definitions for linux hardware timestamping #3155

folkertdev · 2023-03-19T20:31:31Z

Definitions can be found here https://elixir.bootlin.com/linux/latest/source/include/uapi/linux/net_tstamp.h#L76

These definitions are relevant for (PTP and NTP) hardware timestamping

rustbot · 2023-03-19T20:31:35Z

(rustbot has picked a reviewer for you, use r? to override)

folkertdev · 2023-03-19T20:57:33Z

I'm not sure what this CI error means

  cargo:warning=/checkout/target/i686-unknown-linux-gnu/debug/build/libc-test-1d150dc772ee3115/out/main.c:35676:80: error: 'HWTSTAMP_FLAG_BONDED_PHC_INDEX' undeclared here (not in a function)
  cargo:warning=35676 |             static const int __test_const_HWTSTAMP_FLAG_BONDED_PHC_INDEX_val = HWTSTAMP_FLAG_BONDED_PHC_INDEX;
  cargo:warning=      |                                                                                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  cargo:warning=/checkout/target/i686-unknown-linux-gnu/debug/build/libc-test-1d150dc772ee3115/out/main.c:35682:68: error: 'HWTSTAMP_FLAG_LAST' undeclared here (not in a function)
  cargo:warning=35682 |             static const int __test_const_HWTSTAMP_FLAG_LAST_val = HWTSTAMP_FLAG_LAST;
  cargo:warning=      |                                                                    ^~~~~~~~~~~~~~~~~~
  cargo:warning=/checkout/target/i686-unknown-linux-gnu/debug/build/libc-test-1d150dc772ee3115/out/main.c:35688:68: error: 'HWTSTAMP_FLAG_MASK' undeclared here (not in a function)
  cargo:warning=35688 |             static const int __test_const_HWTSTAMP_FLAG_MASK_val = HWTSTAMP_FLAG_MASK;
  cargo:warning=      |                                                                    ^~~~~~~~~~~~~~~~~~
  cargo:warning=cc1: note: unrecognized command-line option '-Wno-unknown-warning-option' may have been intended to silence earlier diagnostics

maybe the system libc on CI does not (yet?) have these symbols? They are not important to me so I'd be OK with removing them (just thought I might as well add them for completeness)

JohnTitor · 2023-03-23T07:43:01Z

maybe the system libc on CI does not (yet?) have these symbols?

Correct.

Could you squash commits into one?

JohnTitor · 2023-03-23T11:03:58Z

Thank you, @bors r+

bors · 2023-03-23T11:03:59Z

📌 Commit b409bdc has been approved by JohnTitor

It is now in the queue for this repository.

bors · 2023-03-23T11:04:06Z

⌛ Testing commit b409bdc with merge 38ff53a...

definitions for linux hardware timestamping Definitions can be found here https://elixir.bootlin.com/linux/latest/source/include/uapi/linux/net_tstamp.h#L76 These definitions are relevant for (PTP and NTP) hardware timestamping

bors · 2023-03-23T11:15:40Z

💔 Test failed - checks-actions

folkertdev · 2023-03-24T10:03:08Z

I'm not sure what went wrong here. I assume the action needs to just be re-run?

folkertdev · 2023-03-28T08:22:28Z

I added two extra constants that are needed (SIOCSHWTSTAMP and SIOCGHWTSTAMP) and rebased. Maybe that makes bors happier?

though CI seems to be broken at the moment.

JohnTitor · 2023-03-28T09:00:39Z

The FreeBSD failure is unrelated and it's already fixed on master.
But #3155 (comment) is legit. That's because (probably) musl toolchain on CI is old, could you skip tests on that env?

folkertdev · 2023-03-28T09:57:56Z

ok, rebased again to fix that freebsd problem

could you skip tests on that env

how would I do this? do you mean a cfg for the undefined constants?

JohnTitor · 2023-04-01T04:19:25Z

https://github.com/rust-lang/libc/blob/master/libc-test/build.rs is a place to tweak, like:

libc/libc-test/build.rs

Lines 1739 to 1754 in 68e06ad

    
           cfg.skip_const(move |name| { 
        
               match name { 
        
                   // The IPV6 constants are tested in the `linux_ipv6.rs` tests: 
        
                   | "IPV6_FLOWINFO" 
        
                   | "IPV6_FLOWLABEL_MGR" 
        
                   | "IPV6_FLOWINFO_SEND" 
        
                   | "IPV6_FLOWINFO_FLOWLABEL" 
        
                   | "IPV6_FLOWINFO_PRIORITY" 
        
                   // The F_ fnctl constants are tested in the `linux_fnctl.rs` tests: 
        
                   | "F_CANCELLK" 
        
                   | "F_ADD_SEALS" 
        
                   | "F_GET_SEALS" 
        
                   | "F_SEAL_SEAL" 
        
                   | "F_SEAL_SHRINK" 
        
                   | "F_SEAL_GROW" 
        
                   | "F_SEAL_WRITE" => true,

folkertdev · 2023-04-03T07:07:23Z

allright, I added logic to ignore the constant for i686 and mips. that should make the tests that failed before pass, but there might be other (cancelled) ci runs that also should skip this constant. Is there a better way to figure that out than trial and error?

JohnTitor · 2023-04-03T07:17:16Z

There's a trade-off, we could disable the fail-fast but we may want to find failures fast.
@bors r+

JohnTitor · 2023-04-03T09:48:14Z

libc-test/build.rs

@@ -3724,6 +3724,8 @@ fn test_linux(target: &str) {
                => true,
            "SCTP_FUTURE_ASSOC" | "SCTP_CURRENT_ASSOC" | "SCTP_ALL_ASSOC" | "SCTP_PEER_ADDR_THLDS_V2" => true, // linux 5.5+

+            "HWTSTAMP_TX_ONESTEP_P2P" if i686 || mips || aarch64_musl => true,


I think the below should work like others:

Suggested change

"HWTSTAMP_TX_ONESTEP_P2P" if i686 || mips || aarch64_musl => true,

// FIXME: Requires more recent kernel headers

"HWTSTAMP_TX_ONESTEP_P2P" if mips || musl => true,

yes, fixed. A couple more linux flavors to get through

The condition doesn't seem to be changed. It should mips || musl as we use older kernels for these targets.

the tests bors also fail for aarch with musl here https://github.com/rust-lang/libc/actions/runs/4593954920/jobs/8112512429

based on https://elixir.bootlin.com/linux/v5.6/C/ident/HWTSTAMP_TX_ONESTEP_P2P, linux version 5.6 is the first one that defines this symbol.

though maybe musl is more relevant than aarch here?

so i686 || mips || musl, or only musl even?

Ah, I mean the condition should be sparc64 || musl, not mips.

For sparc64, see:

libc/ci/docker/sparc64-unknown-linux-gnu/Dockerfile

Lines 1 to 2 in c64bfc5

# FIXME: Update to 22.04 once Debian image of sparc64 has a newer glibc.

FROM ubuntu:20.04

i686 itself isn't related to the failure, it fails due to the musl toolchain.

makes sense, I've updated it

JohnTitor · 2023-04-03T11:36:17Z

Let's try again, @bors r+

bors · 2023-04-03T11:36:18Z

📌 Commit 58ffdf4 has been approved by JohnTitor

It is now in the queue for this repository.

bors · 2023-04-03T11:36:25Z

⌛ Testing commit 58ffdf4 with merge c8de3b4...

definitions for linux hardware timestamping Definitions can be found here https://elixir.bootlin.com/linux/latest/source/include/uapi/linux/net_tstamp.h#L76 These definitions are relevant for (PTP and NTP) hardware timestamping

bors · 2023-04-03T11:57:54Z

💔 Test failed - checks-actions

folkertdev · 2023-04-03T12:08:53Z

we're getting closer. I'm not sure what this is trying to do though:

RUNNING ALL TESTS
rust[8] = 0 != 8 (C): Rust "hwtstamp_config" -> C
rust[9] = 0 != 9 (C): Rust "hwtstamp_config" -> C
rust[10] = 0 != 10 (C): Rust "hwtstamp_config" -> C
rust[11] = 0 != 11 (C): Rust "hwtstamp_config" -> C
thread 'main' panicked at 'some tests failed', /checkout/target/sparc64-unknown-linux-gnu/debug/build/libc-test-298fd54f1fef40cd/out/main.rs:12:21
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
error: test failed, to rerun pass `--test main`

Caused by:
  process didn't exit successfully: `/test-runner-linux sparc64 /checkout/target/sparc64-unknown-linux-gnu/debug/deps/main-973f2bcc439c5f8d` (exit status: 1)

should we skip the hwtstamp_config struct on this target?

folkertdev · 2023-04-03T12:44:18Z

rather looks like this is an ABI roundtrip test. That this fails is extremely surprising to me, given that the struct is just three ints, so nothing tricky is going on. Should we just ignore the roundtrip for sparc64?

JohnTitor · 2023-04-03T20:44:46Z

I guess #1558 is related. Yeah, skipping would be fine 👍

folkertdev · 2023-04-03T21:08:25Z

allright let's hope that's it

JohnTitor · 2023-04-03T21:22:41Z

@bors r+

bors · 2023-04-03T21:22:43Z

📌 Commit 4949998 has been approved by JohnTitor

It is now in the queue for this repository.

bors · 2023-04-03T21:49:58Z

⌛ Testing commit 4949998 with merge 2dc31a7...

bors · 2023-04-03T22:48:28Z

☀️ Test successful - checks-actions, checks-cirrus-freebsd-12, checks-cirrus-freebsd-13, checks-cirrus-freebsd-14
Approved by: JohnTitor
Pushing 2dc31a7 to master...

rustbot assigned JohnTitor Mar 19, 2023

rustbot added the S-waiting-on-review label Mar 19, 2023

folkertdev force-pushed the hardware-timestamping branch 2 times, most recently from 687f7bc to b409bdc Compare March 23, 2023 10:16

JohnTitor approved these changes Mar 23, 2023

View reviewed changes

bors added S-waiting-on-bors and removed S-waiting-on-review labels Mar 23, 2023

bors added S-waiting-on-review and removed S-waiting-on-bors labels Mar 23, 2023

folkertdev force-pushed the hardware-timestamping branch from b409bdc to 26071ae Compare March 28, 2023 08:20

folkertdev force-pushed the hardware-timestamping branch from 26071ae to 2f4ef84 Compare March 28, 2023 09:54

JohnTitor added S-waiting-on-author and removed S-waiting-on-review labels Apr 1, 2023

folkertdev force-pushed the hardware-timestamping branch from 2f4ef84 to c9dbeda Compare April 3, 2023 07:05

folkertdev force-pushed the hardware-timestamping branch from c9dbeda to 5842b24 Compare April 3, 2023 07:08

bors added S-waiting-on-review and removed S-waiting-on-bors labels Apr 3, 2023

folkertdev force-pushed the hardware-timestamping branch from 5842b24 to bdd5b57 Compare April 3, 2023 08:22

JohnTitor reviewed Apr 3, 2023

View reviewed changes

folkertdev force-pushed the hardware-timestamping branch 2 times, most recently from 89b13ab to 58ffdf4 Compare April 3, 2023 11:17

bors added S-waiting-on-bors and removed S-waiting-on-review labels Apr 3, 2023

bors added S-waiting-on-review and removed S-waiting-on-bors labels Apr 3, 2023

definitions for linux hardware timestamping

4949998

folkertdev force-pushed the hardware-timestamping branch from 58ffdf4 to 4949998 Compare April 3, 2023 21:07

bors added S-waiting-on-bors and removed S-waiting-on-review labels Apr 3, 2023

bors merged commit 2dc31a7 into rust-lang:master Apr 3, 2023
10 checks passed

This was referenced Apr 3, 2023

linux SCTP support part 2 #3117

Closed

Additional linux AF_PACKET functionality #3120

Closed

	"HWTSTAMP_TX_ONESTEP_P2P" if i686 \|\| mips \|\| aarch64_musl => true,
	// FIXME: Requires more recent kernel headers
	"HWTSTAMP_TX_ONESTEP_P2P" if mips \|\| musl => true,

	# FIXME: Update to 22.04 once Debian image of sparc64 has a newer glibc.
	FROM ubuntu:20.04

definitions for linux hardware timestamping #3155

definitions for linux hardware timestamping #3155

Conversation

folkertdev commented Mar 19, 2023

rustbot commented Mar 19, 2023

folkertdev commented Mar 19, 2023

JohnTitor commented Mar 23, 2023

JohnTitor commented Mar 23, 2023

bors commented Mar 23, 2023

bors commented Mar 23, 2023

bors commented Mar 23, 2023

folkertdev commented Mar 24, 2023

folkertdev commented Mar 28, 2023

JohnTitor commented Mar 28, 2023

folkertdev commented Mar 28, 2023

JohnTitor commented Apr 1, 2023

folkertdev commented Apr 3, 2023

JohnTitor commented Apr 3, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JohnTitor Apr 3, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JohnTitor commented Apr 3, 2023

bors commented Apr 3, 2023

bors commented Apr 3, 2023

bors commented Apr 3, 2023

folkertdev commented Apr 3, 2023

folkertdev commented Apr 3, 2023

JohnTitor commented Apr 3, 2023

folkertdev commented Apr 3, 2023

JohnTitor commented Apr 3, 2023

bors commented Apr 3, 2023

bors commented Apr 3, 2023

bors commented Apr 3, 2023

JohnTitor Apr 3, 2023 •

edited