Let's stop using --json-schema and instead have claude write a JSON
file in the repo root which we pass around as an artifact similar to
how we pass around the input. This works around the bug where claude
receives task notifications after producing structured output which
breaks the structured output.
We often have a pattern where the same verb function is used for
multiple actions. This leads to an antipattern where we figure out what
action needs to be taken from argv[0] multiple times: often once in
arse_argv() to figure out what options are allowed, then once again
implicitly in dispatch_verb(), and then again in the action verb itself.
Let's allow passing a parameter into the verb to simplify this.
As it seems impossible to prevent claude from receiving notifications
about subagents finishing after it has produced structured output, which
breaks the structured output as it has to be the final reply, let's stop
using subagents and background tasks completely to avoid the issue.
tpm_log_tagged_event() only measures the event to the TPM while
tpm_log_ipl_event() measures the event both to the TPM and CC. Fix the
inconsistency.
Note, this is a potentially breaking change for TDX guests as systemd will
now measure more stuff to the MRTD/RTMRs, reference values for attestation may
need to be adjusted.
Found by Claude Code Review.
If this new scheme of adding dependencies is supposed to be used more
widely we need to start making it easy to add them. So add a new
self-contained header that projects can simply include without the need
to link against libsystemd itself. This will allow them to generate
`.notes.dlopen` sections:
```
> readelf -p .note.dlopen ./l2md
String dump of section '.note.dlopen':
[ a] |@FDO
[ 10] [{"feature":"manifest-json","description":"Manifest-based change detection via gzip and JSON parsing","priority":"suggested","soname":["libz.so.1","libsystemd.so.0"]}]
[ c2] |@FDO
[ c8] [{"feature":"manifest-http","description":"HTTP transport for lore.kernel.org manifest fetch","priority":"suggested","soname":["libcurl.so.4"]}]
```
The SYSTEMD_COLORS=true/1/yes no longer forced colors
when stdout was not a TTY (e.g. piped), because the COLOR_TRUE bypass
of the terminal_is_dumb() check was accidentally dropped.
Restore the old behavior by guarding the TTY check with
`m != COLOR_TRUE`, so an explicit boolean "true" value continues to
unconditionally force color output regardless of whether stdout is a TTY
or whether $NO_COLOR is set.
claude keeps failing by its subagents completing after it has already
written the review for large prs. It seems to run out of turns, tries
to get the subagents to post partial reviews but doesn't seem to stop
them.
Let's insist that it waits for background tasks to stop but let's also
increase the max turns a bit so it doesn't run out as quickly.
claude wants to use python to access the JSON context so let's allow
it. Since python3 basically allows you to reimplement every other tool,
let's just enable all tools except the web related ones but enable network
isolation so it can't try to exfiltrate anything via python.
Add a --grain-size= CLI option to override the default 4 KiB partition
alignment grain. Setting --grain-size=1M matches the alignment used by
fdisk/parted and fixes misaligned partitions after small fixed-size
partitions like the 16 KiB verity-sig partition.
Also fix context_place_partitions() to re-align the start offset after
each partition, not just once per free area. Without this, a small
partition would cause all subsequent partitions in the same free area
to start at an unaligned offset.
In CI we get spurious failures about unitialized variables with gcc
versions older then (depending on the case) 12, 13, or 14. Let's only
try to do this check with newer gcc which returns more useful results.
At the same time, do compile with both gcc and clang at -O2, just
disable the warning.
The old logic seems to have been confused. We compile with -Wall, at
least in some cases, which includes -Wmaybe-unitialized. So if we
_don't_ want it, we need to explicitly disable it.
We often have a pattern where the same verb function is used for
multiple actions. This leads to an antipattern where we figure out what
action needs to be taken from argv[0] multiple times: often once in
parse_argv() to figure out what options are allowed, then once again
implicitly in dispatch_verb(), and then again in the action verb itself.
Let's allow passing a parameter into the verb to simplify this.
This matches a pattern we have in conf-parser.h, where we have both
void *userdata (more global) and void *data (per-config item). Here,
I opted for uintptr_t userdata. It seems that most of the time we'll
want to just pass an enum value. This works OK with no casts. I also
tried a void* and union. In both cases, much more boilerplate is needed:
either a cast or a macro to help avoid compiler warnings. uintptr_t
seems generic enough to cover foreseeable usecases with no fuss.
This is a noop refactoring. See next commit for an example.
This series of renaming patches has a few overlapping motivations:
- when functions are named uniformly, it code is more obvious
- I want to add a parameter to all verb functions
- in #40880 uniform naming of verb functions will be necessary too.
So let's do this cleanup. Some tools had a mix of functions w/ and
w/o "verb_", which looked messy.
Relicense sd-dlopen.h from LGPL-2.1-or-later to MIT-0 so that
downstream projects can copy/paste the macros directly without
introducing a build dependency on the systemd headers.
Acked-by: Lennart Poettering <lennart@amutable.com>
Acked-by: Luca Boccassi <luca.boccassi@gmail.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Document the new public sd-dlopen.h header and SD_ELF_NOTE_DLOPEN()
macro with associated constants. Includes usage examples for single
and multiple soname annotations.
Signed-off-by: Christian Brauner <brauner@kernel.org>
Switch all internal callers from the private ELF_NOTE_DLOPEN() macro to
the new public SD_ELF_NOTE_DLOPEN() API from sd-dlopen.h, and remove
the now-redundant macro definitions from dlfcn-util.h.
Signed-off-by: Christian Brauner <brauner@kernel.org>
Expose ELF note dlopen annotation macros as a public header-only API in
sd-dlopen.h. This allows any project to embed .note.dlopen metadata in
their ELF binaries by simply including the header - no runtime linkage
against libsystemd is required.
The header provides SD_ELF_NOTE_DLOPEN() and associated macros/constants
implementing the ELF dlopen metadata specification for declaring optional
shared library dependencies loaded via dlopen() at runtime.
Signed-off-by: Christian Brauner <brauner@kernel.org>
Add %I (TID in initial PID namespace) to the core_pattern, so the
kernel passes the crashing thread's TID to systemd-coredump. Use it
to read the thread's comm name from /proc/<tid>/comm and log both as
new journal fields:
COREDUMP_TID= — TID of the crashing thread
COREDUMP_THREAD_NAME= — comm name of the crashing thread
These fields are also stored as xattrs on external coredump files
(user.coredump.tid, user.coredump.thread_name) and displayed by
coredumpctl info alongside the PID line.
For single-threaded processes the TID equals the PID and thread_name
equals comm; for multi-threaded programs with named worker threads
(pthread_setname_np / PR_SET_NAME) this identifies which thread
crashed without needing to open the coredump file itself.
The new fields are optional in the socket forwarding path, so older
systemd-coredump senders are handled gracefully.
Co-developed-by: Claude <claude@anthropic.com>
These functions take so many return paramaters, and in many of our cases
we don't actually needt them. Hence introduce _full() flavours of the
funcs, and hide the params by default.
The labelling approach introduced in 6089075265
means contributors can now trigger the workflow on their own when the label
is added by a maintainer and they update the PR. Hence we need to allow all
users to access the claude code action. This is safe because we already gate
the workflow ourselves to only the contributors that we want to allow.
Additionally, the claude code job has no permissions anymore except read access
to the repository and can execute very limited tools, so this should be safe.
The name doesn't actually matter, it gets replaced with the name
of the file when not archiving. So stop passing a name and pass in
the filename as the name when downloading the artifact.
