mingw-w64's pthread_unistd.h had a bug that mistakenly (because there is
no support for the *lockfile() functions required[1]) defined
_POSIX_THREAD_SAFE_FUNCTIONS and that was being worked around since
3ecd153a3b (compat/mingw: support MSys2-based MinGW build, 2016-01-14).
The bug was fixed in winphtreads, but as a side effect, leaves the
reentrant functions from time.h no longer visible and therefore breaks
the build.
Since the intention all along was to avoid using the fallback functions,
formalize the use of POSIX by setting the corresponding feature flag and
compile out the implementation for the fallback functions.
[1] https://unix.org/whitepapers/reentrant.html
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Cygwin emulates Unix sockets by writing files with custom contents and
then marking them as system files.
The tricky problem is that while the file is written and its `system`
bit is set, it is still identified as a file. This caused test failures
when Git is too fast looking for the Unix sockets and then complains
that there is a plain file in the way.
Let's work around this by adding a delayed retry loop, specifically for
Cygwin.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Tested-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This compatilibity implementation has been returning a wrong type,
ever since 731043fd (Add compat/unsetenv.c ., 2006-01-25) added to
the system, yet nobody noticed it in the past 16 years, presumably
because no code checks failures in their unsetenv() calls. Sigh.
For now, make it always succeed.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, git will share its console with all its children (unless
they create their own), and is therefore possible that any of them
that might change the settings for it could affect its operations once
completed.
Refactor the platform specific functionality to save the terminal
settings and expand it to also do so for the output handler.
This will allow for the state of the terminal to be saved and
restored around a child that might misbehave (ex vi) which will
be implemented next.
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
gcc will helpfully raise a -Wcast-function-type warning when casting
between functions that might have incompatible return types
(ex: GetUserNameExW returns bool which is only half the size of the
return type from FARPROC which is long long), so create a new type that
could be used as a completely generic function pointer and cast through
it instead.
Additionaly remove the -Wno-incompatible-pointer-types temporary
flag added in 27e0c3c (win32: allow building with pedantic mode
enabled, 2021-09-03), as it will be no longer needed.
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Here, GCC warns about every use of the INIT_PROC_ADDR macro, for example:
In file included from compat/mingw.c:8:
compat/mingw.c: In function 'mingw_strftime':
compat/win32/lazyload.h:38:12: warning: assignment to
'size_t (*)(char *, size_t, const char *, const struct tm *)'
{aka 'long long unsigned int (*)(char *, long long unsigned int,
const char *, const struct tm *)'} from incompatible pointer type
'FARPROC' {aka 'long long int (*)()'} [-Wincompatible-pointer-types]
38 | (function = get_proc_addr(&proc_addr_##function))
| ^
compat/mingw.c:1014:6: note: in expansion of macro 'INIT_PROC_ADDR'
1014 | if (INIT_PROC_ADDR(strftime))
| ^~~~~~~~~~~~~~
(message wrapped for convenience). Insert a cast to keep the compiler
happy. A cast is fine in these cases because they are generic function
pointer values that have been looked up in a DLL.
Helped-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change various places that hardcode the names of these two files to
refer to either $(GENERATED_H), or to a new generated-hdrs
target. That target is consistent with the *-objs targets I recently
added in 029bac01a8 (Makefile: add {program,xdiff,test,git,fuzz}-objs
& objects targets, 2021-02-23).
A subsequent commit will add a new generated hook-list.h. By doing
this refactoring we'll only need to add the new file to the
GENERATED_H variable, not EXCEPT_HDRS, the vcbuild/README etc.
Hardcoding command-list.h there seems to have been a case of
copy/paste programming in 976aaedca0 (msvc: add a Makefile target to
pre-generate the Visual Studio solution, 2019-07-29). The
config-list.h was added later in 709df95b78 (help: move
list_config_help to builtin/help, 2020-04-16).
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Set an ACL on the named pipe to allow the well-known group EVERYONE
to read and write to the IPC server's named pipe. In the event that
the daemon was started with elevation, allow non-elevated clients
to communicate with the daemon.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create "ipc-debug" category events to log unexpected errors
when creating Simple-IPC connections.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add `command_len` argument to the Simple IPC API.
In my original Simple IPC API, I assumed that the request would always
be a null-terminated string of text characters. The `command`
argument was just a `const char *`.
I found a caller that would like to pass a binary command to the
daemon, so I am amending the Simple IPC API to receive `const char
*command, size_t command_len` arguments.
I considered changing the `command` argument to be a `void *`, but the
IPC layer simply passes it to the pkt-line layer which takes a `const
char *`, so to avoid confusion I left it as is.
Note, the response side has always been a `struct strbuf` which
includes the buffer and length, so we already support returning a
binary answer. (Yes, it feels a little weird returning a binary
buffer in a `strbuf`, but it works.)
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 2f732bf15e (tr2: log parent process name, 2021-07-21) we started
logging parent process names, but only logged all parents on Windows.
on Linux only the name of the immediate parent process was logged.
Extend the functionality added there to also log full parent chain on
Linux.
This requires us to lookup "/proc/<getppid()>/stat" instead of
"/proc/<getppid()>/comm". The "comm" file just contains the name of the
process, but the "stat" file has both that information, and the parent
PID of that process, see procfs(5). We parse out the parent PID of our
own parent, and recursively walk the chain of "/proc/*/stat" files all
the way up the chain. A parent PID of 0 indicates the end of the
chain.
It's possible given the semantics of Linux's PID files that we end up
getting an entirely nonsensical chain of processes. It could happen if
e.g. we have a chain of processes like:
1 (init) => 321 (bash) => 123 (git)
Let's assume that "bash" was started a while ago, and that as shown
the OS has already cycled back to using a lower PID for us than our
parent process. In the time it takes us to start up and get to
trace2_collect_process_info(TRACE2_PROCESS_INFO_STARTUP) our parent
process might exit, and be replaced by an entirely different process!
We'd racily look up our own getppid(), but in the meantime our parent
would exit, and Linux would have cycled all the way back to starting
an entirely unrelated process as PID 321.
If that happens we'll just silently log incorrect data in our ancestry
chain. Luckily we don't need to worry about this except in this
specific cycling scenario, as Linux does not have PID
randomization. It appears it once did through a third-party feature,
but that it was removed around 2006[1]. For anyone worried about this
edge case raising PID_MAX via "/proc/sys/kernel/pid_max" will mitigate
it, but not eliminate it.
One thing we don't need to worry about is getting into an infinite
loop when walking "/proc/*/stat". See 353d3d77f4 (trace2: collect
Windows-specific process information, 2019-02-22) for the related
Windows code that needs to deal with that, and [2] for an explanation
of that edge case.
Aside from potential race conditions it's also a bit painful to
correctly parse the process name out of "/proc/*/stat". A simpler
approach is to use fscanf(), see [3] for an implementation of that,
but as noted in the comment being added here it would fail in the face
of some weird process names, so we need our own parse_proc_stat() to
parse it out.
With this patch the "ancestry" chain for a trace2 event might look
like this:
$ GIT_TRACE2_EVENT=/dev/stdout ~/g/git/git version | grep ancestry | jq -r .ancestry
[
"bash",
"screen",
"systemd"
]
And in the case of naughty process names like the following. This uses
perl's ability to use prctl(PR_SET_NAME, ...). See
Perl/perl5@7636ea95c5 (Set the legacy process name with prctl() on
assignment to $0 on Linux, 2010-04-15)[4]:
$ perl -e '$0 = "(naughty\nname)"; system "GIT_TRACE2_EVENT=/dev/stdout ~/g/git/git version"' | grep ancestry | jq -r .ancestry
[
"sh",
"(naughty\nname)",
"bash",
"screen",
"systemd"
]
1. https://grsecurity.net/news#grsec2110
2. https://lore.kernel.org/git/48a62d5e-28e2-7103-a5bb-5db7e197a4b9@jeffhostetler.com/
3. https://lore.kernel.org/git/87o8agp29o.fsf@evledraar.gmail.com/
4. 7636ea95c5
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change code added in 2f732bf15e (tr2: log parent process name,
2021-07-21) to use a switch statement without a "default" branch to
have the compiler error if this code ever drifts out of sync with the
members of the "enum trace2_process_info_reason".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a subsequent commit I'll be replacing most of this code to log N
parents, but let's first fix bugs introduced in the recent
2f732bf15e (tr2: log parent process name, 2021-07-21).
It was using the strbuf_read_file() in the wrong way, its return value
is either a length or a negative value on error. If we didn't have a
procfs, or otherwise couldn't access it we'd end up pushing an empty
string to the trace2 ancestry array.
It was also using the strvec_push() API the wrong way. That API always
does an xstrdup(), so by detaching the strbuf here we'd leak
memory. Let's instead pass in our pointer for strvec_push() to
xstrdup(), and then free our own strbuf. I do have some WIP changes to
make strvec_push_nodup() non-static, which makes this and some other
callsites nicer, but let's just follow the prevailing pattern of using
strvec_push() for now.
We'll also need to free that "procfs_path" strbuf whether or not
strbuf_read_file() succeeds, which was another source of memory leaks
in 2f732bf15e, i.e. we'd leak that memory as well if we weren't on a
system where we could read the file from procfs.
Let's move all the freeing of the memory to the end of the
function. If we're still at STRBUF_INIT with "name" due to not having
taken the branch where the strbuf_read_file() succeeds freeing it is
redundant. So we could move it into the body of the "if", but just
handling freeing the same way for all branches of the function makes
it more readable.
In combination with the preceding commit this makes all of
t[0-9]*trace2*.sh pass under SANITIZE=leak on Linux.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rewrite a comment added in 2f732bf15e (tr2: log parent process name,
2021-07-21) to describe what we might do under
TRACE2_PROCESS_INFO_EXIT in the future, instead of vaguely referring
to "something extra".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
I'm fairly sure that there is no way on Linux to inspect the process
tree without using procfs, any tool such as ps(1), top(1) etc. that
shows this sort of information ultimately looks the information up in
procfs.
So let's remove this comment added in 2f732bf15e (tr2: log parent
process name, 2021-07-21), it's setting us up for an impossible task.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In preparation to building with pedantic mode enabled, change a couple
of places where the current mingw gcc compiler provided with the SDK
reports issues.
A full fix for the incompatible use of (void *) to store function
pointers has been punted, with the minimal change to instead use a
generic function pointer (FARPROC), and therefore the (hopefully)
temporary need to disable incompatible pointer warnings.
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
xmalloc() dies on error, allows zero-sized allocations and enforces
GIT_ALLOC_LIMIT for testing. Our mmap replacement doesn't need any of
that. Let's cut out the wrapper, reject zero-sized requests as required
by POSIX and use malloc(3) directly. Allocation errors were needlessly
handled by git_mmap() before; this code becomes reachable now.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When performing a rebase, rmdir() is called on the folder .git/logs. On
Unix rmdir() exits without deleting anything in case .git/logs is a
symbolic link but the equivalent functions on Windows (_rmdir, _wrmdir
and RemoveDirectoryW) do not behave the same and remove the folder if it
is symlinked even if it is not empty.
This creates issues when folders in .git/ are symlinks which is
especially the case when git-repo[1] is used: It replaces `.git/logs/`
with a symlink.
One such issue is that the _target_ of that symlink is removed e.g.
during a `git rebase`, where `delete_reflog("REBASE_HEAD")` will not
only try to remove `.git/logs/REBASE_HEAD` but then recursively try to
remove the parent directories until an error occurs, a technique that
obviously relies on `rmdir()` refusing to remove a symlink.
This was reported in https://github.com/git-for-windows/git/issues/2967.
This commit updates mingw_rmdir() so that its behavior is the same as
Linux rmdir() in case of symbolic links.
To verify that Git does not regress on the reported issue, this patch
adds a regression test for the `git rebase` symptom, even if the same
`rmdir()` behavior is quite likely to cause potential problems in other
Git commands as well.
[1]: git-repo is a python tool built on top of Git which helps manage
many Git repositories. It stores all the .git/ folders in a central
place by taking advantage of symbolic links.
More information: https://gerrit.googlesource.com/git-repo/
Signed-off-by: Thomas Bétous <tomspycell@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It can be useful to tell who invoked Git - was it invoked manually by a
user via CLI or script? By an IDE? In some cases - like 'repo' tool -
we can influence the source code and set the GIT_TRACE2_PARENT_SID
environment variable from the caller process. In 'repo''s case, that
parent SID is manipulated to include the string "repo", which means we
can positively identify when Git was invoked by 'repo' tool. However,
identifying parents that way requires both that we know which tools
invoke Git and that we have the ability to modify the source code of
those tools. It cannot scale to keep up with the various IDEs and
wrappers which use Git, most of which we don't know about. Learning
which tools and wrappers invoke Git, and how, would give us insight to
decide where to improve Git's usability and performance.
Unfortunately, there's no cross-platform reliable way to gather the name
of the parent process. If procfs is present, we can use that; otherwise
we will need to discover the name another way. However, the process ID
should be sufficient to look up the process name on most platforms, so
that code may be shareable.
Git for Windows gathers similar information and logs it as a "data_json"
event. However, since "data_json" has a variable format, it is difficult
to parse effectively in some languages; instead, let's pursue a
dedicated "cmd_ancestry" event to record information about the ancestry
of the current process and a consistent, parseable way.
Git for Windows also gathers information about more than one generation
of parent. In Linux further ancestry info can be gathered with procfs,
but it's unwieldy to do so. In the interest of later moving Git for
Windows ancestry logging to the 'cmd_ancestry' event, and in the
interest of later adding more ancestry to the Linux implementation - or
of adding this functionality to other platforms which have an easier
time walking the process tree - let's make 'cmd_ancestry' accept an
array of parentage.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To pave the way for non-Windows platforms to define
trace2_collect_process_info(), reorganize the stub-or-definition schema
to something which doesn't directly reference Windows.
Platforms which want to collect parent process information in the
future should:
1. Add an implementation to compat/ (e.g. compat/somearch/procinfo.c)
2. Add that object to COMPAT_OBJS to config.mak.uname
(e.g. COMPAT_OBJS += compat/somearch/procinfo.o)
3. Define HAVE_PLATFORM_PROCINFO in config.mak.uname
In the Windows case, this definition lives in
compat/win32/trace2_win32_process_info.c, which is already conditionally
added to COMPAT_OBJS; so let's add HAVE_PLATFORM_PROCINFO to hint to the
build that compat/stub/procinfo.c should not be used.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Simple IPC always requires threads (in addition to various
platform-specific IPC support). Fix the ifdefs in the Makefile
to define SUPPORTS_SIMPLE_IPC when appropriate.
Previously, the Unix version of the code would only verify that
Unix domain sockets were available.
This problem was reported here:
https://lore.kernel.org/git/YKN5lXs4AoK%2FJFTO@coredump.intra.peff.net/T/#m08be8f1942ea8a2c36cfee0e51cdf06489fdeafc
Reported-by: Randall S. Becker <rsbecker@nexbridge.com>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apparently this is not supported with Microsoft's Universal C Runtime.
So let's not actually do that.
Instead, just return success because we _know_ that we expect the `NUL`
device to be present.
Side note: it is possible to turn off the "Null device driver" and
thereby disable `NUL`. Too many things are broken if this driver is
disabled, therefore it is not worth bothering to try to detect its
presence when `access()` is called.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
commit 5c327502 (MacOS: precompose_argv_prefix(), 2021-02-03) uses
the function precompose_string_if_needed() internally. It is only
used from precompose_argv_prefix() and therefore static in
compat/precompose_utf8.c
Expose this function, it will be used in the next commit.
While there, allow passing a NULL pointer, which will return NULL.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create Unix domain socket based implementation of "simple-ipc".
A set of `ipc_client` routines implement a client library to connect
to an `ipc_server` over a Unix domain socket, send a simple request,
and receive a single response. Clients use blocking IO on the socket.
A set of `ipc_server` routines implement a thread pool to listen for
and concurrently service client connections.
The server creates a new Unix domain socket at a known location. If a
socket already exists with that name, the server tries to determine if
another server is already listening on the socket or if the socket is
dead. If socket is busy, the server exits with an error rather than
stealing the socket. If the socket is dead, the server creates a new
one and starts up.
If while running, the server detects that its socket has been stolen
by another server, it automatically exits.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Create Windows implementation of "simple-ipc" using named pipes.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On some platforms, open() reportedly returns EINTR when opening regular
files and we receive a signal (usually SIGALRM from our progress meter).
This shouldn't happen, as open() should be a restartable syscall, and we
specify SA_RESTART when setting up the alarm handler. So it may actually
be a kernel or libc bug for this to happen. But it has been reported on
at least one version of Linux (on a network filesystem):
https://lore.kernel.org/git/c8061cce-71e4-17bd-a56a-a5fed93804da@neanderfunk.de/
as well as on macOS starting with Big Sur even on a regular filesystem.
We can work around it by retrying open() calls that get EINTR, just as
we do for read(), etc. Since we don't ever _want_ to interrupt an open()
call, we can get away with just redefining open, rather than insisting
all callsites use xopen().
We actually do have an xopen() wrapper already (and it even does this
retry, though there's no indication of it being an observed problem back
then; it seems simply to have been lifted from xread(), etc). But it is
used hardly anywhere, and isn't suitable for general use because it will
die() on error. In theory we could combine the two, but it's awkward to
do so because of the variable-args interface of open().
This patch adds a Makefile knob for enabling the workaround. It's not
enabled by default for any platforms in config.mak.uname yet, as we
don't have enough data to decide how common this is (I have not been
able to reproduce on either Linux or Big Sur myself). It may be worth
enabling preemptively anyway, since the cost is pretty low (if we don't
see an EINTR, it's just an extra conditional).
However, note that we must not enable this on Windows. It doesn't do
anything there, and the macro overrides the existing mingw_open()
redirection. I've added a preemptive #undef here in the mingw header
(which is processed first) to just quietly disable it (we could also
make it an #error, but there is little point in being so aggressive).
Reported-by: Aleksey Kliger <alklig@microsoft.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before checking out a file, we have to confirm that all of its leading
components are real existing directories. And to reduce the number of
lstat() calls in this process, we cache the last leading path known to
contain only directories. However, when a path collision occurs (e.g.
when checking out case-sensitive files in case-insensitive file
systems), a cached path might have its file type changed on disk,
leaving the cache on an invalid state. Normally, this doesn't bring
any bad consequences as we usually check out files in index order, and
therefore, by the time the cached path becomes outdated, we no longer
need it anyway (because all files in that directory would have already
been written).
But, there are some users of the checkout machinery that do not always
follow the index order. In particular: checkout-index writes the paths
in the same order that they appear on the CLI (or stdin); and the
delayed checkout feature -- used when a long-running filter process
replies with "status=delayed" -- postpones the checkout of some entries,
thus modifying the checkout order.
When we have to check out an out-of-order entry and the lstat() cache is
invalid (due to a previous path collision), checkout_entry() may end up
using the invalid data and thrusting that the leading components are
real directories when, in reality, they are not. In the best case
scenario, where the directory was replaced by a regular file, the user
will get an error: "fatal: unable to create file 'foo/bar': Not a
directory". But if the directory was replaced by a symlink, checkout
could actually end up following the symlink and writing the file at a
wrong place, even outside the repository. Since delayed checkout is
affected by this bug, it could be used by an attacker to write
arbitrary files during the clone of a maliciously crafted repository.
Some candidate solutions considered were to disable the lstat() cache
during unordered checkouts or sort the entries before passing them to
the checkout machinery. But both ideas include some performance penalty
and they don't future-proof the code against new unordered use cases.
Instead, we now manually reset the lstat cache whenever we successfully
remove a directory. Note: We are not even checking whether the directory
was the same as the lstat cache points to because we might face a
scenario where the paths refer to the same location but differ due to
case folding, precomposed UTF-8 issues, or the presence of `..`
components in the path. Two regression tests, with case-collisions and
utf8-collisions, are also added for both checkout-index and delayed
checkout.
Note: to make the previously mentioned clone attack unfeasible, it would
be sufficient to reset the lstat cache only after the remove_subtree()
call inside checkout_entry(). This is the place where we would remove a
directory whose path collides with the path of another entry that we are
currently trying to check out (possibly a symlink). However, in the
interest of a thorough fix that does not leave Git open to
similar-but-not-identical attack vectors, we decided to intercept
all `rmdir()` calls in one fell swoop.
This addresses CVE-2021-21300.
Co-authored-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
The following sequence leads to a "BUG" assertion running under MacOS:
DIR=git-test-restore-p
Adiarnfd=$(printf 'A\314\210')
DIRNAME=xx${Adiarnfd}yy
mkdir $DIR &&
cd $DIR &&
git init &&
mkdir $DIRNAME &&
cd $DIRNAME &&
echo "Initial" >file &&
git add file &&
echo "One more line" >>file &&
echo y | git restore -p .
Initialized empty Git repository in /tmp/git-test-restore-p/.git/
BUG: pathspec.c:495: error initializing pathspec_item
Cannot close git diff-index --cached --numstat
[snip]
The command `git restore` is run from a directory inside a Git repo.
Git needs to split the $CWD into 2 parts:
The path to the repo and "the rest", if any.
"The rest" becomes a "prefix" later used inside the pathspec code.
As an example, "/path/to/repo/dir-inside-repå" would determine
"/path/to/repo" as the root of the repo, the place where the
configuration file .git/config is found.
The rest becomes the prefix ("dir-inside-repå"), from where the
pathspec machinery expands the ".", more about this later.
If there is a decomposed form, (making the decomposing visible like this),
"dir-inside-rep°a" doesn't match "dir-inside-repå".
Git commands need to:
(a) read the configuration variable "core.precomposeunicode"
(b) precocompose argv[]
(c) precompose the prefix, if there was any
The first commit,
76759c7dff "git on Mac OS and precomposed unicode"
addressed (a) and (b).
The call to precompose_argv() was added into parse-options.c,
because that seemed to be a good place when the patch was written.
Commands that don't use parse-options need to do (a) and (b) themselfs.
The commands `diff-files`, `diff-index`, `diff-tree` and `diff`
learned (a) and (b) in
commit 90a78b83e0 "diff: run arguments through precompose_argv"
Branch names (or refs in general) using decomposed code points
resulting in decomposed file names had been fixed in
commit 8e712ef6fc "Honor core.precomposeUnicode in more places"
The bug report from above shows 2 things:
- more commands need to handle precomposed unicode
- (c) should be implemented for all commands using pathspecs
Solution:
precompose_argv() now handles the prefix (if needed), and is renamed into
precompose_argv_prefix().
Inside this function the config variable core.precomposeunicode is read
into the global variable precomposed_unicode, as before.
This reading is skipped if precomposed_unicode had been read before.
The original patch for preocomposed unicode, 76759c7dff, placed
precompose_argv() into parse-options.c
Now add it into git.c::run_builtin() as well. Existing precompose
calls in diff-files.c and others may become redundant, and if we
audit the callflows that reach these places to make sure that they
can never be reached without going through the new call added to
run_builtin(), we might be able to remove these existing ones.
But in this commit, we do not bother to do so and leave these
precompose callsites as they are. Because precompose() is
idempotent and can be called on an already precomposed string
safely, this is safer than removing existing calls without fully
vetting the callflows.
There is certainly room for cleanups - this change intends to be a bug fix.
Cleanups needs more tests in e.g. t/t3910-mac-os-precompose.sh, and should
be done in future commits.
[1] git-bugreport-2021-01-06-1209.txt (git can't deal with special characters)
[2] https://lore.kernel.org/git/A102844A-9501-4A86-854D-E3B387D378AA@icloud.com/
Reported-by: Daniel Troger <random_n0body@icloud.com>
Helped-By: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Something changed in `vcpkg` (which we use in our Visual C++ build to
provide the dependencies such as libcurl) and our `vs-build` job started
failing in CI. The reason is that we had a work-around in place to help
CMake find iconv, and this work-around is neither needed nor does it
work anymore.
For the full discussion with the vcpkg project, see this comment:
https://github.com/microsoft/vcpkg/issues/14780#issuecomment-735368280
Signed-off-by: Dennis Ameling <dennis@dennisameling.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 1af265f0 (compat/bswap.h: simplify MSVC endianness
detection, 2020-11-08) we attempted to simplify code by assuming MSVC
builds will be for little-endian machines, since only unusably old
versions of MSVC supported big-endian MIPS and m68k architectures.
However, it's possible that MSVC could be ported to build for a
big-endian architecture again, so the simplification wasn't as
future-proof as hoped.
So let's go back to the old way of detecting MSVC, and then checking
architecture from a list of little-endian architecture macros.
Note that MSVC does not treat ARM64 as bi-endian, so we can safely treat
it as little-endian.
Helped-by: brian m. carlson <sandals@crustytoothpaste.net>
Helped-by: Jeff King <peff@peff.net>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Daniel Gurney <dgurney99@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Modern MSVC or Windows versions don't support big-endian, so it's
unnecessary to consider architectures when using it.
This also makes ARM64 MSVC builds succeed.
Helped-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Daniel Gurney <dgurney99@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 554544276a (*.[ch]: remove extern from function declarations using
spatch, 2019-04-29), `extern` on function declarations were declared to
be redundant and thus removed from the codebase. An `extern` was
accidentally reintroduced in 08809c09aa (mingw: add a helper function to
attach GDB to the current process, 2020-02-13).
Remove this spurious `extern`.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our put_be32() routine and its variants (get_be32(), put_be64(), etc)
has two implementations: on some platforms we cast memory in place and
use nothl()/htonl(), which can cause unaligned memory access. And on
others, we pick out the individual bytes using bitshifts.
This introduces extra complexity, and sometimes causes compilers to
generate warnings about type-punning. And it's not clear there's any
performance advantage.
This split goes back to 660231aa97 (block-sha1: support for
architectures with memory alignment restrictions, 2009-08-12). The
unaligned versions were part of the original block-sha1 code in
d7c208a92e (Add new optimized C 'block-sha1' routines, 2009-08-05),
which says it is:
Based on the mozilla SHA1 routine, but doing the input data accesses a
word at a time and with 'htonl()' instead of loading bytes and shifting.
Back then, Linus provided timings versus the mozilla code which showed a
27% improvement:
https://lore.kernel.org/git/alpine.LFD.2.01.0908051545000.3390@localhost.localdomain/
However, the unaligned loads were either not the useful part of that
speedup, or perhaps compilers and processors have changed since then.
Here are times for computing the sha1 of 4GB of random data, with and
without -DNO_UNALIGNED_LOADS (and BLK_SHA1=1, of course). This is with
gcc 10, -O2, and the processor is a Core i9-9880H.
[stock]
Benchmark #1
An error occurred
: t/helper/test-tool sha1 <foo.rand
Time (mean ± σ): 6.638 s ± 0.081 s [User: 6.269 s, System: 0.368 s]
Range (min … max): 6.550 s … 6.841 s 10 runs
[-DNO_UNALIGNED_LOADS]
Benchmark #1
An error occurred
: t/helper/test-tool sha1 <foo.rand
Time (mean ± σ): 6.418 s ± 0.015 s [User: 6.058 s, System: 0.360 s]
Range (min … max): 6.394 s … 6.447 s 10 runs
And here's the same test run on an AMD A8-7600, using gcc 8.
[stock]
Benchmark #1
An error occurred
: t/helper/test-tool sha1 <foo.rand
Time (mean ± σ): 11.721 s ± 0.113 s [User: 10.761 s, System: 0.951 s]
Range (min … max): 11.509 s … 11.861 s 10 runs
[-DNO_UNALIGNED_LOADS]
Benchmark #1
An error occurred
: t/helper/test-tool sha1 <foo.rand
Time (mean ± σ): 11.744 s ± 0.066 s [User: 10.807 s, System: 0.928 s]
Range (min … max): 11.637 s … 11.863 s 10 runs
So the unaligned loads don't seem to help much, and actually make things
worse. It's possible there are platforms where they provide more
benefit, but:
- the non-x86 platforms for which we use this code are old and obscure
(powerpc and s390).
- the main caller that cares about performance is block-sha1. But
these days it is rarely used anyway, in favor of sha1dc (which is
already much slower, and nobody seems to have cared that much).
Let's just drop unaligned versions entirely in the name of simplicity.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Orgad Shaneh <orgads@gmail.com>
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Orgad Shaneh <orgads@gmail.com>
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update mingw_unlink() to first try to delete the file with existing
permissions before trying to force it.
Windows throws an error when trying to delete a read-only file. The
mingw_unlink() compatibility wrapper always tries to _wchmod(666) the
file before calling _wunlink() to avoid that error. However, since
most files in the worktree are already writable, this is usually
wasted effort.
Update mingw_unlink() to just call DeleteFileW() directly and if that
succeeds return. If that fails, fall back into the existing code path
to update the permissions and use _wunlink() to get the existing
error code mapping.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We eventually want to drop the argv_array name and just use strvec
consistently. There's no particular reason we have to do it all at once,
or care about interactions between converted and unconverted bits.
Because of our preprocessor compat layer, the names are interchangeable
to the compiler (so even a definition and declaration using different
names is OK).
This patch converts remaining files from the first half of the alphabet,
to keep the diff to a manageable size.
The conversion was done purely mechanically with:
git ls-files '*.c' '*.h' |
xargs perl -i -pe '
s/ARGV_ARRAY/STRVEC/g;
s/argv_array/strvec/g;
'
and then selectively staging files with "git add '[abcdefghjkl]*'".
We'll deal with any indentation/style fallouts separately.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 897d68e7af (Makefile: use curl-config --cflags, 2020-03-26), we
taught the build process to use `curl-config --cflags` to make sure that
it can find cURL's headers.
In the MSVC build, this is completely bogus because we're running in a
Git for Windows SDK whose `curl-config` supports the _GCC_ build.
Let's just ignore each and every `-I<path>` option where `<path>` points
to GCC/Clang specific headers.
Reported by Jeff Hostetler in
https://github.com/microsoft/git/issues/275.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
ccd469450a (date.c: switch to reentrant {gm,local}time_r, 2019-11-28)
removes the only gmtime() call we had and moves to gmtime_r() which
doesn't have the same portability problems.
Remove the compat gmtime code since it is no longer needed, and confirm
by successfull running t4212 in FreeBSD 9.3 amd64 (the oldest I could
get a hold off).
Further work might be needed to ensure 32bit time_t systems (like FreeBSD
i386) will handle correctly the overflows tested in t4212, but that is
orthogonal to this change, and it doesn't change the current behaviour
as neither gmtime() or gmtime_r() will ever return NULL on those systems
because time_t is unsigned.
Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In Linux with musl libc, we have this inclusion chain:
compat/regex/regex.c:69
`-> compat/regex/regex_internal.h
`-> /usr/include/stdlib.h
`-> /usr/include/features.h
`-> /usr/include/alloca.h
In that inclusion chain, `<features.h>` claims it's _BSD_SOURCE
compatible when it's NOT asked to be either
{_POSIX,_GNU,_XOPEN,_BSD}_SOURCE, or __STRICT_ANSI__.
And, `<stdlib.h>` will include `<alloca.h>` to be compatible with
software written for GNU and BSD. Thus, redefine `alloca` macro,
which was defined before at compat/regex/regex.c:66.
Considering this is only compat code, we've taken from other project,
it's not our business to decide which source should we adhere to.
Include `<stdlib.h>` early to prevent the redefinition of alloca.
This also remove a potential warning about alloca not defined on:
#undef alloca
Helped-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Đoàn Trần Công Danh <congdanhqx@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To help pinpoint the source of a regression, it is useful to know some
info about the compiler which the user's Git client was built with. By
adding a generic get_compiler_info() in 'compat/' we can choose which
relevant information to share per compiler; to get started, let's
demonstrate the version of glibc if the user built with 'gcc'.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Helped-by: Đoàn Trần Công Danh <congdanhqx@gmail.com>
Helped-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Starting in 3ac68a93fd, help.o began to depend on builtin/branch.o,
builtin/clean.o, and builtin/config.o. This meant that help.o was
unusable outside of the context of the main Git executable.
To make help.o usable by other commands again, move list_config_help()
into builtin/help.c (where it makes sense to assume other builtin libraries
are present).
When command-list.h is included but a member is not used, we start to
hear a compiler warning. Since the config list is generated in a fairly
different way than the command list, and since commands and config
options are semantically different, move the config list into its own
header and move the generator into its own script and build rule.
For reasons explained in 976aaedc (msvc: add a Makefile target to
pre-generate the Visual Studio solution, 2019-07-29), some build
artifacts we consider non-source files cannot be generated in the
Visual Studio environment, and we already have some Makefile tweaks
to help Visual Studio to use generated command-list.h header file.
Do the same to a new generated file, config-list.h, introduced by
this change.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
On certain network filesystems (currently encountered with Isilon, but
in theory more network storage solutions could be causing the same
issue), when the directory in question is missing,
`raceproof_create_file()` fails with an `ERROR_INVALID_PARAMETER`
instead of an `ERROR_PATH_NOT_FOUND`.
Since it is highly unlikely that we produce such an error by mistake
(the parameters we pass are fairly benign), we can be relatively certain
that the directory is missing in this instance. So let's just translate
that error automagically.
This fixes https://github.com/git-for-windows/git/issues/1345.
Signed-off-by: Nathan Sanders <spekbukkem@gmail.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
MSYS2's strace facility is very useful for debugging... With this patch,
the bash will be executed through strace if the environment variable
GIT_STRACE_COMMANDS is set, which comes in real handy when investigating
issues in the test suite.
Also support passing a path to a log file via GIT_STRACE_COMMANDS to
force Git to call strace.exe with the `-o <path>` argument, i.e. to log
into a file rather than print the log directly.
That comes in handy when the output would otherwise misinterpreted by a
calling process as part of Git's output.
Note: the values "1", "yes" or "true" are *not* specifying paths, but
tell Git to let strace.exe log directly to the console.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 4dc42c6c18 (mingw: refuse paths containing reserved names,
2019-12-21), we started disallowing file names that are reserved, e.g.
`NUL`, `CONOUT$`, etc.
This included `COM<n>` where `<n>` is a digit. Unfortunately, this
includes `COM0` but only `COM1`, ..., `COM9` are reserved, according to
the official documentation, `COM0` is mentioned in the "NT Namespaces"
section but it is explicitly _omitted_ from the list of reserved names:
https://docs.microsoft.com/en-us/windows/win32/fileio/naming-a-file#naming-conventions
Tests corroborate this: it is totally possible to write a file called
`com0.c` on Windows 10, but not `com1.c`.
So let's tighten the code to disallow only the reserved `COM<n>` file
names, but to allow `COM0` again.
This fixes https://github.com/git-for-windows/git/issues/2470.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Microsoft introduced a new "Universal C Runtime Library" (UCRT) with
Visual Studio 2015. The UCRT comes with a new strftime() implementation
that supports more date formats. We link git against the older
"Microsoft Visual C Runtime Library" (MSVCRT), so to use the UCRT
strftime() we need to load it from ucrtbase.dll using
DECLARE_PROC_ADDR()/INIT_PROC_ADDR().
Most supported Windows systems should have recieved the UCRT via Windows
update, but in some cases only MSVCRT might be available. In that case
we fall back to using that implementation.
With this change, it is possible to use e.g. the `%g` and `%V` date
format specifiers, e.g.
git show -s --format=%cd --date=format:‘%g.%V’ HEAD
Without this change, the user would see this error message on Windows:
fatal: invalid strftime format: '‘%g.%V’'
This fixes https://github.com/git-for-windows/git/issues/2495
Signed-off-by: Matthias Aßhauer <mha1993@live.de>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Cygwin, the codepath for POSIX-like systems is taken in
run-command.c::start_command(). The prepare_cmd() helper
function is called to decide if the command needs to be looked
up in the PATH. The logic there is to do the PATH-lookup if
and only if it does not have any slash '/' in it. If this test
passes we end up attempting to run the command by appending the
string after each colon-separated component of PATH.
The Cygwin environment supports both Windows and POSIX style
paths, so both forwardslahes '/' and back slashes '\' can be
used as directory separators for any external program the user
supplies.
Examples for path strings which are being incorrectly searched
for in the PATH instead of being executed as is:
- "C:\Program Files\some-program.exe"
- "a\b\c.exe"
To handle these, the PATH lookup detection logic in prepare_cmd()
is taught to know about this Cygwin quirk, by introducing
has_dir_sep(path) helper function to abstract away the difference
between true POSIX and Cygwin systems.
Signed-off-by: Andras Kucsma <r0maikx02b@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Explanation
-----------
The problem here is flawed `poll()` implementation. When it tries to
see if pipe can be written without blocking, it eventually calls
`NtQueryInformationFile()` and tests `WriteQuotaAvailable`. However,
the meaning of quota was misunderstood. The value of quota is reduced
when either some data was written to a pipe, *or* there is a pending
read on the pipe. Therefore, if there is a pending read of size >= than
the pipe's buffer size, poll() will think that pipe is not writable and
will hang forever, usually that means deadlocking both pipe users.
I have studied the problem and found that Windows pipes track two values:
`QuotaUsed` and `BytesInQueue`. The code in `poll()` apparently wants to
know `BytesInQueue` instead of quota. Unfortunately, `BytesInQueue` can
only be requested from read end of the pipe, while `poll()` receives
write end.
The git's implementation of `poll()` was copied from gnulib, which also
contains a flawed implementation up to today.
I also had a look at implementation in cygwin, which is also broken in a
subtle way. It uses this code in `pipe_data_available()`:
fpli.WriteQuotaAvailable = (fpli.OutboundQuota - fpli.ReadDataAvailable)
However, `ReadDataAvailable` always returns 0 for the write end of the pipe,
turning the code into an obfuscated version of returning pipe's total
buffer size, which I guess will in turn have `poll()` always say that pipe
is writable. The commit that introduced the code doesn't say anything about
this change, so it could be some debugging code that slipped in.
These are the typical sizes used in git:
0x2000 - default read size in `strbuf_read()`
0x1000 - default read size in CRT, used by `strbuf_getwholeline()`
0x2000 - pipe buffer size in compat\mingw.c
As a consequence, as soon as child process uses `strbuf_read()`,
`poll()` in parent process will hang forever, deadlocking both
processes.
This results in two observable behaviors:
1) If parent process begins sending STDIN quickly (and usually that's
the case), then first `poll()` will succeed and first block will go
through. MAX_IO_SIZE_DEFAULT is 8MB, so if STDIN exceeds 8MB, then
it will deadlock.
2) If parent process waits a little bit for any reason (including OS
scheduler) and child is first to issue `strbuf_read()`, then it will
deadlock immediately even on small STDINs.
The problem is illustrated by `git stash push`, which will currently
read the entire patch into memory and then send it to `git apply` via
STDIN. If patch exceeds 8MB, git hangs on Windows.
Possible solutions
------------------
1) Somehow obtain `BytesInQueue` instead of `QuotaUsed`
I did a pretty thorough search and didn't find any ways to obtain
the value from write end of the pipe.
2) Also give read end of the pipe to `poll()`
That can be done, but it will probably invite some dirty code,
because `poll()`
* can accept multiple pipes at once
* can accept things that are not pipes
* is expected to have a well known signature.
3) Make `poll()` always reply "writable" for write end of the pipe
Afterall it seems that cygwin (accidentally?) does that for years.
Also, it should be noted that `pump_io_round()` writes 8MB blocks,
completely ignoring the fact that pipe's buffer size is only 8KB,
which means that pipe gets clogged many times during that single
write. This may invite a deadlock, if child's STDERR/STDOUT gets
clogged while it's trying to deal with 8MB of STDIN. Such deadlocks
could be defeated with writing less than pipe's buffer size per
round, and always reading everything from STDOUT/STDERR before
starting next round. Therefore, making `poll()` always reply
"writable" shouldn't cause any new issues or block any future
solutions.
4) Increase the size of the pipe's buffer
The difference between `BytesInQueue` and `QuotaUsed` is the size
of pending reads. Therefore, if buffer is bigger than size of reads,
`poll()` won't hang so easily. However, I found that for example
`strbuf_read()` will get more and more hungry as it reads large inputs,
eventually surpassing any reasonable pipe buffer size.
Chosen solution
---------------
Make `poll()` always reply "writable" for write end of the pipe.
Hopefully one day someone will find a way to implement it properly.
Reproduction
------------
printf "%8388608s" X >large_file.txt
git stash push --include-untracked -- large_file.txt
I have decided not to include this as test to avoid slowing down the
test suite. I don't expect the specific problem to come back, and
chances are that `git stash push` will be reworked to avoid sending the
entire patch via STDIN.
Signed-off-by: Alexandr Miloslavskiy <alexandr.miloslavskiy@syntevo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We can check if certain characters are present in a string by calling
strchr(3) on each of them, or we can pass them all to a single
strpbrk(3) call. The latter is shorter, less repetitive and slightly
more efficient, so let's do that instead.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When debugging Git, the criss-cross spawning of processes can make
things quite a bit difficult, especially when a Unix shell script is
thrown in the mix that calls a `git.exe` that then segfaults.
To help debugging such things, we introduce the `open_in_gdb()` function
which can be called at a code location where the segfault happens (or as
close as one can get); This will open a new MinTTY window with a GDB
that already attached to the current process.
Inspired by Derrick Stolee.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Carlo Marcelo Arenas Belón