Page MenuHomeFreeBSD

rcm (R. Christian McDonald)
User

Projects (7)

User Details

User Since
Apr 13 2023, 7:00 PM (187 w, 6 d)

Recent Activity

Today

rcm added inline comments to D61725: arp: Teach arp to attach and run itself in a jail.
Thu, Nov 19, 6:30 AM
rcm added inline comments to D61725: arp: Teach arp to attach and run itself in a jail.
Thu, Nov 19, 6:24 AM
rcm closed D61196: netipsec: Set the outer IPv6 flow label from the SA.
Thu, Nov 19, 3:00 AM
rcm committed rGd1e14cba4a79: netipsec: Set the outer IPv6 flow label from the SA.
Thu, Nov 19, 3:00 AM

Yesterday

rcm closed D61674: pf: Do not report a packet the outbound pass sent as dropped.
Wed, Nov 18, 9:45 PM
rcm committed rG3ccfcdacd6ae: pf: Do not report a packet the outbound pass sent as dropped.
Wed, Nov 18, 9:45 PM
rcm committed rG20e652ba25b2: netinet6: Do not count a packet a hook consumed as discarded in ip6_tryforward().
Wed, Nov 18, 9:44 PM
rcm closed D61042: netinet6: Do not count a packet a hook consumed as discarded in ip6_tryforward().
Wed, Nov 18, 9:44 PM
rcm committed rGbf8783a32faf: netinet6: Decrement the hop limit before the outbound hooks in ip6_tryforward().
Wed, Nov 18, 9:41 PM
rcm closed D61039: netinet6: Decrement the hop limit before the outbound hooks in ip6_tryforward().
Wed, Nov 18, 9:41 PM
rcm committed rGfb9969ca61ca: netinet6: Keep a pfil-chosen next hop in ip6_tryforward().
Wed, Nov 18, 9:38 PM
rcm closed D61038: netinet6: Keep a pfil-chosen next hop in ip6_tryforward().
Wed, Nov 18, 9:38 PM
rcm added a member for pkgbase: rcm.
Wed, Nov 18, 9:26 PM
rcm committed rGb74577a2eaa2: if_epair: Hash the IP header, not the Ethernet header, for fanout.
Wed, Nov 18, 9:20 PM
rcm closed D61037: if_epair: Hash the IP header, not the Ethernet header, for fanout.
Wed, Nov 18, 9:20 PM
rcm added a reviewer for D61726: tests/net/bpf: run the bpf:inject test with RSS again: pouria.
Wed, Nov 18, 9:15 PM
rcm added a reviewer for D61725: arp: Teach arp to attach and run itself in a jail: pouria.
Wed, Nov 18, 9:14 PM
rcm added a comment to D61037: if_epair: Hash the IP header, not the Ethernet header, for fanout.

Please check if this fixes flakyness of tests/sys/net/bpf:inject in presence of RSS and if it does, please remove the check there.

Wed, Nov 18, 9:11 PM
rcm requested review of D61726: tests/net/bpf: run the bpf:inject test with RSS again.
Wed, Nov 18, 9:09 PM
rcm requested review of D61725: arp: Teach arp to attach and run itself in a jail.
Wed, Nov 18, 9:02 PM
rcm updated the diff for D61041: netinet6: Send redirects from ip6_tryforward().

Rebased on D61039 Diff 2: the new min-ttl 255 leg in fragmentation_pass:v6_refrag_hlim chooses the path with vnet_ip6_forward_path, like the rest of the tests here; no other change.

Wed, Nov 18, 7:37 PM
rcm added a comment to D61039: netinet6: Decrement the hop limit before the outbound hooks in ip6_tryforward().
Wed, Nov 18, 7:35 PM
rcm updated the diff for D61039: netinet6: Decrement the hop limit before the outbound hooks in ip6_tryforward().

Give the hop back only below IPV6_MAXHLIM. When an outbound scrub (min-ttl 255) had raised the hop limit to 255, the give-back wrapped it to 0 and ip6_forward() dropped the refragmented packet with time exceeded (thanks, @pouria). Such a packet now leaves with 254, as it did before this change. fragmentation_pass:v6_refrag_hlim gets a min-ttl 255 leg over both paths; summary updated to match.

Wed, Nov 18, 7:33 PM
rcm closed D61516: opencrypto: Do not requeue a re-dispatched request as ordered.
Wed, Nov 18, 7:21 PM
rcm committed rGcb9e8ccf4ff0: opencrypto: Do not requeue a re-dispatched request as ordered.
Wed, Nov 18, 7:21 PM
rcm updated the diff for D61041: netinet6: Send redirects from ip6_tryforward().

Rebased on D61038 Diff 3 (fwd.sh uses sysctl -j there); no other change.

Wed, Nov 18, 6:45 PM
rcm updated the diff for D61676: netinet: Send to the next hop a pfil hook chose in ip_tryforward().

nexthop_v4 now uses ifconfig -j, route -j and sysctl -j, same as nexthop_v6 in D61038.

Wed, Nov 18, 6:42 PM
rcm updated the diff for D61038: netinet6: Keep a pfil-chosen next hop in ip6_tryforward().

The new test now uses ifconfig -j, route -j and sysctl -j instead of jexec; the change to ip6_fastfwd.c is unchanged

Wed, Nov 18, 6:40 PM
rcm added a member for Src Committers: rcm.
Wed, Nov 18, 5:07 PM
rcm added a reviewer for D61715: ixl/iavf: Start the head write-back completion index at the head: kgalazka.
Wed, Nov 18, 3:21 PM
rcm requested review of D61715: ixl/iavf: Start the head write-back completion index at the head.
Wed, Nov 18, 3:20 PM

Tue, Nov 17

rcm added a comment to D61038: netinet6: Keep a pfil-chosen next hop in ip6_tryforward().
In D61038#1412777, @ae wrote:

Hi, recently I also has been debugging the problem with ipfw fwd and IPv6 on inbound pfil processing. We just use fwd rules with out opcode to solve this problem.
Does this behavior match the IPv4 case?

Tue, Nov 17, 3:49 PM
rcm requested review of D61676: netinet: Send to the next hop a pfil hook chose in ip_tryforward().
Tue, Nov 17, 3:43 PM
rcm requested review of D61674: pf: Do not report a packet the outbound pass sent as dropped.
Tue, Nov 17, 2:39 PM
rcm updated the diff for D61041: netinet6: Send redirects from ip6_tryforward().

Rebased onto the two new fixes below it in the stack, and reworked the tests. The change to ip6_fastfwd.c and ip6_input.c is the same.

Tue, Nov 17, 2:20 PM
rcm requested review of D61673: netinet6: Do not forward packets with a loopback address in ip6_tryforward().
Tue, Nov 17, 2:11 PM
rcm requested review of D61672: netinet6: Deliver packets redirected to a local address in ip6_tryforward().
Tue, Nov 17, 2:08 PM
rcm updated the diff for D61042: netinet6: Do not count a packet a hook consumed as discarded in ip6_tryforward().

Rewrote the test; the kernel change is the same.

Tue, Nov 17, 2:04 PM

Mon, Nov 16

rcm updated the diff for D61196: netipsec: Set the outer IPv6 flow label from the SA.

Rebased onto 1b9a8e1dcc69 (D61194 as committed), plus one fix from a final review pass: the bundle relabel tested the final SA's own mode, but an SA added without -m has mode "any" and still serves a transport request, so such a bundle kept the label of the tunnel SA beneath. It now tests the policy request's mode (sp->req[idx - 1]), as the encapsulation decision does.

Mon, Nov 16, 6:48 PM
rcm requested review of D61645: netipsec: Apply the rest of a bundle when its last request has no SA.
Mon, Nov 16, 6:25 PM
rcm retitled D61638: netipsec: Send the packet when the rest of a bundle is skipped from ipsec: Send the packet when the rest of a bundle is skipped
Mon, Nov 16, 5:50 PM
rcm requested review of D61638: netipsec: Send the packet when the rest of a bundle is skipped.
Mon, Nov 16, 5:49 PM
rcm closed D61515: netipsec: Clear the crypto error before dispatching a request again.
Mon, Nov 16, 5:43 PM
rcm committed rGb9e45c2dca5c: netipsec: Clear the crypto error before dispatching a request again.
Mon, Nov 16, 5:43 PM
rcm updated the diff for D61393: ip6_forward: Do not reference the nexthop.

rebased and drop an initializer

Mon, Nov 16, 4:55 PM
rcm updated the diff for D61515: netipsec: Clear the crypto error before dispatching a request again.

The six callbacks now call ipsec_crypto_retry(), which records the new session in the SA, updates xd->cryptoid and clears crp_etype. ipsec_updateid() has no other callers and becomes static. The callbacks still call crypto_dispatch() themselves, because the input callbacks have to release the sahtree lock first.

Mon, Nov 16, 3:57 PM
rcm committed rG698d94219061: netipsec: Release the SA and SP when mb_unmapped_to_ext() fails.
Mon, Nov 16, 3:18 PM
rcm committed rG1b9a8e1dcc69: netipsec: Select outbound transmit queues by SA instead of inner flow.
Mon, Nov 16, 1:20 PM
rcm closed D61194: netipsec: Select outbound transmit queues by SA instead of inner flow.
Mon, Nov 16, 1:20 PM

Sun, Nov 15

rcm added a comment to D61515: netipsec: Clear the crypto error before dispatching a request again.

I don't object to the change, but this strikes me as something that should be either handled in an ipsec subroutine, given that these error handlers all look identical, or in crypto_dispatch() itself.

Sun, Nov 15, 7:59 PM

Fri, Nov 13

rcm requested review of D61521: x86: Do not take a CPUID leaf 1Fh module for the package.
Fri, Nov 13, 8:39 PM
rcm closed D61509: ip6_forward: Free the ICMPv6 copy on all error paths.
Fri, Nov 13, 5:40 PM
rcm committed rG3d88ee6af296: ip6_forward: Free the ICMPv6 copy on all error paths.
Fri, Nov 13, 5:40 PM
rcm requested review of D61516: opencrypto: Do not requeue a re-dispatched request as ordered.
Fri, Nov 13, 5:35 PM
rcm requested review of D61515: netipsec: Clear the crypto error before dispatching a request again.
Fri, Nov 13, 5:27 PM
rcm requested review of D61509: ip6_forward: Free the ICMPv6 copy on all error paths.
Fri, Nov 13, 3:06 PM
rcm requested review of D61493: syslogd: Create one forwarding socket per address pair.
Fri, Nov 13, 12:27 AM

Thu, Nov 12

rcm added a comment to D61476: loader.efi: Configure the boot interface when looking for an initmd.
In D61476#1409297, @imp wrote:

I'll take a look at this. I also have a series that has a fix for this in it. I'll see which is better.

Thu, Nov 12, 11:40 PM
rcm requested review of D61477: libsa: Fall back to DISCOVER promptly when INIT-REBOOT is refused.
Thu, Nov 12, 8:41 PM
rcm requested review of D61476: loader.efi: Configure the boot interface when looking for an initmd.
Thu, Nov 12, 7:54 PM
rcm planned changes to D61385: ip_forward: Take the nexthop reference only for ip_output().

Thanks all, I am working on a new revision!

Thu, Nov 12, 6:58 PM
rcm closed D61384: netipsec: Do not reference the default policy for packets without a PCB.
Thu, Nov 12, 12:58 PM
rcm committed rG53d8733b3dd7: ipsec: Do not reference the default policy for packets without a PCB.
Thu, Nov 12, 12:58 PM

Wed, Nov 11

rcm committed rG5c237bdbf7a4: esp: Allocate the crypto request and its callback state together.
Wed, Nov 11, 8:58 PM
rcm closed D61395: esp: Allocate the crypto request and its callback state together.
Wed, Nov 11, 8:58 PM
rcm added inline comments to D61395: esp: Allocate the crypto request and its callback state together.
Wed, Nov 11, 8:47 PM
rcm updated the diff for D61395: esp: Allocate the crypto request and its callback state together.
  • Size the AAD buffer for the high-order half of the sequence number as well.
  • Allocate struct esp_req from a dedicated UMA zone, created in esp_attach() and destroyed in esp_detach() after xform_detach(). M_ESP is gone.
  • Pass the esp_req as crp_opaque, so esp_freereq() takes it directly and no longer needs __containerof().
Wed, Nov 11, 8:44 PM
rcm committed rG41bdbe7dc06c: carp: Check if_carp under carp_sx in carp_nl_get().
Wed, Nov 11, 7:29 PM
rcm added inline comments to D61385: ip_forward: Take the nexthop reference only for ip_output().
Wed, Nov 11, 5:41 PM
rcm added a comment to D61384: netipsec: Do not reference the default policy for packets without a PCB.
In D61384#1408255, @kib wrote:
In D61384#1408168, @rcm wrote:
In D61384#1408071, @kib wrote:

Could this be done for NONE default policy as well?

It already is: without a PCB the default policy is never referenced, whether it is NONE or DISCARD. The helper name hid that, so the updated diff drops the helper. With a PCB the reference stays for both, because the PCB caches the policy; I would do that case as a follow-up.

I see, the NONE is indicated by sp == NULL and error == 0, am I right?

Wed, Nov 11, 5:08 PM
rcm added a comment to D61384: netipsec: Do not reference the default policy for packets without a PCB.
In D61384#1408071, @kib wrote:

Could this be done for NONE default policy as well?

Wed, Nov 11, 3:17 PM
rcm updated the diff for D61384: netipsec: Do not reference the default policy for packets without a PCB.

Drop the ipsec_default_discard() helper and compare V_def_policy->policy with IPSEC_POLICY_DISCARD in the callers. Say in the comment on ipsec[46]_getpolicy() what a NULL return means.

Wed, Nov 11, 3:15 PM
rcm added a reviewer for D61385: ip_forward: Take the nexthop reference only for ip_output(): zlei.
Wed, Nov 11, 2:06 PM
rcm added inline comments to D60545: pf: Hash rule and pool addresses by type.
Wed, Nov 11, 3:04 AM
rcm updated the diff for D60545: pf: Hash rule and pool addresses by type.

The enum is now named enum pf_addr_type, and pf_hash_addr() switches on (enum pf_addr_type)addr->type with no default case. Leaving a type out is now a -Wswitch error in the kernel build (checked by dropping PF_ADDR_URPFFAILED).

Wed, Nov 11, 3:03 AM
rcm committed rG2af08ab3f125: netipsec: Release the sahtree lock on the AH and IPComp EAGAIN paths.
Wed, Nov 11, 2:32 AM
rcm closed D61399: netipsec: Release the sahtree lock on the AH and IPComp EAGAIN paths.
Wed, Nov 11, 2:32 AM

Tue, Nov 10

rcm added inline comments to D61385: ip_forward: Take the nexthop reference only for ip_output().
Tue, Nov 10, 8:02 PM
rcm updated the diff for D61385: ip_forward: Take the nexthop reference only for ip_output().

Add a comment at the lookup saying why no reference is needed (@zlei's question). No code change. Test plan extended with the nexthop lifetime and failed try-ref runs.

Tue, Nov 10, 7:59 PM
rcm requested review of D61399: netipsec: Release the sahtree lock on the AH and IPComp EAGAIN paths.
Tue, Nov 10, 5:32 PM
rcm requested review of D61395: esp: Allocate the crypto request and its callback state together.
Tue, Nov 10, 3:46 PM
rcm requested review of D61393: ip6_forward: Do not reference the nexthop.
Tue, Nov 10, 2:35 PM
rcm closed D61386: netipsec: Release the SA inside the vnet context in AH and IPComp callbacks.
Tue, Nov 10, 1:42 PM
rcm committed rG5d2e81f29a8a: netipsec: Release the SA inside the vnet context in AH and IPComp callbacks.
Tue, Nov 10, 1:42 PM
rcm committed rG88b7b8512107: ixgbe: Attach the MDIO bus after iflib drops the context lock.
Tue, Nov 10, 12:49 PM
rcm closed D61199: ixgbe: Attach the MDIO bus after iflib drops the context lock.
Tue, Nov 10, 12:49 PM
rcm added a reviewer for D61386: netipsec: Release the SA inside the vnet context in AH and IPComp callbacks: mjg.
Tue, Nov 10, 11:43 AM
rcm requested review of D61386: netipsec: Release the SA inside the vnet context in AH and IPComp callbacks.
Tue, Nov 10, 11:42 AM
rcm requested review of D61385: ip_forward: Take the nexthop reference only for ip_output().
Tue, Nov 10, 11:36 AM
rcm requested review of D61384: netipsec: Do not reference the default policy for packets without a PCB.
Tue, Nov 10, 11:32 AM

Mon, Nov 9

rcm updated the diff for D61275: libfetch: Send If-Range with ranged requests.

Reworked to keep the change minimal, and to answer the ABI point properly:

Mon, Nov 9, 5:21 PM
rcm updated the test plan for D61267: libfetch: Fail truncated or malformed chunked bodies.
Mon, Nov 9, 4:53 PM
rcm updated the diff for D61267: libfetch: Fail truncated or malformed chunked bodies.

fetch_getln() no longer returns a line the peer cut off before
its newline as if it were complete. A chunk size line cut after a
leading "0" used to end the body as complete, and a truncated status,
header or FTP reply line was parsed as whatever was left; it is EPROTO
now. An empty remainder still returns an empty line, so headers
without the final blank line end as before.

Mon, Nov 9, 4:51 PM
rcm updated the diff for D61267: libfetch: Fail truncated or malformed chunked bodies.

Updated: a read error while waiting for a chunk header is no longer
reported as EPROTO. http_new_chunk() sets errno to EPROTO itself for
a malformed size, and http_fillbuf() records errno, so a stall between
chunks fails as "transfer timed out" like a stall inside a chunk does.

Mon, Nov 9, 4:17 PM
rcm committed rG488bf4473772: netipsec: Do not leak the SAH when an SA is cloned.
Mon, Nov 9, 3:32 PM
rcm closed D61101: netipsec: Do not leak the SAH when an SA is cloned.
Mon, Nov 9, 3:32 PM

Sun, Nov 8

rcm updated the diff for D61275: libfetch: Send If-Range with ranged requests.

Enable symbol versioning

Sun, Nov 8, 3:13 AM

Fri, Nov 6

rcm requested review of D61275: libfetch: Send If-Range with ranged requests.
Fri, Nov 6, 8:39 PM
rcm added inline comments to D61267: libfetch: Fail truncated or malformed chunked bodies.
Fri, Nov 6, 6:42 PM
rcm updated the diff for D61267: libfetch: Fail truncated or malformed chunked bodies.
Fri, Nov 6, 6:42 PM