Two virtual wireless devices are instantiated during network devices
initialization.
A new flag (-wifi) is added that controls whether these virtual wifi
devices are instantiated and configured during proc initialization.
Also, two new pseudo syscalls are added:
1. syz_80211_inject_frame(mac_addr, packet, packet_len) -- injects an
arbitrary packet into the wireless stack. It is injected as if it
originated from the device identitied by mac_addr.
2. syz_80211_join_ibss(interface_name, ssid, ssid_len, mode) --
puts a specific network interface into IBSS state and joins an IBSS
network.
Arguments of syz_80211_join_ibss:
1) interface_name -- null-terminated string that identifies
a wireless interface
2) ssid, ssid_len -- SSID of an IBSS network to join to
3) mode -- mode of syz_80211_join_ibss operation (see below)
Modes of operation:
JOIN_IBSS_NO_SCAN (0x0) -- channel scan is not performed and
syz_80211_join_ibss waits until the interface reaches IF_OPER_UP.
JOIN_IBSS_BG_SCAN (0x1) -- channel scan is performed (takes ~ 9
seconds), syz_80211_join_ibss does not await IF_OPER_UP.
JOIN_IBSS_BG_NO_SCAN (0x2) -- channel scan is not performed,
syz_80211_join_ibss does not await IF_OPER_UP.
Local testing ensured that these syscalls are indeed able to set up an
operating network and inject packets into mac80211.
It's better to keep functionality in packages rather than in main.
It makes it reusable and better organized.
Move machine info functionality to pkg/host and do some cosmetic refactoring.
* syz-manager: finish a prototype
Extract machine info from /proc/cpuinfo and /sys/kvm*/parameters/* and
send it from syz-fuzzer to syz-manager. Append the machine info after
crash reports.
* syz-manager: refactor the code
- Add kvm parameters machine info.
- Store the machine info in the RPCServer instead of the manager.
- Store the machine info in another field instead of appending it after
the original report
- Save the machine info locally in machineInfo*.
* syz-manager: fix coding-style problems
* syz-fuzzer: improve the output from /proc/cpuinfo
Improve the machine info extracted from /proc/cpuinfo by grouping lines
with the same key.
* syz-manager: fix race condition in runInstance
* syz-fuzzer: add tests for collecting machine info
- Add some tests to test collecting machine information.
- Split readCPUInfo into scanCPUInfo so that we can test it.
* syz-fuzzer: refactor scanCPUInfo
Refactor scanCPUInfo so that no sorting is needed.
* syz-fuzzer: refactor some code
Fix some issue that was pointed out on Github.
* all: initialize vhci in linux
* executor/common_linux.h: improve vhci initialization
* pkg/repro/repro.go: add missing vhci options
* executor/common_linux.h: fix type and add missing header
* executor, pkg: do it like NetInjection
* pkg/csource/csource.go: do not emit syz_emit_vhci if vhci is not enabled
* executor/common_linux.h: fix format string
* executor/common_linux.h: initialize with memset
For som reason {0} gets complains about missing braces...
* executor/common_linux.h: simplify vhci init
* executor/common_linux.h: try to bring all available hci devices up
* executor/common_linux.h: find which hci device has been registered
* executor/common_linux.h: use HCI_VENDOR_PKT response to retrieve device id
* sys/linux/dev_vhci.txt: fix structs of inquiry and report packets
* executor/common_linux.h: remove unnecessary return statement and check vendor_pkt read size
* executor/common_linux.h: remove unnecessary return statement and check vendor_pkt read size
* sys/linux/dev_vhci.txt: pack extended_inquiry_info_t
* sys/linux/l2cap.txt: add l2cap_conf_opt struct
* executor/common_linux.h: just fill bd addr will 0xaa
* executor/common_linux.h: just fill bd addr will 0xaa
These checks still fire episodically [on gvisor instance only?].
I've done several attempts to debug this/extend checks.
But so far I have no glue and we are still seeing them.
They are rare enough to be directly debuggable and to be
something trivial. This may be some memory corruption
(kernel or our race), or some very episodic condition.
They are rare enough to be a problem, so don't include
syscall name so that they all go into a single bug bucket.
Some terms are normalised on the technical level
but may be oppressive on a societal level.
Replace them with more technically neutral terms.
See the following doc for a longer version:
https://tools.ietf.org/id/draft-knodel-terminology-00.html
We are seeing some panics that say that some disabled
syscalls somehow get into corpus.
I don't see where/how this can happen.
Add a check to syz-fuzzer to panic whenever we execute
a program with disabled syscall. Hopefull the panic
stack will shed some light.
Also add a check in manager as the last defence line
so that bad programs don't get into the corpus.
Make MakeMmap return more than 1 call.
This is a preparation for future changes.
Also remove addr/size as they are effectively
always the same and can be inferred from the target
(will also conflict with the future changes).
Also rename to MakeDataMmap to better represent
the new purpose: it's just some arbitrary mmap,
but rather mapping of the data segment.
We have _some_ limits on program length, but they are really soft.
When we ask to generate a program with 10 calls, sometimes we get
100-150 calls. There are also no checks when we accept external
programs from corpus/hub. Issue #1630 contains an example where
this crashes VM (executor limit on number of 1000 resources is
violated). Larger programs also harm the process overall (slower,
consume more memory, lead to monster reproducers, etc).
Add a set of measure for hard control over program length.
Ensure that generated/mutated programs are not too long;
drop too long programs coming from corpus/hub in manager;
drop too long programs in hub.
As a bonus ensure that mutation don't produce programs with
0 calls (which is currently possible and happens).
Fixes#1630
We are seeing some one-off panics during Deserialization
and it's unclear if it's machine memory corrpution or
an actual bug in prog. I leam towards machine memory corruption
but it's impossible to prove without seeing the orig program.
Move git revision to prog and it's more base package
(sys can import prog, prog can't import sys).
1. Show all syscalls even if they don't have coverage yet.
2. Show full syscall names.
3. Show prio/corpus/cover for paticular syscall descrimination.
This allows to check what exactly syscalls are enabled
and see prio/corpus/cover for a single syscall.
In case there is a fixed pci devlink handle "pci/pci/0000:00:10.0"
on the system (initial network namespace), it is moved to a working
network namespace.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Leak checking support was half done and did not really work.
This is heavy-lifting to make it work.
1. Move leak/fault setup into executor.
pkg/host was a wrong place for them because we need then in C repros too.
The pkg/host periodic callback functionality did not work too,
we need it in executor so that we can reuse it in C repros too.
Remove setup/callback functions in pkg/host entirely.
2. Do leak setup/checking in C repros.
The way leak checking is invoked is slightly different from fuzzer,
but much better then no support at all.
At least the checking code is shared.
3. Add Leak option to pkg/csource and -leak flag to syz-prog2c.
4. Don't enalbe leak checking in fuzzer while we are triaging initial corpus.
It's toooo slow.
5. Fix pkg/repro to do something more sane for leak bugs.
Few other minor fixes here and there.
Currently we don't fill enabled/disabled calls when we return
"all system calls are disabled" error. As the result manager
does not print any explanation for the error.
Fill in enabled/disabled calls when all calls are disabled.
Instead of always closing open fds (number 3 to 30) after each program,
add an options called EnableCloseFds. It can be passed to syz-execprog,
syz-prog2c and syz-stress via the -enable and -disable flags. Set the
default value to true. Also minimize C repros over it, except for when
repeat is enabled.
Prefix image testing errors with BUG: so that they are detected
by whoever is analyzing the output. Otherwise currently they are
dignosed as just "lost connection to test machine".
This change makes all syz-execprog, syz-prog2c and syz-stress accept
-enable and -disable flags to enable or disable additional features
(tun, net_dev, net_reset, cgroups and binfmt_misc) instead of having
a separate flag for each of them.
The default (without any flags) behavior isn't changed: syz-execprog
and syz-stress enabled all the features (provided the runtime supports
them) and syz-prog2c disables all of them.
This ability was never used but we maintain a bunch of code for it.
syzkaller also recently learned to spoof this error code
with some ptrace magic (probably intercepted control flow again
and exploited executor binary).
Drop all of it.
It's not clear from the log what program caused the executor-detected bug.
Log proc number so that it can be matched with logged program.
Also exit immediately otherwise we crash later because info is nil.
Right now syzkaller only supports coverage collected from the threads that
execute syscalls. However some useful things happen in background threads,
and it would be nice to collect coverage from those threads as well.
This change adds extra coverage support to syzkaller. This coverage is not
associated with a particular syscall, but rather with the whole program.
Executor passes extra coverage over the same ipc mechanism to syz-fuzzer
with syscall number set to -1. syz-fuzzer then passes this coverage to
syz-manager with the call name "extra".
This change requires the following kcov patch:
https://github.com/xairy/linux/pull/2
Right now syz-fuzzer does a search through /proc/kallsyms for each syscall
to check whether it's supported. Do one search instead and save the results
to a map. This speeds up syscall detection ~60 times when testing arm64 kernel
on x86. Also add another search pattern for arm64 and add some logging.
Currently syz-runtest fails to start because -debug flag is defined
both in syz-runtest and ipcconfig.
But moving sandbox functions we prevent ipcconfig from being imported into syz-runtest.
Add bulk of checks for strict parsing mode.
Probably not complete, but we can extend then in future as needed.
Turns out we can't easily use it for serialized programs
as they omit default args and during deserialization it looks like missing args.
Over time we relaxed parsing to handle all kinds of invalid programs
(excessive/missing args, wrong types, etc).
This is useful when reading old programs from corpus.
But this is harmful for e.g. reading test inputs as they can become arbitrary outdated.
For runtests which creates additional problem of executing not
what is actually written in the test (or at least what author meant).
Add strict parsing mode that does not tolerate any errors.
For now it just checks excessive syscall arguments.
We used to use len([]CallInfo) to check both, whether the slice is nil or
whether its length is zero. Since ProgInfo is not a slice, we need a
separate check for nil.
This patch add a new struct ProgInfo that for now holds info about each
call in a program []CallInfo, but in the future will be expanded with remote
coverage info. Update all the callers to use the new interface as well.
Rewind kmemleak fd before reading it second time,
otherwise we will read truncated reports.
Auto-learn what leak reports we've already seen
and ignore them in future. This is required because
there are some false positives and some fire too frequently.
So now we will hit each leak only once per manager run,
but we still will try to reproduce them.
I guess the condition was always meant to be "!ok || try > 10".
So we always ignored non-executor failures, and failed on executor ones.
I think now it's better to just retry always. If the error is something
persistent we will fail after 10 tries anyway.
executor: add support for android_untrusted_app sandbox
This adds a new sandbox type, 'android_untrusted_app', which restricts
syz-executor to the privileges which are available to third-party applications,
e.g. those installed from the Google Play store.
In particular, this uses the UID space reserved for applications (instead of
the 'setuid' sandbox, which uses the traditional 'nobody' user / 65534)
as well as a set of groups which the Android-specific kernels are aware of,
and finally ensures that the SELinux context is set appropriately.
Dependencies on libselinux are avoided by manually implementing the few
functions that are needed to change the context of the current process,
and arbitrary files. The underlying mechanisms are relatively simple.
Fixesgoogle/syzkaller#643
Test: make presubmit
Bug: http://b/112900774
The tool is run as:
$ syz-runtest -config manager.config
This runs all programs from sys/*/test/* in different modes
on actual VMs and checks results.
Fixes#603
Write coverage from unfinished syscalls.
Also detect when a syscall was blocked during execution,
even if it finished. Helpful for fallback coverage.
Fixes#580
