Date: Thu, 4 Oct 2012 18:49:10 GMT From: John Baldwin <jhb@FreeBSD.org> To: freebsd-gnats-submit@FreeBSD.org Subject: ports/172332: [exp-run] Expanding stdio's internal file descriptors from short to int Message-ID: <201210041849.q94InA3H025428@red.freebsd.org> Resent-Message-ID: <201210041850.q94IoBai002875@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 172332 >Category: ports >Synopsis: [exp-run] Expanding stdio's internal file descriptors from short to int >Confidential: no >Severity: non-critical >Priority: low >Responsible: freebsd-ports-bugs >State: open >Quarter: >Keywords: >Date-Required: >Class: change-request >Submitter-Id: current-users >Arrival-Date: Thu Oct 04 18:50:10 UTC 2012 >Closed-Date: >Last-Modified: >Originator: John Baldwin >Release: HEAD >Organization: >Environment: >Description: On Friday, September 28, 2012 6:47:39 pm John Baldwin wrote: > Four years or so ago I cleaned up some of the stdio internals as fallout from > running into problems with stdio using a short instead of an int to hold file > descriptors. Back then I got sidetracked with attempting to make FILE opaque > and ended up never getting around to bumping _file from a short to an int. I > recently ran back into the SHRT_MAX limit at work again and came up with a > patch to fix this. > > To preserve the ABI, it is necessary to leave the existing short _file in > place and add a new int _file to the end of the FILE structure. Also, for old > applications, the old _file (_ofile in the patch) must still be valid. The > approach I have taken is to bump the symbol version for routines that create > FILE objects with a non-fake _file (fopen, fdopen, and freopen). The old > FBSD_1.0 variants still fail if an fd is greater than SHRT_MAX (and thus > cannot be safely stored in _ofile). The new FBSD_1.3 variants assign to both > _file and _ofile if the fd is less than SHRT_MAX. I also changed fileno() > to no longer be an inline macro in <stdio.h> but to always be a function call > going forward. > > If folks think this is ok, I'll hack up a modified version that hides _file > from outside consumers (rename it to _nfile or some such) and send it for a > ports-exp run before committing to make sure there aren't any 3rd party apps > accessing _file directly. I have a slightly modified version of my original patch that should hide _file completely from any package builds. No ports should be directly accessing the internals of FILE. They should be using things like fileno() instead. Can you do an exp-run with a patched world to see if there are any such abusers? The patch is MI, so I think just doing one architecture should be sufficient. http://www.FreeBSD.org/~jhb/patches/stdio_file_exp.patch >How-To-Repeat: >Fix: >Release-Note: >Audit-Trail: >Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201210041849.q94InA3H025428>