Closed spikespaz closed 2 years ago
It works fine for me and also Unix socket related test work fine, which wouldn't work if Unix sockets weren't created. Maybe the additional -t
on your command line's end causes the issue.
I tried without. I am wondering if it's Apple's quarantine xattr
? What are your related settings?
I don't think it has anything to do with xattr
. I am not aware of any special setting but -t
means terminate after connection loss and this is what may be happening here that it terminates on you after the first connection, even connection via the control channel, is lost.
I tried without. I am wondering if it's Apple's quarantine
xattr
? What are your related settings?
I have already tried it without -t
.
I have created a Homebrew tap so that you can attempt to reproduce my issue. Try updating all your dependencies and installing swtpm
from here: https://github.com/spikespaz/homebrew-jacob
brew update
brew upgrade
brew tap spikespaz/jacob
brew install swtpm
which swtpm
This should help bring our machines to parity.
Also, make sure to use Homebrew from /opt/homebrew
. You could be using Rosetta and compiling for x86_64
.
I shouldn't have to install this from your repo. I compiled from the same sources as you did and it should lead to the same result.
Does this test also pass for you with unix+unix
? You ran make test
it sounds like, so it should work.
$ cd tests/
$ ./test_tpm2_locality
[...]
==== Starting swtpm with interfaces socket+socket ====
Locality outside valid range of [0..4].
OK
==== Starting swtpm with interfaces socket+unix ====
Locality outside valid range of [0..4].
OK
==== Starting swtpm with interfaces unix+unix ====
Locality outside valid range of [0..4].
OK
A number of issues with your assumption that we get the same binary. Are your formulae up-to-date? Are you sure that we are using the same environment variables? One of ours may be polluted. Are you sure your compilation environment isn't using Rosetta by default?
I agree, if the make tests pass, it should work. But it doesn't. That is the exact reason we are troubleshooting right now.
I am now testing using Rosetta 2 to remove that potential discontinuity.
$ arch -x86_64 /usr/local/brew tap spikespaz/jacob
$ arch -x86_64 /usr/local/brew install --interactive swtpm
$ arch -x86_64 ./autogen.sh --prefix=/usr/local/ --with-openssl
$ cd tests/
$ arch -x86_64 ./test_tpm2_locality
Need Linux to run test with CUSE interface.
Need Linux to run test with CUSE interface.
==== Starting swtpm with interfaces socket+socket ====
swtpm: Data client disconnected
Locality outside valid range of [0..4].
OK
==== Starting swtpm with interfaces socket+unix ====
swtpm: Data client disconnected
Locality outside valid range of [0..4].
OK
==== Starting swtpm with interfaces unix+unix ====
Locality outside valid range of [0..4].
swtpm: Data client disconnected
swtpm: Data client disconnected
swtpm: Data client disconnected
swtpm: Data client disconnected
swtpm: Data client disconnected
OK
$ arch -x86_64 make -j4 check >> ~/Desktop/swtpm-make-check.log
I am getting the exact same results without using arch -x86_64
.
Does any part of these two libraries use OpenSSL 1.1? Because I have some links to OpenSSL 1.1 showing here.
Well, it started working. No idea why.
Describe the bug I should expect the socket file specified to be created in the directory specified, but instead, no file is created and the terminal remains blank with no logging output.
To Reproduce Steps to reproduce the behavior:
mkdir /tmp/emulated_tpm/
Expected behavior I should expect the socket file specified to be created in the directory specified, but instead, no file is created and the terminal remains blank with no logging output.
Desktop (please complete the following information):
Versions of relevant components
cd8025fa
)Log files There are no log files (AFAIK).