From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/10957 Path: news.gmane.org!.POSTED!not-for-mail From: Eric Hassold Newsgroups: gmane.linux.lib.musl.general Subject: Re: Fix pthread_create on some devices failing to initialize guard area Date: Fri, 20 Jan 2017 13:04:28 -0800 Message-ID: <30588c41-eb3d-627e-c5eb-91e19ef56790@gmail.com> References: <20170120195649.GS1533@brightrain.aerifal.cx> Reply-To: musl@lists.openwall.com NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit X-Trace: blaine.gmane.org 1484946305 5323 195.159.176.226 (20 Jan 2017 21:05:05 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Fri, 20 Jan 2017 21:05:05 +0000 (UTC) User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:45.0) Gecko/20100101 Thunderbird/45.6.0 To: musl@lists.openwall.com Original-X-From: musl-return-10972-gllmg-musl=m.gmane.org@lists.openwall.com Fri Jan 20 22:05:01 2017 Return-path: Envelope-to: gllmg-musl@m.gmane.org Original-Received: from mother.openwall.net ([195.42.179.200]) by blaine.gmane.org with smtp (Exim 4.84_2) (envelope-from ) id 1cUgMS-00080I-He for gllmg-musl@m.gmane.org; Fri, 20 Jan 2017 22:04:40 +0100 Original-Received: (qmail 15710 invoked by uid 550); 20 Jan 2017 21:04:43 -0000 Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-ID: Original-Received: (qmail 15689 invoked from network); 20 Jan 2017 21:04:42 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=subject:to:references:from:message-id:date:user-agent:mime-version :in-reply-to:content-transfer-encoding; bh=GHnmBASubVDeZdH5h5VyCsXkq8oxO3MJsjqUEKyJlvA=; b=VfYWVCO1fJL/POZ387ROWgDe6GHCsHuodI5jFokKVcCYSiRzR9hKyERNeECDxliqjH a6wmAUcSkGP67kba+6GKMoR5dBVhvreY3JehvMK2f4xqYlJLALIOBxWydc5aK18CXC+P fak4pfzTprFwgcPy+tv5+DxDRTi3Rp6bplNUZx6wVGKOPBhLCTa6NYVvV1tpn4pertws AaCtisjt8M/RZEzsM1NdOcuH2rbWazitYgeotYOYXqSh59bIxPnfOjJsVEz0OAald/Ow 9ijW5k6u8l05yMJxRAOLhesS6g7b4I02hYAwUniCy5veqkqpwaa+ybQTSvGy6xjUuaN9 niQw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:subject:to:references:from:message-id:date :user-agent:mime-version:in-reply-to:content-transfer-encoding; bh=GHnmBASubVDeZdH5h5VyCsXkq8oxO3MJsjqUEKyJlvA=; b=JNgvum1E9+jSB9PU2eyXoMJKCaC26s0ZINu8zplUb2RK/SIw50uAIABk5kD7iVdJ3E gUGOEGexVFlMVjSMYqZPTToR0zfR8UVxpsrl5XZBqfL9d5RbZh2f+lXCKMMZgyYEcPhh yaf7w+Cu5QI/xGbWfk+pOijn+kIVStGuLq54mjPKnjv7brXS3QaJ/YnJ4BXh5t2wxzGx 3j5xqwJhLVo/wUvkDLzyELgpkIQNaNMgWTivA1DB2Gl4XTKgZfhH7or5V/kAFwcaLTwv nN5Pczf2VE8XOe4DfMLD/MWQW/8gWAXXa7JIQROmhxl5dRAvxGiN7BlyIlnv6M23zGgS YbTA== X-Gm-Message-State: AIkVDXJXxkRzcA2rnH8JG/gNcfOWXDJQvzTp0eVuLaJ2JOLo36PmxuYjqRe1wDbW2W/gfg== X-Received: by 10.84.216.91 with SMTP id f27mr24506478plj.92.1484946270483; Fri, 20 Jan 2017 13:04:30 -0800 (PST) In-Reply-To: <20170120195649.GS1533@brightrain.aerifal.cx> Xref: news.gmane.org gmane.linux.lib.musl.general:10957 Archived-At: On 1/20/17 11:56 AM, Rich Felker wrote: > On Fri, Jan 20, 2017 at 11:45:09AM -0800, Eric Hassold wrote: >> Hi All, >> >> While deploying test static executable across farm of different >> embedded systems, found out that pthread_create() is failing >> systematically on some (very few) arm-linux devices whenever non >> null stack guard is enabled (that is, also when calling >> pthread_create with default - i.e. null - attributes since default >> is a one page of guard). One of those device is for example a >> Marvell Armada 375 running Linux 3.10.39. Same test code, built with >> alternative libc implementations (glibc, uClibc) works as expected >> on those devices. >> >> >> Issue >> >> This occurs because of call to mprotect() in pthread_create fails. >> In current implementation, if guard size is non null, memory for >> (guard + stack + ...) is first allocated (mmap'ed) with no >> accessibility (PROT_NONE), then mprotect() is called to re-enable >> read/write access to (memory + guardsize). Since call to mprotect() >> systematically fails in this scenario (returning error code EINVAL), >> it is impossible to create thread. > Failure is ignored and the memory is assumed to be writable in this > case, since EINVAL is assumed to imply no MMU. Is this assumption > wrong in your case, and if so, can you explain why? In my case, devices exhibiting issue are not MMU-less, they are Cortex-A9 devices with valid mmu / page protection working as expected otherwise. Note that current Musl code assumes ENOSYS means no MMU and handles it by assuming the system has no page protection at all. For the case I observe, it is EINVAL which is returned, this is not ignored, so memory is unmap'ed and pthread_create() fails. > >> Patch >> >> In proposed patch (attached below), memory for (guard + stack + ...) >> is first mmap'ed with read/write accessibility, then guard area is >> protected by calling mprotect() with PROT_NONE on guardsize first >> bytes of returned memory. This call to mprotect() to remove all >> accessibility on guard area, with guard area being at beginning of >> previously mmap'ed memory, works correctly on those platforms having >> issue with current implementation. Incidentally, this makes the >> logic more concise to handle both cases (with or without guard) is a >> more consistent way, and handle systems with partial/invalid page >> protection implementation (e.g. mprotect() returning ENOSYS) more >> gracefully since the stack is explicitly created with read/write >> access. > This doesn't work correctly on normal systems with mmu, because the > size of the guard pages is accounted against commit charge. Linux > should, but AFAIK doesn't, subtract it from commit charge once it's > changed to PROT_NONE without having been dirtied, but even if this bug > is fixed on the kernel side, there would still be a moment where > excess commit charge is consumed and thus where pthread_create might > spuriously fail or cause allocations in other processes/threads to > fail. > > If the kernel is not allocating actually-usable address ranges for > PROT_NONE on all nommu systems, I think the only solution is to handle > EINVAL from mprotect by going back and re-doing the mmap with > PROT_READ|PROT_WRITE. Do you have any better ideas? > > Rich Had this "deja vu" feeling... reminds me conversation you had in this thread some time ago elsewhere... https://sourceware.org/ml/libc-alpha/2015-09/msg00447.html Your proposition seems reasonable on nommu system, but again, the issue observed here is on legit systems with mmu, with mprotect failing with EINVAL (and not ENOSYS), for some other reason than system not supporting page protection. Catching EINVAL error returned by mprotect and falling back to re-doing the mmap would mean actually silently running without stack guard on system supporting it, so I believe it is actually legitimate to fail and return error in that case. But that's difference use case than the issue I'm observing. I took note of Balazs's suggestion (in the thread referenced above) to switch to a pattern similar to Musl's current one (mmap(PROT_NONE) + mprotect(stack, READ|WRITE)) in order to avoid those guard pages to actually occupy resources. But I can indeed observe that this approach fails on some devices (which have valid mmu), while I'm not sure I'm seeing the issue with first mapping PROT_READ|PROT_WRITE then mprotect(PROT_NONE) guard area. Latter approach (as implemented by patch) is, at least, consistent with all the other implementations out there (I checked glibc's allocatestack.c, but also e.g. bionic), and couldn't find report of those failures you are envisaging. Eric