- dev - passt

[PATCH v2 0/2] SO_PEEK_OFF support
by Jon Maloy 06 May '24

06 May '24

SO_PEEK_OFF support and a fix for the zero-window problem. Jon Maloy (2): tcp: leverage support of SO_PEEK_OFF socket option when available tcp: allow retransmit when peer receive window is zero tcp.c | 87 ++++++++++++++++++++++++++++++++++++++++++++++-------- tcp_conn.h | 2 ++ 2 files changed, 77 insertions(+), 12 deletions(-) -- 2.42.0

3 9

[PATCH v2 0/7] Rework UDP buffers
by David Gibson 02 May '24

02 May '24

Laurent recently reworked the TCP buffer handling to be split into various pieces tracked by iovecs. We'll want that for various future changes. This series makes a similar split for UDP buffers, which we'll want in order to allow dual-stack UDP sockets, amongst other things. This is based on my earlier series of cleanups for the TCP buffer handling. Changes since v1: * Rebase on v2 of the TCP iov and buffer handling cleanups * Rename udp_payload and udp_meta structures to _t for more consistency with tcp.c * Other minor changes based on Stefano's review David Gibson (7): test: Allow sftp via vsock-ssh in tests udp: Split tap-bound UDP packets into multiple buffers using io vector udp: Combine initialisation of IPv4 and IPv6 iovs udp: Explicitly set checksum in guest-bound UDP headers udp: Share payload buffers between IPv4 and IPv6 udp: Use the same buffer for the L2 header for all frames udp: Single buffer for IPv4, IPv6 headers and metadata tap.h | 38 ------ test/passt.mbuto | 6 +- udp.c | 299 +++++++++++++++++++++++------------------------ 3 files changed, 152 insertions(+), 191 deletions(-) -- 2.44.0

2 8

[PATCH v2 00/10] Small improvements to IOV handling
by David Gibson 02 May '24

02 May '24

Laurent's changes to split TCP buffers into various components with IOVs is now merged. This series has a batch of small cleanups to make the handling of this slightly nicer. These are preliminaries to doing something similar with the UDP buffers. Note that patch 10/10 might interfere with the experiments to work out what is going wrong with the odd batching / performance issues we've seen. We can leave it off for the time being if that's a problem. Changes since v1: * Added new patch removing some unused structures * Added two new patches cleaning up some endianness confusion * Added a bunch of missing cases to standardisation of length variable names * Assorted minor revisions based on Stefano's review David Gibson (10): checksum: Use proto_ipv6_header_psum() for ICMPv6 as well tap: Split tap specific and L2 (ethernet) headers tap: Remove unused structs tap_msg, tap_l4_msg treewide: Remove misleading and redundant endianness notes checksum: Make csum_ip4_header() take a host endian length treewide: Standardise variable names for various packet lengths tcp: Simplify packet length calculation when preparing headers tap, tcp: (Re-)abstract TAP specific header handling iov: Helper macro to construct iovs covering existing variables or fields tcp: Update tap specific header too in tcp_fill_headers[46]() arp.c | 6 +- checksum.c | 68 ++++++++++----------- checksum.h | 12 ++-- dhcp.c | 6 +- icmp.c | 12 ++-- iov.h | 3 + ndp.c | 6 +- passt.h | 28 ++------- pcap.c | 14 ++--- pcap.h | 2 +- tap.c | 146 ++++++++++++++++++++++---------------------- tap.h | 60 ++++++++++++------ tcp.c | 176 +++++++++++++++++++++++------------------------------ udp.c | 52 ++++++++-------- 14 files changed, 290 insertions(+), 301 deletions(-) -- 2.44.0

2 11

[PATCH 0/2] netlink: Don't duplicate routes with mismatching interfaces
by Stefano Brivio 01 May '24

01 May '24

Stefano Brivio (2): netlink: Fix iterations over nexthop objects netlink: Don't duplicate routes referring to unrelated host interfaces netlink.c | 53 ++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 44 insertions(+), 9 deletions(-) -- 2.43.0

2 4

[PATCH 0/7] Rework UDP buffers
by David Gibson 01 May '24

01 May '24

Laurent recently reworked the TCP buffer handling to be split into various pieces tracked by iovecs. We'll want that for various future changes. This series makes a similar split for UDP buffers, which we'll want in order to allow dual-stack UDP sockets, amongst other things. This is based on my earlier series of cleanups for the TCP buffer handling. David Gibson (7): test: Allow sftp via vsock-ssh in tests udp: Split tap-bound UDP packets into multiple buffers using io vector udp: Combine initialisation of IPv4 and IPv6 iovs udp: Explicitly set checksum in guest-bound UDP headers udp: Share payload buffers between IPv4 and IPv6 udp: Use the same buffer for the L2 header for all frames udp: Single buffer for IPv4, IPv6 headers and metadata tap.h | 38 ------ test/passt.mbuto | 6 +- udp.c | 304 +++++++++++++++++++++++------------------------ 3 files changed, 153 insertions(+), 195 deletions(-) -- 2.44.0

2 13

[PATCH 0/7] Small improvements to IOV handling
by David Gibson 01 May '24

01 May '24

Laurent's changes to split TCP buffers into various components with IOVs is now merged. This series has a batch of small cleanups to make the handling of this slightly nicer. These are preliminaries to doing something similar with the UDP buffers. Note that patch 7/7 might interfere with the experiments to work out what is going wrong with the odd batching / performance issues we've seen. We can leave it off for the time being if that's a problem. David Gibson (7): checksum: Use proto_ipv6_header_psum() for ICMPv6 as well tap: Split tap specific and L2 (ethernet) headers treewide: Standardise variable names for various packet lengths tcp: Simplify packet length calculation when preparing headers tap, tcp: (Re-)abstract TAP specific header handling iov: Helper macro to construct iovs covering existing variables or fields tcp: Update tap specific header too in tcp_fill_headers[46]() checksum.c | 46 ++++++++--------- checksum.h | 10 ++-- icmp.c | 8 +-- iov.h | 3 ++ passt.h | 4 +- tap.c | 51 ++++++++++--------- tap.h | 46 +++++++++++++---- tcp.c | 144 +++++++++++++++++++++++------------------------------ udp.c | 50 ++++++++++--------- 9 files changed, 184 insertions(+), 178 deletions(-) -- 2.44.0

2 17

[PATCH 0/2] Support for SO_PEEK_OFF when a available
by Jon Maloy 29 Apr '24

29 Apr '24

First commit introduces support for SO_PEEK_OFF socket option. Second commit adds a workaround for a kernel TCP bug. Jon Maloy (2): tcp: leverage support of SO_PEEK_OFF socket option when available tcp: allow retransmit when peer receive window is zero tcp.c | 63 +++++++++++++++++++++++++++++++++++++++++++----------- tcp_conn.h | 2 ++ 2 files changed, 53 insertions(+), 12 deletions(-) -- 2.42.0

3 14

passt: new version 2024_04_26.d03c4e2 available
by Stefano Brivio 26 Apr '24

26 Apr '24

The new version with tag 2024_04_26.d03c4e2 includes the following changes: d03c4e2 netlink: Use IFA_F_NODAD also while duplicating addresses from the host bfc83b5 netlink: For IPv4, IFA_LOCAL is the interface address, not IFA_ADDRESS ff2ff2f test: Make log truncation test more robust 2681366 test: Slight simplification to pasta log tests 0804fdb udp: Correctly look up outbound socket with port remappings 9560123 tcp: Replace TCP buffer structure by an iovec array 27f1c76 conf: Don't fail if the template interface doesn't have a MAC address eca8baa conf: We're interested in the MAC address, not in the MAC itself ee338a2 pasta, util: Align stack area for clones to maximum natural alignment 5d5208b treewide: Compilers' name for armv6l and armv7l is "arm" https://passt.top/passt/log/?qt=range&q=2024_04_05.954589b..2024_04_26.d03c… Packages: - Alpine Linux: https://pkgs.alpinelinux.org/packages?name=passt - Arch Linux: https://www.archlinux.org/packages/extra/x86_64/passt/ https://archlinuxarm.org/packages/aarch64/passt https://archlinuxarm.org/packages/armv7h/passt - Chimera: https://pkgs.chimera-linux.org/packages?name=passt - Debian tracker: https://tracker.debian.org/pkg/passt - Copr (CentOS Stream, EPEL, Fedora, Mageia): https://copr.fedorainfracloud.org/coprs/sbrivio/passt/build/7374013/ permanent mirror: https://passt.top/builds/copr/0^20240426.gd03c4e2/ - Fedora updates: https://bodhi.fedoraproject.org/updates/?packages=passt - Gentoo versions: https://packages.gentoo.org/packages/net-misc/passt - GNU Guix: https://packages.guix.gnu.org/packages/passt/ - openSUSE: https://build.opensuse.org/package/requests/Virtualization:containers/passt - Ubuntu tracker: https://launchpad.net/ubuntu/+source/passt - Void Linux: https://voidlinux.org/packages/?q=passt - Static builds: - Package for other RPM-based distributions, x86_64 only: https://passt.top/builds/latest/x86_64/passt-gd03c4e2-1.x86_64.rpm - x86_64 static binaries: https://passt.top/builds/latest/x86_64/ - Debian package, from x86_64 static build: https://passt.top/builds/latest/x86_64/passt_0d03c4e2-1_all.deb -- Stefano

1 0

[PATCH] netlink: For IPv4, IFA_LOCAL is the interface address, not IFA_ADDRESS
by Stefano Brivio 26 Apr '24

26 Apr '24

See the comment to the unnamed enum in linux/if_addr.h, which currently states: /* * Important comment: * IFA_ADDRESS is prefix address, rather than local interface address. * It makes no difference for normally configured broadcast interfaces, * but for point-to-point IFA_ADDRESS is DESTINATION address, * local address is supplied in IFA_LOCAL attribute. * * [...] */ if we fetch IFA_ADDRESS, and we have a point-to-point link with a peer address configured, we'll source the peer address as "our" address, and refuse to resolve it in arp(). This was reported with pasta and a tun upstream interface configured by OpenVPN in "p2p" topology: the target namespace will have similar addresses and routes as the host, which is fine, and will try to resolve the point-to-point peer address (because it's the default gateway). Given that we configure it as our address (only internally, not for visibly in the namespace), we'll fail to resolve that and traffic doesn't go anywhere. Note that this is not the case for IPv6: there, IFA_ADDRESS is the actual, local address of the interface, and IFA_LOCAL is not necessarily present, so the comment in linux/if_addr.h doesn't apply either. Link: https://github.com/containers/podman/issues/22320 Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> --- netlink.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/netlink.c b/netlink.c index 89c0641..447fea2 100644 --- a/netlink.c +++ b/netlink.c @@ -668,7 +668,8 @@ int nl_addr_get(int s, unsigned int ifi, sa_family_t af, for (rta = IFA_RTA(ifa), na = IFA_PAYLOAD(nh); RTA_OK(rta, na); rta = RTA_NEXT(rta, na)) { - if (rta->rta_type != IFA_ADDRESS) + if ((af == AF_INET && rta->rta_type != IFA_LOCAL) || + (af == AF_INET6 && rta->rta_type != IFA_ADDRESS)) continue; if (af == AF_INET && ifa->ifa_prefixlen > prefix_max) { -- 2.43.0

2 1

[PATCH 0/4] mbuto: Assorted fixes and simplifications
by David Gibson 26 Apr '24

26 Apr '24

While debugging a problem Laurent Vivier was having with the passt test suite, I discovered mbuto's archivemount support is entirely broken: if archivemount is on the system, mbuto won't work. While fixing that, I discovered some slightly related simplfications that can be made. Here they all are. David Gibson (4): ${wd} is always set, no need to test for it Remove stale archivemount support Split "auto" compression mode into its own path Remove unnecessary cpio_init function mbuto | 120 +++++++++++++++------------------------------------------- 1 file changed, 30 insertions(+), 90 deletions(-) -- 2.44.0

2 15

2024

2023

2022

2021

dev