From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp1.atof.net (smtp1.atof.net [52.86.233.228]) by sourceware.org (Postfix) with ESMTPS id C11283858D37 for ; Thu, 20 Apr 2023 18:46:24 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org C11283858D37 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=gluelogic.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=gluelogic.com X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-Spam-Language: en X-Spam-Relay-Country: X-Spam-DCC: B=MGTINTERNET; R=smtp1.atof.net 1170; Body=1 Fuz1=1 Fuz2=1 X-Spam-RBL: X-Spam-PYZOR: Reported 0 times. Date: Thu, 20 Apr 2023 14:46:21 -0400 From: gs-cygwin.com@gluelogic.com To: Bruno Haible , cygwin@cygwin.com Subject: Re: posix_spawn facility Message-ID: References: <1752276.7aRn1RRit1@nimes> <4892432.0VBMTVartN@nimes> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: X-Spam-Status: No, score=-2.1 required=5.0 tests=BAYES_00,KAM_DMARC_STATUS,SPF_HELO_NONE,SPF_PASS,TXREP,T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 List-Id: On Thu, Apr 20, 2023 at 05:40:52PM +0200, Corinna Vinschen via Cygwin wrote: > On Apr 20 16:58, Bruno Haible via Cygwin wrote: > > Corinna Vinschen wrote: > > > > Hmm. Your code uses lpReserved2 for that, but the functionality is > > > > one implemented in MSVCRT. For obvious reasons, Cygwin executables > > > > are not linked against msvcrt.dll and we're using lpReserved2 for our > > > > own purposes. > > > > > > Oh, btw., did you know that there's a newer mechanism for defining > > > specific inheritable handles to CreateProcess, which is implemented > > > in kernel32.dll, so it does not depend on MSVCRT? > > > > > > There's a STARTUPINFOEX structure which allows to specify the > > > additional handles. See > > > > > > https://learn.microsoft.com/en-us/windows/win32/api/winbase/ns-winbase-startupinfoexa > > > > > > and the PROC_THREAD_ATTRIBUTE_HANDLE_LIST argument described in > > > > > > https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-updateprocthreadattribute > > > > Indeed, this appears to be a more "official" way to pass handles for fd ≥ 3, > > instead of lpReserved2 — albeit without associated 'flags'. Not sure how > > O_APPEND is handled then... > > Yeah, theoretically, that should be handled by CreateFile opening the > file with FILE_APPEND_DATA attribute, and in the child MSVCRT should > test with NtQueryInformationFile(FILE_ACCESS_INFORMATION) if the > FILE_APPEND_DATA flag is set. > > But then again, if MSVCRT implements fcntl (F_SETFL) to allow > manipulating the O_APPEND flag... unfortunately there's no such > operation via Win32 or native calls. That would require to reopen the > file with different access mask and replace the HANDLE under the hood of > the descriptor. I'm not aware if and how MSVCRT performs this action. If you are carefully controlling and allowing an explicit set of file handles to be inherited, and the entire program uses this interface to create new processes, then you can safely _sopen_s() or otherwise to create new handles, pass them to CreateProcess() using STARTUPINFOEX, and then close any new handles created solely for inheritance in child. Cheers, Glenn