aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* pkg/updater: temporary disable testing of builded syzkalleractualGrigory Bazilevich19 hours1-0/+2
|
* pkg/fuzzer: ProbCover initialization was missing in testGrigory Bazilevich31 hours1-2/+3
|
* pkg/corpus: increase the priority of programs that were rarely selectedGrigory Bazilevich31 hours2-1/+31
|
* syz-manager: information about probability-only inputs in dbGrigory Bazilevich31 hours4-5/+19
|
* pkg/corpus: upgrade corpus minimization algorithmGrigory Bazilevich31 hours3-4/+49
|
* pkg/fuzzer,pkg/corpus: detection and preservation of programs with ↵Grigory Bazilevich31 hours4-40/+78
| | | | probability coverage
* pkg/signal: new utility methodsGrigory Bazilevich31 hours1-0/+39
|
* pkg/corpus: update Programs List priority storageGrigory Bazilevich31 hours5-17/+171
| | | | | | | | | | | Static prefix sums have been replaced with a Fenwick tree. In the current syzkaller, program priority was set based on a Signal received by a single system call. This commit allows priority to be changed dynamically, making it possible to maintain priority based on Signals from all system calls. Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru>
* pkg/csource: build tests without warningsGrigory Bazilevich31 hours3-5/+7
| | | | | | Otherwise syz-ci will fail to build the project. Signed-off-by: Denis Efremov <efremov@ispras.ru>
* dashboard/app: use local scripts instead of downloading from cdnGrigory Bazilevich31 hours15-6/+2597
| | | | Signed-off-by: Sergey Zotov <svzotov@ispras.ru>
* dashboard/app: add direct link to a manager's coverageGrigory Bazilevich31 hours1-1/+1
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* syz-manager: save kernel commit tag in the crash metainfoGrigory Bazilevich31 hours2-0/+30
| | | | | | | Add hash of actual kernel commit to crash metainfo if available. Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru> Signed-off-by: Sergey Zotov <svzotov@ispras.ru>
* pkg/report: ignore timer-related and mark_buffer_dirty framesGrigory Bazilevich31 hours6-762/+328
| | | | | | | Fix corresponding test, remove incorrect tests Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru> Signed-off-by: Denis Efremov <efremov@ispras.ru>
* syz-ci: add syzkaller test timeout retrieval from configGrigory Bazilevich31 hours2-1/+27
| | | | | | | The timeout for "go test ..." can now be specified in the syz-ci settings in the `SyzkallerBuildParams` dictionary under `test_timeout`. Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru>
* dashboard/app: add kernel-mirror.intra.ispra.ru to cgit hostsDenis Efremov31 hours1-1/+1
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* dashboard: repeat bug reproduction every 30 daysAlexey Khoroshilov31 hours1-2/+2
|
* pkg/vcs: show proper link to LVC syzkallerDenis Efremov31 hours1-1/+12
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* tools/syz-hubtool: add chunked and consistent corpus upload logicGrigory Bazilevich31 hours1-3/+58
| | | | Signed-off-by: Alexey Panov <apanov@astralinux.ru>
* syz-manager: don't use hostname for public web addressGrigory Bazilevich31 hours1-3/+0
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* pkg/db: skip invalid corpus entries instead of failingGrigory Bazilevich31 hours1-1/+1
| | | | Signed-off-by: Vladimir Telezhnikov <vtelezhnikov@astralinux.ru>
* pkg/vcs/linux: use gcc-10.3.0 for kernels +v5.10Grigory Bazilevich31 hours1-0/+2
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* pkg/mgrconfig: Make slowdown configurableGrigory Bazilevich31 hours2-0/+10
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* syz-ci: make rebuild periods configurableGrigory Bazilevich31 hours3-55/+66
| | | | Signed-off-by: Denis Efremov <efremov@ispras.ru>
* tools/create-image.sh: add arguments for apt-repo and keyringGrigory Bazilevich31 hours2-107/+121
| | | | Signed-off-by: Sergey Zotov <svzotov@ispras.ru>
* tools: add kcidb and hubtool targets to MakefileGrigory Bazilevich31 hours1-1/+5
| | | | | Signed-off-by: Denis Efremov <efremov@ispras.ru> Signed-off-by: SVZotov <svzotov@ispras.ru>
* executor: fix fd leaksGrigory Bazilevich31 hours1-1/+7
| | | | | | | | | Close file descriptors in read_btf_vmlinux, find_vf_interface and syz_usbip_server_init. Otherwise sykaller fails when executor exeeds a limit for available fds. Signed-off-by: Pavel Nikulshin <p.nikulshin@ispras.ru>
* syz-ci/managercmd: fix graceful syz-manager terminationGrigory Bazilevich31 hours1-2/+4
| | | | | | | Due to a combination of circumstances, syz-ci may have failed to send a termination signal to syz-manager, which can cause a system overload. Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru>
* pkg/build/linux: add build time loggingGrigory Bazilevich31 hours1-0/+3
| | | | Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru>
* pkg/build/linux: add timeout retrieval from configGrigory Bazilevich31 hours1-1/+23
| | | | | | | The build timeout can now be specified in the manager settings in the ‘build’ dictionary under ‘timeout’. Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru>
* syz-manager: always save crashes and repros locallyGrigory Bazilevich31 hours3-8/+65
| | | | | | | | | | | Save crashes and repros with corresponding kernel and syzkaller configs locally in syz-manager even when they are published to syz-dashboard, to be able to share the crashes with Linux Verification Center (linuxtesting.org). Signed-off-by: Grigory Bazilevich <g.bazilevich@ispras.ru> Signed-off-by: Sergey Zotov <svzotov@ispras.ru> Co-authored-by: Sergey Zotov <svzotov@ispras.ru>
* dashboard/app: use '/upstream' in local_ui_test.goAlexander Potapenko2 days1-3/+3
| | | | | The default dashboard namespace is called '/upstream', so let's make it the same in the test.
* syz-agent: systematically show verbose error messagesDmitry Vyukov2 days7-7/+30
| | | | | | | | | | | | | Currently we added custom code to kernel build action, and few others to expose verbose errors from executed binaries (notably make). But lots of other binary executions missing this logic, e.g. for git failure we currently see unuseful: failed to run ["git" "fetch" "--force" "--tags" exit status 128 Instead of adding more and more custom code to do the same, remove the custom code and always add verbose output in syz-agent and tools/syz-aflow.
* pkg/gce, vm/gce: create InstanceConfig struct used by CreateInstance methodPimyn Girgis2 days2-16/+33
|
* sys/linux: run make extractPimyn Girgis2 days3-17/+13
|
* sys/linux: update descriptionsPimyn Girgis2 days2-3/+3
| | | | | dev_kvm_amd64: TDX_ATTR_* were renamed to TDX_TD_ATTR_*. landlock: LANDLOCK_ACCESS_NET_*_UDP and LANDLOCK_SCOPE_PATHNAME_UNIX_SOCKET are no longer present.
* dashboard/app: repair AI job access checkDmitry Vyukov2 days1-3/+13
| | | | | | | | | Currently the query fails with: API error 1 (datastore_v3: BAD_REQUEST): cannot get more than 1000 keys in a single call back Don't try to select more than 1000 bugs.
* pkg/aflow/tool/grepper: disable tests on non-linuxDmitry Vyukov2 days1-0/+3
| | | | | | | | | The tests fail on OpenBSD with: expected: "bad expression: fatal: command line, 'bad expression (': Unmatched ( or \\(" actual : "bad expression: fatal: command line, 'bad expression (': parentheses not balanced" Disable the tests on non-linux for now.
* pkg/aflow: return syzkaller program as outputAlexander Potapenko3 days1-3/+1
| | | | | Requesting to return the program as one of the agent's outputs enforces its structure and prevents LLM from using garbage formatting.
* pkg/aflow: ensure we don't register MCP tools with duplicate namesDmitry Vyukov3 days5-4/+40
| | | | | | | If we have duplicate names, then only one of the duplicates will be used at random. Add a check that we don't have duplicate names. Currently it's only "crash-reproducer" (both action and a tool). Also ignore "set-results" tool, and all tools created in tests.
* prog: add support for decimal/octal arguments in text formatAlexander Potapenko3 days3-5/+15
| | | | | | | | | | | Previously, the parser only expected the '0' character to begin an integer argument, which effectively meant it only handled hexadecimal formatting (e.g., `0x...`) when parsing arguments. This change modifies parseArgImpl() to route any starting digit ('0'-'9') to the integer parsing logic. Since `strconv.ParseUint` already handles base-10 parsing using the "0" base flag, this cleanly enables the parser to natively deserialize decimal and octal arguments.
* pkg/aflow: handle more genai errorsDmitry Vyukov3 days2-6/+41
| | | | Fixes #6897
* dashboard/app: filter by workflows in NextStaleJobAleksandr Nogikh3 days2-10/+28
| | | | | | | We were ignoring the workflows passed by the agent in the job poll request in the context of restarting failed jobs. Fix it and add test coverage.
* sys/linux: add descriptions for NILFS2 subsystemJiaming Zhang3 days2-0/+168
|
* sys/linux: update descriptions of BTRFSJiaming Zhang3 days3-0/+42
|
* sys/linux: add descriptions for OCFS2 subsystemJiaming Zhang3 days2-0/+202
|
* dashboard/app: redefine the workflows tableAleksandr Nogikh3 days6-45/+59
| | | | | | Make Workflows be interleaved into Agents. This enables a more complete data representation and enables a more precise tracking of the active workflow types.
* dashboard: recreate stall AI jobsAleksandr Nogikh3 days3-38/+245
| | | | | | | | | | | If an agent has taken a job and then, without finishing it, comes for the next one, recreate that job. If an agent has started working on a job, not finished it and completely disappeared for more than 8 hours, give the job to another agent. At the same time, don't look at the stale jobs in the automatic job creation logic - let it focus on recreating explicitly failed jobs.
* dashboard: remember last active time of each agentAleksandr Nogikh3 days9-5/+116
| | | | Update the timestamp on each job poll or each trajectory update.
* dashboard/app: remember syz-agent namesAleksandr Nogikh3 days10-0/+18
| | | | | | | For now, use a dashboard client name. In a prod deployment, that would be a persistent name of the machine where the syz-agent is hosted. This information be used later for restarting the aborted jobs.
* pkg/aflow: add Flow.Consts instead of ProvideDmitry Vyukov4 days12-64/+133
| | | | | | | | | There is no point in using Provide more than once, and anywhere besides the first action of a flow. So it's not really an action, but more of a flow property. Add Flow.Consts field to handle this case better. Also provide slightly less verbose syntax by using a map instead of a struct, and add tests.
generated by ggit (джигит) mrf-deployment (git 2.46.0) at 2026-03-13 15:52:59 +0000