aboutsummaryrefslogtreecommitdiff
path: root/win32/process.c (follow)
Commit message (Collapse)AuthorAgeFilesLines
* win32: another attempt at a UCRT hackucrt_hackRon Yorston2021-11-121-8/+0
| | | | | | | | | | | | | | | | | The problem with UCRT seems to be that if a process has been started with a non-NULL environment block passed to CreateProcess() any subsequent call to spawnve() with a non-NULL environment pointer fails. Commit 5b48ca53b (win32: pass NULL to spawnve, not environ) fixed the problem in busybox-w32 for those cases where a NULL environment pointer was sufficient. It didn't handle the case where the shell passes a modified environment to its child. All calls to spawnve() in the shell occur in a process which will terminate whether or not the call succeeds. It therefore doesn't matter if we mess with the environment of this process such that spawnve() can be passed a NULL environment pointer. (I think.)
* win32: move UCRT hack to process.cRon Yorston2021-11-041-0/+8
| | | | | | | | Only apply the UCRT hack if we're about to call spawnve() with a non-NULL environment pointer. Add the pid of the process to the name of the environment variable to provide greater assurance that the environment will be changed.
* win32: rename mingw_spawn_1Ron Yorston2021-11-011-8/+8
| | | | | | | | | | The name of the function mingw_spawn_1() wasn't particularly meaningful. Remove its envp argument (as all of its current callers pass a NULL pointer) and rename it mingw_spawnvp() to better reflect what it does. The path search it performs isn't the standard one: it has features specific to busybox-w32.
* win32: pass NULL to spawnve, not environRon Yorston2021-10-281-4/+4
| | | | | | | | | | | | Building busybox-w32 for use with UCRT results in mysterious failures. (GitHub issue #234) These are somehow related to the environment values passed to spawnve. In several places the global environ pointer was being passed to spawnve so the child would inherit its environment from the parent. This can also be achieved by passing a NULL pointer. This prevents the failures in at least some cases and also makes the binary smaller.
* win32: handle -1 return status from execve(2)Ron Yorston2021-06-191-2/+3
| | | | | | | | The implementations of execve(2) and execvp(3) were unable to distinguish between failure and a program returning a status of -1. Check the error return code to discriminate between these cases. See GitHub issue #218.
* ash: code shrinkRon Yorston2021-02-141-9/+0
| | | | | Since there's only one call to mingw_spawn_forkshell() we might as well just call spawnve() directly from ash.
* win32: code shrink Unix-style path handlingRon Yorston2020-08-231-19/+9
| | | | | | | | Replace auto_add_system_drive() with alloc_system_drive() which leaves space for a possible filename extension. This makes it possible to drop alloc_win32_extension() and auto_win32_extension(). Saves 144 bytes.
* win32: use built-in applets for non-existent binaries with Unix-style pathsRon Yorston2020-08-131-2/+13
| | | | | | | | | | | | | | | | | | | | Shell scripts moved from Unix may contain hard-coded paths to binaries such as /bin/sh. A recent commit made it possible to execute such binaries reliably, but that does require them to be installed. As an alternative solution: if a binary with a standard Unix path prefix can't be found but is available as a built-in applet, run the applet. Add the function unix_path() to detect paths starting with /bin, /usr/bin, /sbin or /usr/sbin. Use this function in: - the 'which' applet - shellexec(), describe_command() and find_command() in ash - mingw_spawn_1() See GitHub issue #195.
* win32: handle Unix-style absolute paths for executablesRon Yorston2020-08-131-1/+10
| | | | | | | | | | | | | | | | | | | | | | As noted in commit 548ec7045 (win32: interpret absolute paths as relative to %SYSTEMDRIVE%) a path starting with a '/' in the Unix world is treated as relative to the current drive by Windows. To avoid ambiguity that commit considered certain such paths to be relative to %SYSTEMDRIVE%. Extend this to paths representing executables. Add the functions need_system_drive() and auto_add_system_drive() to detect the need for a system drive prefix and to add it if necessary. Use these functions in: - the 'which' applet - the find_executable() function - tab-completion code - PATH look-up, shellexec(), describe_command() and find_command() in ash - parse_interpreter() and mingw_spawn_1() With these changes executable paths starting with a slash are handled consistently, whatever the current drive.
* win32: code shrinkRon Yorston2020-08-131-1/+1
| | | | | | | | | | Add a new function, has_path(), to detect that an executable name doesn't require a path look-up. Also, since is_absolute_path() is now only used in shell/ash.c move its definition there from include/mingw.h. Saves 128 bytes.
* win32: code shrink kill(2)Ron Yorston2020-07-251-31/+17
| | | | | | | | - Drop exit_code argument from kill_SIGTERM_by_handle() - Pass signal number rather than exit code to other functions - Merge kill_SIGKILL() and kill_SIGTEST() Saves 112 bytes.
* win32: update link to Microsoft document on command-line argumentsRon Yorston2020-07-211-1/+1
|
* win32: minor improvements to process spawningRon Yorston2020-06-121-3/+2
| | | | | | | | | | | | | | The function mingw_spawn_forkshell() was introduced to handle spawning a forkshell process. Since we know that: - the binary being executed doesn't meet any of the special cases handled by spawnveq() - the arguments don't require quoting we can call spawnve() directly instead of spawnveq(). Also, use xzalloc() to allocate new_argv so we don't need to set the final NULL pointer explicitly.
* time: detect when spawn failsRon Yorston2020-06-041-1/+1
| | | | | | | In the 64-bit build the time applet reported garbage when it was unable to run the program to be timed. The error return from mingw_spawn_pid() was incorrect.
* win32: use lazy loading for certain DLLsRon Yorston2020-06-021-2/+13
| | | | | | | | | Only a handful of functions are used from shell32.dll, userenv.dll and psapi.dll. Mostly these functions are in out of the way places. By loading the functions only when required we can avoid the startup cost of linking the three DLLs in the common case that they aren't needed.
* ash: simplify spawning during forkshellRon Yorston2020-06-011-0/+9
| | | | | | | | | | | | | spawn_forkshell() uses mingw_spawn_proc() to start the child shell. mingw_spawn_proc() then calls mingw_spawn_1() which determines that "sh" is an applet, thus calling mingw_spawn_applet() which finally calls spawnveq(). Not only is this convoluted it also won't work if PREFER_APPLETS and SH_STANDALONE aren't enabled. Simplify matters by adding a new function, mingw_spawn_forkshell(), which is tailored for just this case.
* win32: changes to '#!' supportRon Yorston2020-05-241-10/+12
| | | | | | | | | | | | | Since the earliest days of busybox-w32 '#!' has searched PATH for the interpreter. This doesn't seem to be supported by precedent. In testing I also found that additional code was needed for the case where PATH has been altered in the current shell or is modified on the command line: PATH has to be extracted from envp rather than getenv("PATH"). Drop this non-standard feature. *NIX implementations disagree on whether the interpreter can itself be a script. Follow Linux and allow this with a limit of four levels of nesting.
* time: WIN32 portRon Yorston2020-04-061-7/+34
| | | | | | | | Port the time applet to WIN32. This requires the implemntation of a replacement for wait3(2). Only elapsed, user and system times are supported, not the memory and i/o statistics reported by GNU time.
* win32: new functions: getpeername(2), mingw_spawn_detach()Ron Yorston2020-04-051-3/+15
| | | | | | | Implement getpeername(2). Add mingw_spawn_detach() to allow the spawned process to detach from the console.
* win32: more improvements to 'c:file' pathsRon Yorston2020-03-291-2/+3
| | | | | | | | Tab-completion of 'c:file' paths was leaking memory and incorrectly included BusyBox applets and shell built-ins. Fix handling of 'c:file' paths used as arguments to spawn functions. Otherwise things like 'timeout 5 c:busybox sleep 99' don't work.
* win32: trim leading and trailing spaces from shebang optionsRon Yorston2019-05-031-0/+4
| | | | | | | | | | | | Commit 97e2c4a05 (win32: changes to treatment of scripts) attempted to use strtok(3) to simplify the parsing of shebang lines. Unfortunately it resulted in leading and trailing whitespace being left in the option string. Fix this by trimming the options before they're returned. Reported-by: Niklas DAHLQUIST <niklas.dahlquist@st.com> Signed-off-by: Ron Yorston <rmy@pobox.com>
* win32: add function to convert slashes to backslashesRon Yorston2019-03-151-5/+2
| | | | | | | | | There are now two places where slashes are converted to backslashes throughout a string so it makes sense to create a function to do this. To avoid confusion rename convert_slashes() to bs_to_slash() and call the new function slash_to_bs().
* win32: drop argument from err_win_to_posix()Ron Yorston2019-03-061-2/+2
|
* win32: make more use of common lazy loading codeRon Yorston2019-02-271-17/+5
|
* win32: don't try to run non-existent scriptRon Yorston2019-02-071-1/+1
| | | | | | In parse_interpreter return an error when the file doesn't exist. The current code tries to run non-existent scripts with a .sh suffix, thus breaking the test ash-misc/exec.tests.
* ps: fix long output when desktop compatibility is disabledRon Yorston2019-01-081-0/+3
| | | | | | | | In desktop mode the state column is disabled and doesn't affect output. When desktop mode is disabled the state column is displayed in long output. Put some fake data in the column to improve the display.
* ps: display interpreted script name in comm columnRon Yorston2019-01-081-1/+2
| | | | | | | | | | | When an interpreted script is being run the comm column in ps should display the name of the script not the name of the interpreter. Use a fake applet pathname to indicate which argument contains the script. This also allows pidof to obtain the pid of a script.
* ps: add support for the args columnRon Yorston2019-01-081-11/+29
| | | | | | | | | Implement read_cmdline() for WIN32 by storing the command line in the same way as the applet name. The applet name is actually used for the comm column which is truncated to COMM_LEN. Using this as the size of the bb_comm array avoids the need to calculate MAX_APPLET_NAME_LEN.
* ps: add user and group columnsRon Yorston2019-01-081-1/+10
|
* win32: special treatment for PATHRon Yorston2018-12-141-20/+0
| | | | | | | | | | | | | | | | | The PATH shell variable is a special case. It can be exported to the environment where it might be interpreted by native applications which assume the separator is ';'. Hence: - require that the separator used in PATH is ';' - enforce this by intercepting calls to setvareq() that set PATH and adjusting its value if necessary. As a result of this the code to parse PATH can be simplified by replacing the hardcoded Unix ':' path separator by the platform- dependent macro PATH_SEP. The MANPATH variable is also required to use ';' as its separator but since it's less likely to be used this isn't enforced.
* win32: special treatment for files with the '.sh' extensionRon Yorston2018-12-091-27/+36
| | | | | | | | | | | | Files with the extension '.sh' are considered to be executable. Those that start with a '#!' line will be run using the specified interpreter. If no '#!' is present the script will be run by the shell. When searching for an executable the '.sh' extension will be tested in the same way as the standard extensions understood by spawnve(). '.sh' takes precedence over the standard extensions.
* win32: use open_read_close() where possibleRon Yorston2018-12-091-6/+2
|
* win32: allow execution of empty batch file to succeedRon Yorston2018-12-081-0/+8
| | | | | | | Microsoft Windows' spawnve returns ERROR_BAD_EXE_FORMAT when passed an empty batch file. Work around this by skipping spawnve and returning success.
* win32: improve execution of batch filesRon Yorston2018-12-071-26/+8
| | | | | | | | | It appears that when a batch file is executed the first argument must contain backslashes if it's a relative path. Absolute paths work either way. In both cases the extension is optional. This allows for a considerable simplification of the special case in spawnveq.
* win32: rework adding of extensions to filenamesRon Yorston2018-12-061-2/+2
| | | | | | | | | | | | | | | | | | | | | | | Previously there was one function to handle adding extensions to executable filenames, add_win32_extension(). Refactor this into three functions: add_win32_extension() appends the suffix to the argument string in-place. The argument must be long enough to cope with this, as is the case in ash where path_advance() adds 4 bytes to each filename for just this reason. alloc_win32_extension() is equivalent to the old add_win32_extension(). It allocates a string to hold the new filename then calls the new add_win32_extension() function. The caller is responsible for managing the returned string. auto_win32_extension() calls alloc_win32_extension() and saves the resulting string using auto_string(). It's used where the new filename is consumed immediately or the actual value isn't needed. Rewrite code to use the most appropriate function. Also reorder some code in find_executable() and find_command().
* ash: fix a couple of test casesRon Yorston2018-12-041-3/+1
| | | | | | | | | | | | | | The ash tests exitcode_EACCES and exitcode_ENOENT both failed. In commit 92dbd3c09 a test was added to tryexec to check that the file being run existed and was executable. The error codes returned by this test were incorrect. The slightly later commit f5783ef14 added a similar test in spawnveq which got the error codes right. Remove the test from tryexec and some superfluous error messages from spawnveq.
* win32: save a few bytes in process scanningRon Yorston2018-11-271-14/+7
| | | | | | Recalculate the system boot time on every pass through the process scanning loop. It's less efficient than storing the value in a static variable but not noticeably so and it saves a few bytes.
* win32: use string_array_len to obtain size of argv arrayRon Yorston2018-10-261-8/+4
| | | | Saves 16 bytes.
* win32: better error message in kill(2)Ron Yorston2018-07-121-0/+2
| | | | | | When the process architecture of busybox-w32 didn't match that of the target process the error reported was 'Function not implemented'. Change this to 'Permission denied'.
* win32: remove some code from procps_scanRon Yorston2018-04-101-3/+0
| | | | | Now that we're calling memset to clear data for each process it's no longer necessary to zero the process times by hand.
* ps: indicate forkshell processes in listingRon Yorston2018-04-051-1/+1
|
* ps: obtain applet names from other BusyBox processesRon Yorston2018-04-041-18/+89
| | | | | | | | | | | | | | | Remove the code which passed applet names to child processes using environment variables. This only allowed ps to display names for its ancestors. Instead attempt to read applet names from the memory of unrelated processes. The Microsoft documentation alone wasn't enough to figure out how to do this. Additional hints from: https://stackoverflow.com/questions/4298331/exe-or-dll-image-base-address https://stackoverflow.com/questions/14467229/get-base-address-of-process
* win32: return correct exit status from waitpidRon Yorston2018-03-291-1/+5
| | | | | | | | | | | The exit status from _cwait wasn't being correctly returned. This resulted, for example, in the exit status of xargs being incorrect. Running this: $ ls | xargs ls in a directory containing filenames with spaces should cause 'ls' to fail and 'xargs' to return an exit status of 123.
* win32: save a few bytesRon Yorston2018-03-291-4/+3
| | | | | | | | In the recently-added code to pass applet names to child processes use local arrays to build the environment variables rather that allocating them every time. mingw_spawn can call mingw_spawn_proc instead of mingw_spawn_1.
* ps: display applet names in process listingRon Yorston2018-03-281-1/+16
| | | | | | | | | | | | | | | | In standalone shell mode all busybox-w32 applets are displayed as 'busybox.exe' in a process listing. I haven't found a satisfactory way to query a running instance of busybox-w32 to determine which applet it's running. Handle a couple of cases: - the process running the process scan knows its own PID and knows which applet it is; - just before invoking applet_main set an environment variable whose name contains the PID and whose value is the current applet name. Children running a process scan will inherit their parent's environment and can therefore match the parent's PID to its applet name.
* ps: clear status for each processRon Yorston2018-03-281-0/+2
| | | | | | | | | The original procps_scan function takes care to clear the status information before handling each process. Do the same for the WIN32 version. This requires moving the snapshot handle to the part of the structure that isn't cleared on each iteration.
* win32: changes to kill(2)Ron Yorston2018-03-271-14/+20
| | | | | | | | | The names of the callbacks to kill processes with a given signal were confusing because they referred to the implementation rather than the intent. Create the new function kill_SIGTERM_by_handle which is more efficient when a handle to the process to be killed is already available.
* win32: small changes to reduce size of binaryRon Yorston2018-03-221-2/+2
| | | | | | | | Reduce the size of the binary by about 32 bytes: - use xzalloc to allocate static buffers so we don't have to initialise them; - avoid duplicated code in spawnveq.
* win32: reduce amount of static dataRon Yorston2018-03-041-28/+34
| | | | | | | It appears that uninitialised static variables are placed in the data section rather than bss, increasing the size of the binary. Rewrite some code to reduce the amount of static data.
* win32: additional improvements to handling of executablesRon Yorston2018-02-281-10/+24
| | | | | | | | | | | | | Consistent processing of file extensions, as described in the previous commit, has been applied to the 'which' applet and the functions find_executable and mingw_spawn_interpreter. In spawnveq check that the file to be executed exists and is executable, and ensure that it won't have any extensions added by spawnve. It's intended that all files passed to spawnve should have their names fully specified. If this isn't the case the tests here will cause errors which will need to be fixed.
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-21 13:12:26 +0000