Add missing NoSessionPIDFD error to the interface definition; the error
is emitted by vl_method_create_session() when the session leader process
does not have a pidfd available, but was never declared in the IDL.
Drop SD_JSON_MANDATORY from the ReleaseSession.Id dispatch entry so that
omitting the field (which the IDL allows via SD_VARLINK_NULLABLE)
resolves to the caller's own session via session_is_self(NULL).
The error is emitted by vl_method_create_session() when the session leader
process does not have a pidfd available, but was never declared in the IDL.
Fixes: 3180c4d461
The ReleaseSession method's Id field is declared as nullable (?string) in
the IDL, allowing callers to omit it so that the method releases the
caller's own session via session_is_self(NULL). The SD_JSON_MANDATORY flag
in the dispatch table contradicts this and makes omitting Id return -EINVAL
("Invalid argument", parameter "Id") instead.
Drop the flag so omitting Id is treated as passing NULL.
Fixes: 2baca6c22b
If a pidfd id is passed to us, it must have originated from
somewhere, i.e. pidfs should be supported on the system.
Hence let's be honest about -EOPNOTSUPP, which might indicate
some bigger problems under the hood, rather than allowing
it to be turned into mild -ESRCH.
Fix typo: DNS_TYPE_DNSKEY should be used instead of DNS_TYPE_KEY.
- DNS_TYPE_KEY is deprecated
- Source code at resolved-dns-trust-anchor.c#L313 handles
DNS_TYPE_DNSKEY records, but not DNS_TYPE_KEY
So DNS_TYPE_KEY is not referenced anywhere
The error is emitted by `vl_method_resolve_service_complete()` when an
SRV record with a root domain hostname is found, indicating the service
is explicitly not offered on the queried domain (RFC 2782 Usage Rules),
but was not declared in the `io.systemd.Resolve` interface definition.
Follow-up for 3d8f2c1464.
It was noticed that cryptsetup luksDump outputs
tpm2-primary-alg: ecc
regardless of the actual primary key type. Namely, in the situation when
RSA SRK is passed to 'systemd-cryptenroll', the output is incorrect and
misleading. Turns out 'tpm2-primary-alg' is not currently used for
unsealing as 'tpm2_srk' object has it in its properties but the
misleading information problem stays.
Fix the issue with a two-fold fix:
- Put the SRK primary alg type to the JSON token so 'cryptsetup
luksDump' has a chance to actually print the right information without
the need to inspect SRK.
- Avoid printing 'tpm2-primary-alg' when it is unset in the JSON as the
default 'ECC' may not match the reality.
No real functional change for the sealing/unsealing expected.
I found the existing explanation of RestartSteps to be simply
impenetrable. Even providing the full docs context to several of our new
AI overlords resulted in wildly different (and completely incorrect)
explanations of the final restart intervals.
Digging through the code, I found the restart delay computation in
`service_restart_usec_next` in `src/core/service.c`.
I've updated the documentation for RestartSteps with an example,
suggested value range and a detailed enough explanation that accurately
describes the current behavior.
The kfd device is used for running compute workloads on AMD
GPUs. Users that are logged in should be able to run compute
so tag them like other DRM and ACCEL devices are.
History has clearly shown that we're terrible at keeping
the size of the dont_close array up-to-date. Hence let's
step away from a hardcoded max size for that, instead
always collect all fds in the array initializer and
let compiler figure it out, taking advantage of the fact
that close_all_fds() gracefully handles invalid fds in
the 'except' array.
Follow-up for 12001b1bf0
It is part of the keep_fds array already, which is intended
for fds that need to survive until the very end of executor
runtime, i.e. across the second close_all_fds() call.
The owned fds should be closed before close_all_fds() along the lines
of exec_params/runtime_close(), otherwise they would be left dangling
afterwards.
When 'tpm2-primary-alg' is missing in LUKS JSON token, the output of
'cryptsetup luksDump' is always:
tpm2-primary-alg: ecc
because tpm2_parse_luks2_json() returns the default (TPM2_ALG_ECC). This can be
misleading and wrong. Make tpm2_parse_luks2_json() return the reality and move
the default to tpm2_unseal().
'tpm2-primary-alg' field is currently unset in LUKS JSON token both for the
case when SRK was used for enrolling (--tpm2-device-key=) and for the case when
SRK was obtained/generated (--tpm2-device=). While this information is not
really needed for unsealing (the sealed object itself has key type in it), it
may be convenient to see key type in e.g. 'cryptsetup luksDump' in the
situations where key type matters. Since 'tpm2-primary-alg' is already defined,
just set it properly in all cases.
Add the missing 'mstack' value to the ImageType enum in the
io.systemd.sysext Varlink IDL — image_type_table in discover-image.c
already defines IMAGE_MSTACK = "mstack" but it was absent from the IDL.
Export vl_type_ImageType so the test can reference it, and add
TEST_IDL_ENUM(ImageType, ...) to the enums_idl test to catch
similar regressions in the future.
The handler sends the remote URL as a parameter via sd_varlink_errorbo()
but the IDL declared the error with no fields, making the payload
undiscoverable via introspection.
Yu Watanabe
Lennart Poettering
Zbigniew Jędrzejewski-Szmek
noxiouz
Mike Yuan
Thomas Weißschuh
Ronan Pigott
Val Markovic
Mario Limonciello
IntenseWiggling
dependabot[bot]
Weixie Cui
Luca Boccassi
lumingzh
김인수
Léane GRASSER
Efstathios Iosifidis
Vitaly Kuznetsov
Pontus Lundkvist
