From patchwork Tue May 3 13:46:25 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Adhemerval Zanella Netto X-Patchwork-Id: 53420 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id C4D223856DC6 for ; Tue, 3 May 2022 13:46:54 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C4D223856DC6 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1651585614; bh=lR3w3sLS6AXpSdNy19kpyRulA+UGtE7f4rynWdLl8rE=; h=To:Subject:Date:List-Id:List-Unsubscribe:List-Archive:List-Post: List-Help:List-Subscribe:From:Reply-To:From; b=JbILHaAvCmWF4MCm2dQ7A8vUvWBTKKXlCQRCPhXU78duXw6KGnWqotYHeAha24AkS 10ctnl9ldntNj65ms4MLRWMfTNXCI6syTpKUwVOKgW9MmLlDc9UlBAFfjD6q1BvWhe 2C/pWC8VCQfoIVlJwTW9Qr98dckv5yU6XzD9kQLA= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mail-oi1-x234.google.com (mail-oi1-x234.google.com [IPv6:2607:f8b0:4864:20::234]) by sourceware.org (Postfix) with ESMTPS id 59860385736A for ; Tue, 3 May 2022 13:46:32 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 59860385736A Received: by mail-oi1-x234.google.com with SMTP id q8so17780979oif.13 for ; Tue, 03 May 2022 06:46:32 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:subject:date:message-id:mime-version :content-transfer-encoding; bh=lR3w3sLS6AXpSdNy19kpyRulA+UGtE7f4rynWdLl8rE=; b=EUGX146VW5HSt6EMkv1KdnLYAvKch0XuOFwvEDJYa+N9DDz55buTxqoeeheGaEFad+ rd3Imk4gMk9HpdFMO6FuGtM9R7cYsOo74KcnhiTU46K7Xm0bXHu+Kcandydb4AWgmear FPdJhZXU1FWdmAuy9U2baT0/EX6g1fW96Ejh6wFTtSCCfBHgB5LR4PNLDYVuXkvIeMwU 9W9gmsoqRaw8JEXIvhU0pMXwoV3rALrc+4Y5F5NryMSCKHR5qvKn9GOjRWN3Bt91szaK k4WVhr+XlV/DhM38SzmFc7dojM+8JNs84BYU1q46U5oAmIjMLsumh/Oentl4RjGnjfCJ ppeA== X-Gm-Message-State: AOAM532+cYbEgYWdn0ISSo3ZMZq1zQzhclPwnfJ2KufNy/dkxVkiMRWT 3jRKVUZOjVN3D1YuStjKFl0NaTlCce4Y8Q== X-Google-Smtp-Source: ABdhPJz/4EWw7mxY3jQDju+R1VkeY7z327/OpPwU4VHy9hRa1rZsciJuTrjSVNfwmy0Kmij2LbuPZw== X-Received: by 2002:a05:6808:1387:b0:325:837c:696d with SMTP id c7-20020a056808138700b00325837c696dmr1829294oiw.158.1651585591069; Tue, 03 May 2022 06:46:31 -0700 (PDT) Received: from birita.. ([2804:431:c7cb:726:bb0f:4f0c:a44b:4ae1]) by smtp.gmail.com with ESMTPSA id z10-20020a4ad58a000000b0035eb4e5a6c9sm4877124oos.31.2022.05.03.06.46.28 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 03 May 2022 06:46:30 -0700 (PDT) To: libc-alpha@sourceware.org, Alexey Izbyshev , Carlos O'Donell , Florian Weimer Subject: [PATCH v2] linux: Add fallback for clone failure on posix_spawn (BZ#29115) Date: Tue, 3 May 2022 10:46:25 -0300 Message-Id: <20220503134625.2389370-1-adhemerval.zanella@linaro.org> X-Mailer: git-send-email 2.34.1 MIME-Version: 1.0 X-Spam-Status: No, score=-12.9 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Adhemerval Zanella via Libc-alpha From: Adhemerval Zanella Netto Reply-To: Adhemerval Zanella Errors-To: libc-alpha-bounces+patchwork=sourceware.org@sourceware.org Sender: "Libc-alpha" Even though it might characterizeis as a kernel bug (incomplete kernel support for CLONE_VM used along with CLONE_VFORK), time namespace support is already presented in a some kernel releases (since 2020). Although I agree with Carlos that it results in unexpected surprise behaviour, I also see that providing a reasonable fallback that does not add possible issues like race conditions (as some syscall fallbacks in the past) provides a easier transition and make the interfaca to work on multiple scenarios. The fix itself does adds some corner cases that need to be handled, but I still think it is maintanable and a general improvement. --- Linux clone with CLONE_VM may fail for some namespace restriction (for instance if kernel does not allow processes in different time namespaces to share the sameaddress space). In this case clone fails with EINVAL and posix_spawn can not spawn a new process. However the same process can be spawned with fork and exec. The patch fixes by retrying the clone syscall with just CLONE_VFORK if clone fails with a non transient failure (ENOMEM and EAGAIN still returns failure to caller). Error on preparation phase or execve is returned by a pipe now that there is no shared memory that ca be used. It requires some extra care for some file operations: * If the file operation would clobber the pipe file descriptor, the helper process dup the pipe onto an unoccupied file descriptor. * dup2 file action that targets the pipe file descriptor returns EBADF. * If closefrom argument overlaps the pipe file descriptor, it is splited in two calls: [lowdp, pipe - 1] and [pipe + 1, ~Ou]. Failure on prepare phase in helper process does not trigger the fork and exec fallback. Checked on x86_64-linux-gnu. --- sysdeps/unix/sysv/linux/spawni.c | 231 +++++++++++++++++++++++++------ 1 file changed, 185 insertions(+), 46 deletions(-) diff --git a/sysdeps/unix/sysv/linux/spawni.c b/sysdeps/unix/sysv/linux/spawni.c index d6f5ca89cd..e17b1967ef 100644 --- a/sysdeps/unix/sysv/linux/spawni.c +++ b/sysdeps/unix/sysv/linux/spawni.c @@ -44,7 +44,11 @@ third issue is done by a stack allocation in parent, and by using a field in struct spawn_args where the child can write an error code. CLONE_VFORK ensures that the parent does not run until the - child has either exec'ed successfully or exited. */ + child has either exec'ed successfully or exited. + + If the clone with CLONE_VM and CLONE_VFORK fails (due any kernel limitation + such as time namespace), only CLONE_VFORK is used instead and the + preparation and execve failures are communicated with a pipe. */ /* The Unix standard contains a long explanation of the way to signal @@ -67,6 +71,7 @@ struct posix_spawn_args char *const *envp; int xflags; int err; + int pipe[2]; }; /* Older version requires that shell script without shebang definition @@ -94,15 +99,59 @@ maybe_script_execute (struct posix_spawn_args *args) } } +/* If the file operation would clobber the pipe fd used to communite with + parent, dup the pipe onto an unoccupied file descriptor. */ +static inline bool +spawni_fa_handle_pipe (const struct __spawn_action *fa, int p[]) +{ + int fd; + + switch (fa->tag) + { + case spawn_do_close: + fd = fa->action.close_action.fd; + break; + case spawn_do_open: + fd = fa->action.open_action.fd; + break; + case spawn_do_dup2: + fd = fa->action.open_action.fd; + break; + case spawn_do_fchdir: + fd = fa->action.fchdir_action.fd; + default: + return true; + } + + if (fd == p[1]) + { + int r = __fcntl (p[1], F_DUPFD_CLOEXEC); + if (r < 0) + return false; + __close_nocancel (p[1]); + p[1] = r; + } + + return true; +} + +static inline bool +spawni_fa_closerange (int from, int to) +{ + int r = INLINE_SYSCALL_CALL (close_range, from, to, 0); + return r == 0 || __closefrom_fallback (from, false); +} + /* Function used in the clone call to setup the signals mask, posix_spawn attributes, and file actions. It run on its own stack (provided by the posix_spawn call). */ -static int -__spawni_child (void *arguments) +static _Noreturn int +spawni_child (void *arguments) { struct posix_spawn_args *args = arguments; const posix_spawnattr_t *restrict attr = args->attr; const posix_spawn_file_actions_t *file_actions = args->fa; + bool use_pipe = args->pipe[0] != -1; /* The child must ensure that no signal handler are enabled because it shared memory with parent, so the signal disposition must be either SIG_DFL or @@ -113,6 +162,9 @@ __spawni_child (void *arguments) struct sigaction sa; memset (&sa, '\0', sizeof (sa)); + if (use_pipe) + __close (args->pipe[0]); + sigset_t hset; __sigprocmask (SIG_BLOCK, 0, &hset); for (int sig = 1; sig < _NSIG; ++sig) @@ -181,6 +233,9 @@ __spawni_child (void *arguments) { struct __spawn_action *action = &file_actions->__actions[cnt]; + if (use_pipe && !spawni_fa_handle_pipe (action, args->pipe)) + goto fail; + switch (action->tag) { case spawn_do_close: @@ -233,6 +288,11 @@ __spawni_child (void *arguments) break; case spawn_do_dup2: + if (use_pipe && action->action.dup2_action.fd == args->pipe[1]) + { + errno = EBADF; + goto fail; + } /* Austin Group issue #411 requires adddup2 action with source and destination being equal to remove close-on-exec flag. */ if (action->action.dup2_action.fd @@ -264,8 +324,16 @@ __spawni_child (void *arguments) case spawn_do_closefrom: { int lowfd = action->action.closefrom_action.from; - int r = INLINE_SYSCALL_CALL (close_range, lowfd, ~0U, 0); - if (r != 0 && !__closefrom_fallback (lowfd, false)) + /* Skip the pipe descriptor if it is used. No need to handle + it since it created and duplicated with O_CLOEXEC. */ + if (use_pipe + && args->pipe[1] > action->action.closefrom_action.from) + { + if (!spawni_fa_closerange (lowfd, args->pipe[1] - 1) + || !spawni_fa_closerange (args->pipe[1] + 1, ~0U)) + goto fail; + } + else if (!spawni_fa_closerange (lowfd, ~0U)) goto fail; } break; @@ -300,10 +368,114 @@ fail: (EINTERNALBUG) describing that, use ECHILD. Another option would be to set args->err to some negative sentinel and have the parent abort(), but that seems needlessly harsh. */ - args->err = errno ? : ECHILD; + int ret = errno ? : ECHILD; + if (use_pipe) + { + while (__write_nocancel (args->pipe[1], &ret, sizeof (ret)) < 0) + if (errno == EPIPE || errno == EBADF) + break; + } + else + args->err = ret; + _exit (SPAWN_ERROR); } +static pid_t +clone_call (struct posix_spawn_args *args, int flags, void *stack, + size_t stack_size) +{ + struct clone_args clone_args = + { + .flags = flags, + .exit_signal = SIGCHLD, + .stack = (uintptr_t) stack, + .stack_size = stack_size, + }; + return __clone_internal (&clone_args, spawni_child, args); +} + +/* Spawn a new process using clone with CLONE_VM | CLONE_VFORK (to optimize + memory and overcommit) and return TRUE if the helper was created or if the + failure was not due resource exhaustion. */ +static bool +spawni_clone (struct posix_spawn_args *args, pid_t *new_pid, int *ec, + void *stack, size_t stack_size) +{ + /* The clone flags used will create a new child that will run in the same + memory space (CLONE_VM) and the execution of calling thread will be + suspend until the child calls execve or _exit. + + Also since the calling thread execution will be suspend, there is not + need for CLONE_SETTLS. Although parent and child share the same TLS + namespace, there will be no concurrent access for TLS variables (errno + for instance). */ + *new_pid = clone_call (args, CLONE_VM | CLONE_VFORK, stack, stack_size); + + /* It needs to collect the case where the auxiliary process was created + but failed to execute the file (due either any preparation step or + for execve itself). */ + if (*new_pid > 0) + { + /* Also, it handles the unlikely case where the auxiliary process was + terminated before calling execve as if it was successfully. The + args.err is set to 0 as default and changed to a positive value + only in case of failure, so in case of premature termination + due a signal args.err will remain zeroed and it will be up to + caller to actually collect it. */ + *ec = args->err; + if (*ec > 0) + /* There still an unlikely case where the child is cancelled after + setting args.err, due to a positive error value. Also there is + possible pid reuse race (where the kernel allocated the same pid + to an unrelated process). Unfortunately due synchronization + issues where the kernel might not have the process collected + the waitpid below can not use WNOHANG. */ + __waitpid (*new_pid, NULL, 0); + } + else + *ec = errno; + + /* There is no much point in retrying with fork and exec if kernel returns a + failure due resource exhaustion. */ + return *new_pid > 0 || (errno == ENOMEM || errno == EAGAIN); +} + +/* Fallback spawn case which does not use CLONE_VM. Any prepration step or + execve failure is passed with a pipe, which requires additional care by + the helper stating process since it additional file descriptors handle. */ +static void +spawni_vfork (struct posix_spawn_args *args, pid_t *new_pid, int *ec, + char *stack, size_t stack_size) +{ + if (__pipe2 (args->pipe, O_CLOEXEC) != 0) + { + *ec = errno; + return; + } + + /* Do not trigger atfork handler nor any internal state reset since the + helper process will call execve. */ + *new_pid = clone_call (args, CLONE_VFORK, stack, stack_size); + if (*new_pid == 0) + spawni_child (args); + + __close (args->pipe[1]); + + if (*new_pid > 0) + { + if (__read (args->pipe[0], ec, sizeof *ec) != sizeof *ec) + /* A successful execve will close the helper process pipe end. */ + *ec = 0; + else + __waitpid (*new_pid, NULL, 0); + } + else + *ec = errno; + + __close (args->pipe[0]); +} + /* Spawn a new process executing PATH with the attributes describes in *ATTRP. Before running the process perform the actions described in FILE-ACTIONS. */ static int @@ -367,49 +539,16 @@ __spawnix (pid_t * pid, const char *file, args.argc = argc; args.envp = envp; args.xflags = xflags; + args.pipe[0] = args.pipe[1] = -1; __libc_signal_block_all (&args.oldmask); - /* The clone flags used will create a new child that will run in the same - memory space (CLONE_VM) and the execution of calling thread will be - suspend until the child calls execve or _exit. - - Also since the calling thread execution will be suspend, there is not - need for CLONE_SETTLS. Although parent and child share the same TLS - namespace, there will be no concurrent access for TLS variables (errno - for instance). */ - struct clone_args clone_args = - { - .flags = CLONE_VM | CLONE_VFORK, - .exit_signal = SIGCHLD, - .stack = (uintptr_t) stack, - .stack_size = stack_size, - }; - new_pid = __clone_internal (&clone_args, __spawni_child, &args); - - /* It needs to collect the case where the auxiliary process was created - but failed to execute the file (due either any preparation step or - for execve itself). */ - if (new_pid > 0) - { - /* Also, it handles the unlikely case where the auxiliary process was - terminated before calling execve as if it was successfully. The - args.err is set to 0 as default and changed to a positive value - only in case of failure, so in case of premature termination - due a signal args.err will remain zeroed and it will be up to - caller to actually collect it. */ - ec = args.err; - if (ec > 0) - /* There still an unlikely case where the child is cancelled after - setting args.err, due to a positive error value. Also there is - possible pid reuse race (where the kernel allocated the same pid - to an unrelated process). Unfortunately due synchronization - issues where the kernel might not have the process collected - the waitpid below can not use WNOHANG. */ - __waitpid (new_pid, NULL, 0); - } - else - ec = errno; + /* clone with CLONE_VM | CLONE_VFORK may fail for some namespace restriction + (for instance Linux does not allow processes in different time namespaces + to share address space) and in this case clone fails with EINVAL. Retry + with fork and exec. */ + if (!spawni_clone (&args, &new_pid, &ec, stack, stack_size)) + spawni_vfork (&args, &new_pid, &ec, stack, stack_size); __munmap (stack, stack_size);