,
now because passt itself doesn't seem to be picking the interface
with the first default route.

I'm wondering if this is because ip(8) is sorting the output, not just
presenting it in the same order that the underlying netlink interface
does.

> Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
> ---
>  test/dhcp/passt       | 2 +-
>  test/two_guests/basic | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/test/dhcp/passt b/test/dhcp/passt
> index 3d2e939..f45227a 100644
> --- a/test/dhcp/passt
> +++ b/test/dhcp/passt
> @@ -16,7 +16,7 @@ htools	ip jq sed tr head
>  
>  test	Interface name
>  gout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
> -hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
> +hout	HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]'
>  check	[ -n "__IFNAME__" ]
>  
>  test	DHCP: address
> diff --git a/test/two_guests/basic b/test/two_guests/basic
> index cf0b975..f7c016d 100644
> --- a/test/two_guests/basic
> +++ b/test/two_guests/basic
> @@ -18,7 +18,7 @@ htools	ip jq md5sum cut
>  test	Interface names
>  g1out	IFNAME1 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
>  g2out	IFNAME2 ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
> -hout	HOST_IFNAME ip -j -4 route show|jq -rM '.[] | select(.dst == "default").dev'
> +hout	HOST_IFNAME ip -j -4 route show|jq -rM '[.[] | select(.dst == "default").dev] | .[0]'
>  check	[ -n "__IFNAME1__" ]
>  check	[ -n "__IFNAME2__" ]
>  

-- 
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

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[Stefano Brivio]

[-- Attachment #1: Type: text/plain, Size: 2687 bytes --]

On Tue, 19 Jul 2022 16:20:45 +1000
David Gibson <david(a)gibson.dropbear.id.au> wrote:

> On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:
> > By default, passt itself attaches to the first host interface with a
> > default route.  However, when determining the host interface name the tests
> > implicitly select the *last* host interface: they use a jq expression which
> > will list all interfaces with default routes, but the way output detection
> > works in the scripts, it will only pick up the last line.
> > 
> > If there are multiple interfaces with default routes on the host, and they
> > each have a different address, this can cause spurious test
> > failures.  
> 
> It seems this change is not enough to always fix the tests when there
> are multiple default routes.  I'm still sometimes getting failures,
> now because passt itself doesn't seem to be picking the interface
> with the first default route.
> 
> I'm wondering if this is because ip(8) is sorting the output, not just
> presenting it in the same order that the underlying netlink interface
> does.

I don't see that happening at least in my environment (and I also can't
see any code that would sort it, that pretty much comes from
rtnl_dump_filter_l() in lib/libnetlink.c).

The netlink "filter", though, is slightly different. For IPv4:

$ strace -e sendto ip route show >/dev/null
sendto(3, [{{len=36, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=1658257027, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_UNSPEC, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNSPEC, rtm_flags=0}, {{nla_len=8, nla_type=RTA_TABLE}, RT_TABLE_MAIN}}, {len=0, type=0 /* NLMSG_??? */, flags=0, seq=0, pid=0}], 156, 0, NULL, 0) = 156

$ strace -e sendto ./passt -f
sendto(5, {{len=28, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=0, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_MAIN, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNICAST, rtm_flags=0}}, 28, 0, NULL, 0) = 28
[...]

and, while I don't think the FIB trie is actually descended in a
different way, we might still have a slightly different result from the
kernel (if I recall correctly, I didn't check right now).

But letting that aside for a moment: if you have two default routes, I
suppose they have different metrics. If not, what's the intended usage?

If yes, we should probably implement a sorting logic in passt, so that
the route with the lowest metric is picked, and then adjust the jq
expression to also pick that one.

-- 
Stefano

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 3255 bytes --]

On Tue, Jul 19, 2022 at 09:05:52PM +0200, Stefano Brivio wrote:
> On Tue, 19 Jul 2022 16:20:45 +1000
> David Gibson <david(a)gibson.dropbear.id.au> wrote:
> 
> > On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:
> > > By default, passt itself attaches to the first host interface with a
> > > default route.  However, when determining the host interface name the tests
> > > implicitly select the *last* host interface: they use a jq expression which
> > > will list all interfaces with default routes, but the way output detection
> > > works in the scripts, it will only pick up the last line.
> > > 
> > > If there are multiple interfaces with default routes on the host, and they
> > > each have a different address, this can cause spurious test
> > > failures.  
> > 
> > It seems this change is not enough to always fix the tests when there
> > are multiple default routes.  I'm still sometimes getting failures,
> > now because passt itself doesn't seem to be picking the interface
> > with the first default route.
> > 
> > I'm wondering if this is because ip(8) is sorting the output, not just
> > presenting it in the same order that the underlying netlink interface
> > does.
> 
> I don't see that happening at least in my environment (and I also can't
> see any code that would sort it, that pretty much comes from
> rtnl_dump_filter_l() in lib/libnetlink.c).
> 
> The netlink "filter", though, is slightly different. For IPv4:
> 
> $ strace -e sendto ip route show >/dev/null
> sendto(3, [{{len=36, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=1658257027, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_UNSPEC, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNSPEC, rtm_flags=0}, {{nla_len=8, nla_type=RTA_TABLE}, RT_TABLE_MAIN}}, {len=0, type=0 /* NLMSG_??? */, flags=0, seq=0, pid=0}], 156, 0, NULL, 0) = 156
> 
> $ strace -e sendto ./passt -f
> sendto(5, {{len=28, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=0, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_MAIN, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNICAST, rtm_flags=0}}, 28, 0, NULL, 0) = 28
> [...]
> 
> and, while I don't think the FIB trie is actually descended in a
> different way, we might still have a slightly different result from the
> kernel (if I recall correctly, I didn't check right now).
> 
> But letting that aside for a moment: if you have two default routes, I
> suppose they have different metrics. If not, what's the intended usage?

Yes, there are two different metrics.  I was thinking after I sent
this that we should sort by metric.

> If yes, we should probably implement a sorting logic in passt, so that
> the route with the lowest metric is picked, and then adjust the jq
> expression to also pick that one.

Agreed.  I'll see if I can implement this.  Up to you whether you drop
this patch or I just do another update on top of it.

-- 
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

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 6284 bytes --]

On Wed, Jul 20, 2022 at 12:47:57PM +1000, David Gibson wrote:
> On Tue, Jul 19, 2022 at 09:05:52PM +0200, Stefano Brivio wrote:
> > On Tue, 19 Jul 2022 16:20:45 +1000
> > David Gibson <david(a)gibson.dropbear.id.au> wrote:
> > 
> > > On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:
> > > > By default, passt itself attaches to the first host interface with a
> > > > default route.  However, when determining the host interface name the tests
> > > > implicitly select the *last* host interface: they use a jq expression which
> > > > will list all interfaces with default routes, but the way output detection
> > > > works in the scripts, it will only pick up the last line.
> > > > 
> > > > If there are multiple interfaces with default routes on the host, and they
> > > > each have a different address, this can cause spurious test
> > > > failures.  
> > > 
> > > It seems this change is not enough to always fix the tests when there
> > > are multiple default routes.  I'm still sometimes getting failures,
> > > now because passt itself doesn't seem to be picking the interface
> > > with the first default route.
> > > 
> > > I'm wondering if this is because ip(8) is sorting the output, not just
> > > presenting it in the same order that the underlying netlink interface
> > > does.
> > 
> > I don't see that happening at least in my environment (and I also can't
> > see any code that would sort it, that pretty much comes from
> > rtnl_dump_filter_l() in lib/libnetlink.c).
> > 
> > The netlink "filter", though, is slightly different. For IPv4:
> > 
> > $ strace -e sendto ip route show >/dev/null
> > sendto(3, [{{len=36, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=1658257027, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_UNSPEC, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNSPEC, rtm_flags=0}, {{nla_len=8, nla_type=RTA_TABLE}, RT_TABLE_MAIN}}, {len=0, type=0 /* NLMSG_??? */, flags=0, seq=0, pid=0}], 156, 0, NULL, 0) = 156
> > 
> > $ strace -e sendto ./passt -f
> > sendto(5, {{len=28, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=0, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_MAIN, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNICAST, rtm_flags=0}}, 28, 0, NULL, 0) = 28
> > [...]
> > 
> > and, while I don't think the FIB trie is actually descended in a
> > different way, we might still have a slightly different result from the
> > kernel (if I recall correctly, I didn't check right now).

So, I've been staring at this for most of the day.  I don't think the
difference is because of the slightly different filters.  Instead it's
an easy to miss subtletly in nl_get_ext_if().

In nl_get_ext_if() both first_v4 and first_v6 will be set to the
interface which meets the criteria and comes first in the order of
routes returned by netlink from the kernel.  However, there's this early return:

	word = (long *)has_v4;
	for (i = 0; i < ARRAY_SIZE(has_v4) / sizeof(long); i++, word++) {
		tmp = *word;
		while ((n = ffsl(tmp))) {
			int ifi = i * sizeof(long) * 8 + n - 1;

			if (!first_v4)
				first_v4 = ifi;

			tmp &= ~(1UL << (n - 1));
			if (bitmap_isset(has_v6, ifi)) {
				*v4 = *v6 = IP_VERSION_ENABLED;
				return ifi;    <================ HERE
			}
		}
	}

Which means that if we find an interface with both a v4 and a v6
default route, we'll return it immediately.  That loop is across the
has_v4 bitmap, so we'll return the first one in *interface index order*,
which might not necessarily be the same as the order the routes are
returned from netlink.

In the other cases: one of v4 or v6 is disabled, or there's no
interface with both v4 and v6 default routes, then we'll end up
returning the first interface in netlink route order.  Which makes
passt's overall priority of selecting interfaces pretty hard to
follow.

I've been thinking a bunch about what the order should be.  The
obvious cleanup to the current logic would be:
	1. If there are >0 interfaces with both default v4 and v6
	   routes, pick the one with the lowest metric
	2. Otherwise if there are >0 interfaces with a default v4
	   route, pick the one with the lowest metric
	3. Otherwise if there are >0 interfaces with a default v6
	   route, pick the one with the lowest metric
	4. Otherwise fail.

..but, I think that's the wrong choice.  For one thing it gets weird
in the unlikely but possible case that there are two interfaces with
both v4 and v6 routes, but one has a lower v4 metric and the other has
a lower v6 metric.  But more importantly, I think it fails the "do
what you want by default" test in the case of a host that has both v4
and v6 routing, but not on the same interface.  That's quite a
realistic situation if the v6 connectivity is coming from a tunnel
(6in4 or 6rd or 6over4 or whatever).

Instead, I think we should separately pick an external interface for
v4 and v6 connectivity.  Yes, it's weird to have a view of two
different host interfaces on the same virtual interface.  It also
means pasta has to arbitrarily pick one of the host interface names
for the namespace, which will make the tests a bit uglier.  But it
makes picking the right interfaces a lot simpler and clearer, and does
what you probably want for the case of a host with tunnelled ipv6
connectivity.

> > But letting that aside for a moment: if you have two default routes, I
> > suppose they have different metrics. If not, what's the intended usage?
> 
> Yes, there are two different metrics.  I was thinking after I sent
> this that we should sort by metric.
> 
> > If yes, we should probably implement a sorting logic in passt, so that
> > the route with the lowest metric is picked, and then adjust the jq
> > expression to also pick that one.
> 
> Agreed.  I'll see if I can implement this.  Up to you whether you drop
> this patch or I just do another update on top of it.
> 



-- 
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

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[Stefano Brivio]

[-- Attachment #1: Type: text/plain, Size: 6549 bytes --]

On Wed, 20 Jul 2022 17:24:18 +1000
David Gibson <david(a)gibson.dropbear.id.au> wrote:

> On Wed, Jul 20, 2022 at 12:47:57PM +1000, David Gibson wrote:
> > On Tue, Jul 19, 2022 at 09:05:52PM +0200, Stefano Brivio wrote:  
> > > On Tue, 19 Jul 2022 16:20:45 +1000
> > > David Gibson <david(a)gibson.dropbear.id.au> wrote:
> > >   
> > > > On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:  
> > > > > By default, passt itself attaches to the first host interface with a
> > > > > default route.  However, when determining the host interface name the tests
> > > > > implicitly select the *last* host interface: they use a jq expression which
> > > > > will list all interfaces with default routes, but the way output detection
> > > > > works in the scripts, it will only pick up the last line.
> > > > > 
> > > > > If there are multiple interfaces with default routes on the host, and they
> > > > > each have a different address, this can cause spurious test
> > > > > failures.    
> > > > 
> > > > It seems this change is not enough to always fix the tests when there
> > > > are multiple default routes.  I'm still sometimes getting failures,
> > > > now because passt itself doesn't seem to be picking the interface
> > > > with the first default route.
> > > > 
> > > > I'm wondering if this is because ip(8) is sorting the output, not just
> > > > presenting it in the same order that the underlying netlink interface
> > > > does.  
> > > 
> > > I don't see that happening at least in my environment (and I also can't
> > > see any code that would sort it, that pretty much comes from
> > > rtnl_dump_filter_l() in lib/libnetlink.c).
> > > 
> > > The netlink "filter", though, is slightly different. For IPv4:
> > > 
> > > $ strace -e sendto ip route show >/dev/null
> > > sendto(3, [{{len=36, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=1658257027, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_UNSPEC, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNSPEC, rtm_flags=0}, {{nla_len=8, nla_type=RTA_TABLE}, RT_TABLE_MAIN}}, {len=0, type=0 /* NLMSG_??? */, flags=0, seq=0, pid=0}], 156, 0, NULL, 0) = 156
> > > 
> > > $ strace -e sendto ./passt -f
> > > sendto(5, {{len=28, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=0, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_MAIN, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNICAST, rtm_flags=0}}, 28, 0, NULL, 0) = 28
> > > [...]
> > > 
> > > and, while I don't think the FIB trie is actually descended in a
> > > different way, we might still have a slightly different result from the
> > > kernel (if I recall correctly, I didn't check right now).  
> 
> So, I've been staring at this for most of the day.  I don't think the
> difference is because of the slightly different filters.  Instead it's
> an easy to miss subtletly in nl_get_ext_if().
> 
> In nl_get_ext_if() both first_v4 and first_v6 will be set to the
> interface which meets the criteria and comes first in the order of
> routes returned by netlink from the kernel.  However, there's this early return:
> 
> 	word = (long *)has_v4;
> 	for (i = 0; i < ARRAY_SIZE(has_v4) / sizeof(long); i++, word++) {
> 		tmp = *word;
> 		while ((n = ffsl(tmp))) {
> 			int ifi = i * sizeof(long) * 8 + n - 1;
> 
> 			if (!first_v4)
> 				first_v4 = ifi;
> 
> 			tmp &= ~(1UL << (n - 1));
> 			if (bitmap_isset(has_v6, ifi)) {
> 				*v4 = *v6 = IP_VERSION_ENABLED;
> 				return ifi;    <================ HERE
> 			}
> 		}
> 	}
> 
> Which means that if we find an interface with both a v4 and a v6
> default route, we'll return it immediately.  That loop is across the
> has_v4 bitmap, so we'll return the first one in *interface index order*,
> which might not necessarily be the same as the order the routes are
> returned from netlink.

Whoops :) ...right, sorry, I forgot about that.

> In the other cases: one of v4 or v6 is disabled, or there's no
> interface with both v4 and v6 default routes, then we'll end up
> returning the first interface in netlink route order.  Which makes
> passt's overall priority of selecting interfaces pretty hard to
> follow.
> 
> I've been thinking a bunch about what the order should be.  The
> obvious cleanup to the current logic would be:
> 	1. If there are >0 interfaces with both default v4 and v6
> 	   routes, pick the one with the lowest metric
> 	2. Otherwise if there are >0 interfaces with a default v4
> 	   route, pick the one with the lowest metric
> 	3. Otherwise if there are >0 interfaces with a default v6
> 	   route, pick the one with the lowest metric
> 	4. Otherwise fail.
> 
> ..but, I think that's the wrong choice.  For one thing it gets weird
> in the unlikely but possible case that there are two interfaces with
> both v4 and v6 routes, but one has a lower v4 metric and the other has
> a lower v6 metric.  But more importantly, I think it fails the "do
> what you want by default" test in the case of a host that has both v4
> and v6 routing, but not on the same interface.  That's quite a
> realistic situation if the v6 connectivity is coming from a tunnel
> (6in4 or 6rd or 6over4 or whatever).

Definitely, I didn't consider the tunneled IPv6 case, but it looks
still rather real even nowadays.

> Instead, I think we should separately pick an external interface for
> v4 and v6 connectivity.  Yes, it's weird to have a view of two
> different host interfaces on the same virtual interface.  It also
> means pasta has to arbitrarily pick one of the host interface names
> for the namespace, which will make the tests a bit uglier.  But it
> makes picking the right interfaces a lot simpler and clearer, and does
> what you probably want for the case of a host with tunnelled ipv6
> connectivity.

I totally agree. It's still less weird than the alternative.

We'll need to store two interface indices (say, "ifi4" and "ifi6"), but
their usage right now is almost entirely limited to configuration and
netlink functions.

One additional usage is to select a given interface for sockets bound
to link-local addresses by means of setting .sin6_scope_id in struct
sockaddr_in6 -- that happens in tcp_conn_from_tap() and sock_l4().
There, we just need to switch to the new "ifi6" index.

Let me know if I should start looking into this or if you can do it
instead.

-- 
Stefano

Re: [PATCH 17/18] tests: Correct determination of host interface name in tests

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 7377 bytes --]

On Wed, Jul 20, 2022 at 10:23:22AM +0200, Stefano Brivio wrote:
> On Wed, 20 Jul 2022 17:24:18 +1000
> David Gibson <david(a)gibson.dropbear.id.au> wrote:
> 
> > On Wed, Jul 20, 2022 at 12:47:57PM +1000, David Gibson wrote:
> > > On Tue, Jul 19, 2022 at 09:05:52PM +0200, Stefano Brivio wrote:  
> > > > On Tue, 19 Jul 2022 16:20:45 +1000
> > > > David Gibson <david(a)gibson.dropbear.id.au> wrote:
> > > >   
> > > > > On Fri, Jul 15, 2022 at 03:21:40PM +1000, David Gibson wrote:  
> > > > > > By default, passt itself attaches to the first host interface with a
> > > > > > default route.  However, when determining the host interface name the tests
> > > > > > implicitly select the *last* host interface: they use a jq expression which
> > > > > > will list all interfaces with default routes, but the way output detection
> > > > > > works in the scripts, it will only pick up the last line.
> > > > > > 
> > > > > > If there are multiple interfaces with default routes on the host, and they
> > > > > > each have a different address, this can cause spurious test
> > > > > > failures.    
> > > > > 
> > > > > It seems this change is not enough to always fix the tests when there
> > > > > are multiple default routes.  I'm still sometimes getting failures,
> > > > > now because passt itself doesn't seem to be picking the interface
> > > > > with the first default route.
> > > > > 
> > > > > I'm wondering if this is because ip(8) is sorting the output, not just
> > > > > presenting it in the same order that the underlying netlink interface
> > > > > does.  
> > > > 
> > > > I don't see that happening at least in my environment (and I also can't
> > > > see any code that would sort it, that pretty much comes from
> > > > rtnl_dump_filter_l() in lib/libnetlink.c).
> > > > 
> > > > The netlink "filter", though, is slightly different. For IPv4:
> > > > 
> > > > $ strace -e sendto ip route show >/dev/null
> > > > sendto(3, [{{len=36, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=1658257027, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_UNSPEC, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNSPEC, rtm_flags=0}, {{nla_len=8, nla_type=RTA_TABLE}, RT_TABLE_MAIN}}, {len=0, type=0 /* NLMSG_??? */, flags=0, seq=0, pid=0}], 156, 0, NULL, 0) = 156
> > > > 
> > > > $ strace -e sendto ./passt -f
> > > > sendto(5, {{len=28, type=RTM_GETROUTE, flags=NLM_F_REQUEST|NLM_F_DUMP, seq=0, pid=0}, {rtm_family=AF_INET, rtm_dst_len=0, rtm_src_len=0, rtm_tos=0, rtm_table=RT_TABLE_MAIN, rtm_protocol=RTPROT_UNSPEC, rtm_scope=RT_SCOPE_UNIVERSE, rtm_type=RTN_UNICAST, rtm_flags=0}}, 28, 0, NULL, 0) = 28
> > > > [...]
> > > > 
> > > > and, while I don't think the FIB trie is actually descended in a
> > > > different way, we might still have a slightly different result from the
> > > > kernel (if I recall correctly, I didn't check right now).  
> > 
> > So, I've been staring at this for most of the day.  I don't think the
> > difference is because of the slightly different filters.  Instead it's
> > an easy to miss subtletly in nl_get_ext_if().
> > 
> > In nl_get_ext_if() both first_v4 and first_v6 will be set to the
> > interface which meets the criteria and comes first in the order of
> > routes returned by netlink from the kernel.  However, there's this early return:
> > 
> > 	word = (long *)has_v4;
> > 	for (i = 0; i < ARRAY_SIZE(has_v4) / sizeof(long); i++, word++) {
> > 		tmp = *word;
> > 		while ((n = ffsl(tmp))) {
> > 			int ifi = i * sizeof(long) * 8 + n - 1;
> > 
> > 			if (!first_v4)
> > 				first_v4 = ifi;
> > 
> > 			tmp &= ~(1UL << (n - 1));
> > 			if (bitmap_isset(has_v6, ifi)) {
> > 				*v4 = *v6 = IP_VERSION_ENABLED;
> > 				return ifi;    <================ HERE
> > 			}
> > 		}
> > 	}
> > 
> > Which means that if we find an interface with both a v4 and a v6
> > default route, we'll return it immediately.  That loop is across the
> > has_v4 bitmap, so we'll return the first one in *interface index order*,
> > which might not necessarily be the same as the order the routes are
> > returned from netlink.
> 
> Whoops :) ...right, sorry, I forgot about that.
> 
> > In the other cases: one of v4 or v6 is disabled, or there's no
> > interface with both v4 and v6 default routes, then we'll end up
> > returning the first interface in netlink route order.  Which makes
> > passt's overall priority of selecting interfaces pretty hard to
> > follow.
> > 
> > I've been thinking a bunch about what the order should be.  The
> > obvious cleanup to the current logic would be:
> > 	1. If there are >0 interfaces with both default v4 and v6
> > 	   routes, pick the one with the lowest metric
> > 	2. Otherwise if there are >0 interfaces with a default v4
> > 	   route, pick the one with the lowest metric
> > 	3. Otherwise if there are >0 interfaces with a default v6
> > 	   route, pick the one with the lowest metric
> > 	4. Otherwise fail.
> > 
> > ..but, I think that's the wrong choice.  For one thing it gets weird
> > in the unlikely but possible case that there are two interfaces with
> > both v4 and v6 routes, but one has a lower v4 metric and the other has
> > a lower v6 metric.  But more importantly, I think it fails the "do
> > what you want by default" test in the case of a host that has both v4
> > and v6 routing, but not on the same interface.  That's quite a
> > realistic situation if the v6 connectivity is coming from a tunnel
> > (6in4 or 6rd or 6over4 or whatever).
> 
> Definitely, I didn't consider the tunneled IPv6 case, but it looks
> still rather real even nowadays.

Yeah, I gather IPv6 is finally fairly standard in Germany, but here in
Australia I'm with basically the only residential ISP which provides v6.

> > Instead, I think we should separately pick an external interface for
> > v4 and v6 connectivity.  Yes, it's weird to have a view of two
> > different host interfaces on the same virtual interface.  It also
> > means pasta has to arbitrarily pick one of the host interface names
> > for the namespace, which will make the tests a bit uglier.  But it
> > makes picking the right interfaces a lot simpler and clearer, and does
> > what you probably want for the case of a host with tunnelled ipv6
> > connectivity.
> 
> I totally agree. It's still less weird than the alternative.

Excellent.

> We'll need to store two interface indices (say, "ifi4" and "ifi6"), but
> their usage right now is almost entirely limited to configuration and
> netlink functions.

Yes, I looked through all the uses of c->ifi, and it looks fairly
straightforward.

> One additional usage is to select a given interface for sockets bound
> to link-local addresses by means of setting .sin6_scope_id in struct
> sockaddr_in6 -- that happens in tcp_conn_from_tap() and sock_l4().
> There, we just need to switch to the new "ifi6" index.
> 
> Let me know if I should start looking into this or if you can do it
> instead.

I'll work on it.  Hoping I can get that done before heading on holiday
next week.

-- 
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

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 00/18] Test fixes, batch 5

[Stefano Brivio]

[-- Attachment #1: Type: text/plain, Size: 806 bytes --]

On Fri, 15 Jul 2022 15:21:23 +1000
David Gibson <david(a)gibson.dropbear.id.au> wrote:

> Here's yet another batch of fixes to make the tests more robust
> against different environments.  With this lot, I'm now able to run
> the pasta, passt, passt_in_ns and two_guests tests on my Fedora
> system.  I'm still hitting problems with the perf tests.
> 
> This series (specifically 11/18) updates the demo to use socat instead
> of openbsd netcat.  This is needed, or the change to socat in the
> mbuto image would break the demo.  However, I'm hitting unrelated
> problems trying to run the demos, so the switch to socat is, alas,
> untested.

I'm testing this now (excluding 17/18) together with the demo fix-up
series I just sent, hopefully some of the issues you found there are
also gone.

-- 
Stefano