Goes through crash folder that is stated in the workdir.
Collects the crashes, counts and tags.
usage:
./bin/syz-reporter -config manager.cfg
Signed-off-by: Jukka Kaartinen <jukka.kaartinen@unikie.com>
We must pad data arguments with known values when serializing
them into the given destination buffer because it could
be reused and contain random bytes from previous use.
Signed-off-by: Alexander Egorenkov <Alexander.Egorenkov@ibm.com>
* sys/netbsd: adding filesystem and communication syscalls
* sys/netbsd: add fix for struct sockaddr_storage and profil(2)
* sys/netbsd: add common ioctl(2) commands
* sys/netbsd: resolving conflicts
Co-authored-by: Siddharth M <siddharth.muralee@gmail.com>
For implementing sctp_bindx(), FreeBSD uses two IPPROTO_SCTP
level socket options SCTP_BINDX_ADD_ADDR and SCTP_BINDX_REM_ADDR.
The type of the value was changed from struct sctp_getaddresses *
to struct sockaddr_in * or struct sockaddr_in6 * in
https://svnweb.freebsd.org/changeset/base/362451
csum_inet_update does not handle odd number of bytes
on big-endian architectures correctly. When calculating
the checksum of odd number of bytes, the last byte must be
interpreted as LSB on little-endian architectures and
as MSB on big-endian ones in a 16-bit half-word.
Futhermore, the checksum tests assume that the underlying architecture
is always little-endian. When a little-endian machine stores
a calculated checksum into memory, then the checksum's bytes
are automatically swapped. But this is NOT true on a big-endian
architecture.
Signed-off-by: Alexander Egorenkov <Alexander.Egorenkov@ibm.com>
There is little point in printing all targets if no errors happened.
Generation is fast, so this is not even working as a "progress bar".
Only print target if there are any errors.
Ensure that we have at least 1GB per Go compiler/linker invocation.
Go compiler/linker can consume significant amount of memory
(observed to consume at least 600MB). See #1276 for context.
And we have parallelization both on make and on go levels,
this can severe oversubscribe RAM.
Note: the result can be significantly lower than the CPU number,
but this is fine because Go builds/tests are parallelized internally.
1. Use --no-print-directory.
These "Entering directory"/"Leaving directory" messages are completely useless.
2. Use go build instead of go install.
This is just to test build and we don't install anything otherwise.
Don't mess with GOAPTH/bin unnecessarily.
3. Don't export MAKEFLAGS.
It is exported by default.
4. Suppress descriptions up-to-date check output.
It's cryptic and is not particularly useful.
Added the TIPC_NLA_NODE_KEY and TIPC_NLA_NODE_ID fields in the
socket_tipc_netlink.txt file. Created a new struct to hold the key.
Created two new descriptions for TIPC_NL_KEY_SET and TIPC_NL_KEY_FLUSH.
The related kernel commits can be seen in
https://git.kernel.org/linus/134bdac39766 and
https://git.kernel.org/linus/e1f32190cf7d.
Add template type for netlink msg with CMD and POLICY as
arguments, it can be used for all sendmsg calls in this file.
Refactor the existing code.
Signed-off-by: Amit Cohen <amitc@mellanox.com>
Most likely reports without proper stack traces were caused by a bug in the
unwinder and are now fixed in 187b96db5ca7 "x86/unwind/orc: Fix
unwind_get_return_address_ptr() for inactive tasks".
Disable trying to use questionable frames for now.
Fixes#1834
Github started adding me as reviewer on all PRs.
This is quite inconvinient because I can't figure out
if author really wants my review, or I was added just because of
the * entry. And then I can't remove myself.
Since I am looking at all PRs anyway, let's try without the * entry.
Informational mode should always produce "pass" result:
https://docs.codecov.io/docs/commit-status#informational
I can't get anything useful of these notifications.
E.g. now we have patches that add 1 line,
which gives 0% patch coverage, which results in a failure.
But we have some packages/commands not tested at all,
so we can't require contributors to make that covered.
For overall project coverage I also have not seen
any useful pass/fail results. The criteria highly depends
on the nature of the change. If we set the threshold too low,
we will get lots of effectively false failures. The current 50%
setting effectively means "never fail" anyway.
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.
Improve used terminology by using better verbs to express the effect of
the whitelist/blacklist.
This also changes executor to exclusively show respectful log messages,
and as per recent conversion, converts the last such case.
Currently we are specifically picky about what errors
during build process are reported and what are not reported.
E.g. if we fail to create some temp file due to ENOSPSC,
we don't want to send email to all kernel developers.
However, we have not seen lots of infra failures and
we've seen silent errors which were unnoticed and/or
caused confusion as to what happens and why kernels
are not updated.
Report all errors.
If needed later we may explicitly ignore some errors instead.
Fixes#1777
We patched name in struct object, but the dwarf package
caches then and then can return in subsequent invocations.
This causes a struct name to be overwritten by typedef name.
Don't mutate returned struct objects.
