While here also fix a fd leak in got-read-pack. We were dup'ing imsg.fd without closing imsg.fd later; instead just use imsg_get_fd() to extract the file descriptor. Tested by falsifian and Kyle Ackerman, thanks! 'go ahead' stsp@

Move some functions from object_parse.c into hash.c. These functions either require hash.c code anyway or contain object ID implementation internals. Add a new file object_qid.c, for got_object_id_queue and got_object_qid. This new file must be linked to virtually every program.

Simply the SHA detection by not predicating on libcrypto, but instead checking individual header files.

Remove sha1.h as this is found portably across systems.

It will soon grow functions to deal with sha256 too. stsp@ agrees.

In preparation for wide sha256 support; stsp@ agrees. Change done mechanically with find . -iname \*.[cy] -exec sam {} + X ,x/<sha1\.h>/i/\n#include <sha2.h>

This is an initial barebones implementation which provides the absolute minimum of functionality required to serve got(1) and git(1) clients. Basic fetch/send functionality has been tested and seems to work here, but this server is not yet expected to be stable. More testing is welcome. See the man pages for setup instructions. The current design uses one reader and one writer process per repository, which will have to be extended to N readers and N writers in the future. At startup, each process will chroot(2) into its assigned repository. This works because gotd(8) can only be started as root, and will then fork+exec, chroot, and privdrop. At present the parent process runs with the following pledge(2) promises: "stdio rpath wpath cpath proc getpw sendfd recvfd fattr flock unix unveil" The parent is the only process able to modify the repository in a way that becomes visible to Git clients. The parent uses unveil(2) to restrict its view of the filesystem to /tmp and the repositories listed in the configuration file gotd.conf(5). Per-repository chroot(2) processes use "stdio rpath sendfd recvfd". The writer defers to the parent for modifying references in the repository to point at newly uploaded commits. The reader is fine without such help, because Git repositories can be read without having to create any lock-files. gotd(8) requires a dedicated user ID, which should own repositories on the filesystem, and a separate secondary group, which should not have filesystem-level repository access, and must be allowed access to the gotd(8) socket. To obtain Git repository access, users must be members of this secondary group, and must have their login shell set to gotsh(1). gotsh(1) connects to the gotd(8) socket and speaks Git-protocol towards the client on the other end of the SSH connection. gotsh(1) is not an interactive command shell. At present, authenticated clients are granted read/write access to all repositories and all references (except for the "refs/got/" and the "refs/remotes/" namespaces, which are already being protected from modification). While complicated access control mechanism are not a design goal, making it possible to safely offer anonymous Git repository access over ssh(1) is on the road map.

Now that the handling of including sys/queue.h is better, there's no need to remove those lines from the source. Copy the location of those original sys/queue.h lines from upstream at the same line number, so as to avoid any conflicts in the future.

Thanks to the design of Got, the libexec helpers don't need any resource (in fact they run under pledge "stdio recvfd" on OpenBSD) and so using cap_enter(2) on FreeBSD is dead-easy. While the main process can't be sandboxed on FreeBSD (needs to exec the helpers), all the tough work is done by these small libexec helpers which is also the biggest attack surface. tested by naddy, ok thomas

landlock is a new set of linux APIs that is conceptually similar to unveil(2): the idea is to restrict what a process can do on a specified part of the filesystem. There are some differences in the behaviour: the major one being that the landlock ruleset is inherited across execve(2). This just restricts the libexec helpers by completely revoking ANY filesystem access; after all they are the biggest attack surface. got send/fetch/clone *may* end up spawning ssh(1), so at the moment is not possible to landlock the main process. From Omar Polo.

This adds the capability to compile got-portable on FreeBSD.

This commit modifies the GoT main branch to be able to compile it under linux.

ok millert, naddy

ok millert, naddy

POSIX says the limits defined there are available from <limits.h>, which almost all affected source files already included anyway. ok millert stsp

ok stsp