December 2006 - Darwin-dev - lists.apple.com

Re: rsync ACLs and unix permissions (with/without rsrc)
by Giuliano Gavazzi 29 Dec '06

29 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 25 Dec 2006, at 02:51, Corey O'Connor wrote: Giuliano _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I've never had any luck with the rsync that comes with 10.4. I use the rsync build from here: http://www.onthenet.com.au/~q/rsync/ Which has worked fine in my limited testing. You might have better luck with it. -Corey Thanks Corey, I will have a look at that build sooner or later. For the moment I found a solution, with a small change to rsync.c in rsync-24: diff rsync.c.orig rsync.c 197a198,209 > if (!strncmp("._", file->basename, 2)) { > STRUCT_STAT st; > char buf[MAXPATHLEN]; > pathjoin(buf, sizeof buf, file->dirname, (file- >basename)+2); > if (stat(buf, &st) == 0) { > if (do_chmod(fname,(st.st_mode & CHMOD_BITS)) != 0) { > rsyserr(FERROR, errno, "failed to set permissions on %s", > full_fname(fname)); > return 0; > } > } > } this just works by calling stat on the already transferred data fork and then applying its permissions to the temporary copy of the ACL ._ file. A kludge, but it seems to work. I wonder though were it is written that ._* file are not transferred by rsync, at least in Apple's version. This email sent to site_archiver(a)lists.apple.com

1 0

Re: poll() Bug in 10.4 / Darwin 8.0 ?
by Morgan Aldridge 27 Dec '06

27 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thanks in advance, Morgan Aldridge -- morgant(a)makkintosshu.com http://www.makkintosshu.com/ On May 19, 2005, at 8:22 AM, Dave Zarzycki wrote: davez On May 19, 2005, at 8:09 AM, Steven Palm wrote: I end up here by tracing with gdb: The hang is here: do { switch (poll(descriptors, 1, 0)) { case -1: case 0: i = 0; break; default: i = sizeof(extraData) - extraDataLength; if (i == 0) { *buffer = '\000'; return(buffer); } else if ((count = read(0, extraData + extraDataLength, i)) > 0) { extraDataLength+= count; } else i = 0; } } while (i); It appears to hang on the read call. info locals shows: descriptors = {{ fd = 0, events = 1, revents = 32 }} ptr = 0xbffff424 "" i = 1023 j = -1073744832 count = 1 state = 0 The author of wy60 responded with: if (!(descriptors->revents & POLLIN)) i = 0; else before the line that reads if (i == 0) _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Are there still poll/select issues with Darwin 8.x? I've seen references as recently as mid-October to poll() issues when working with any BPF devices (including ttys/ptys). Bug number 3710161 at Apple is tracking this. In short, kqueue and poll on anything in /dev does not behave correctly. I was trying to use a program I've used for a long time in 10.2 and 10.3, a terminal emulator "shim" called "wy60". I haven't tried to wrap this in any sort of simple test case, nor have I filed this as a bug. I thought I'd make sure it was worthy of being filed and not a programming error, although the code does work on various unixes as-is. Also, if this is terribly off-base for this group, I apologize, but it's decidedly at a Darwin level, no GUI (except possibly curses and terminal stuff) involved here. Here are some notes (and workaround) I had posted to the wy60 group, with the author's response: In trying wy60 on MacOS X 10.4 (Tiger), I'm having an odd blocking issue that does not happen with 10.2 or 10.3. #0 0x90013784 in read () #1 0x00007d44 in readResponse (timeout=500, query=0x15a74 "\033 [0c", buffer=0xbffff424 "", firstChar=27, lastChar=99, discard=0, maxLength=80) at wy60.c:3127 #2 0x00010458 in initTerminal (pty=-1) at wy60.c:4919 #3 0x000134e0 in main (argc=1, argv=0xbffff628) at wy60.c:6123 Now, why this is an issue in 10.4 and was not in any earlier version, I do not know. I first attempted to run the binary from an earlier system without luck, and then compiled a fresh copy on the 10.4 machine and it has the same problem. Well, the fix seems easy enough, but the underlying answer not so easy.... I simply edited my config.h to remove HAVE_POLL_H so it wouldn't use the system poll call, and it now works using select and/or poll emulation. My thinking was that poll() was telling the system it had data to read when it really didn't have anything there. I see the author of the tun/tap kext for MacOS X also claims the poll() call is broken for character devices in Darwin8. For some reason, it seems to think that stdin is not a valid file descriptor. Not quite sure why that would happen. But in any case, can you try adding This should make the code stop trying to read from stdin, if poll returned with anything other than POLLIN. Hopefully, if I understand your problem description correctly, this will fix the bug that you are seeing. Although, you might now run the risk of occasionally seeing garbage characters when you press function keys. If that happens, we'll need to take a closer look at what the system calls are actually returning in all of these cases. This email sent to site_archiver(a)lists.apple.com

1 0

Re: statfs and 64bit file lengths?
by Michael Smith 26 Dec '06

26 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Dec 2, 2006, at 12:02 PM, darwin-dev-request(a)lists.apple.com wrote: On Dec 1, 2006, at 2:37 PM, jmzorko(a)mac.com wrote: How do I determine the amount of free disk on a volume on OSX 10.4.x? The *nix statfs() API seems limited to 32bit file lengths, unless i'm misunderstanding something. Can someone send me a quick example of how to do this? >From xnu/bsd/vfs/vfs_syscalls.c:munge_statfs(): This fixup is applied both for statfs(2) and getfsstat(2). = Mike _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… The important detail you're missing here is that statfs(2) will lie about the block size in order to make the answer fit in 31 bits. /* * It's possible for there to be more than 2^^31 blocks in the filesystem, so we * have to fudge the numbers here in that case. We inflate the blocksize in order * to reflect the filesystem size as best we can. */ ... /* * Work out how far we have to shift the block count down to make it fit. * Note that it's possible to have to shift so far that the resulting * blocksize would be unreportably large. At that point, we will clip * any values that don't fit. * * For safety's sake, we also ensure that f_iosize is never reported as * being smaller than f_bsize. */ This email sent to site_archiver(a)lists.apple.com

1 0

Re: darwin equivalent to linux readahead(2) function
by Michael Smith 26 Dec '06

26 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Dec 18, 2006, at 12:02 PM, darwin-dev-request(a)lists.apple.com wrote: Hello list, I was wondering if there exists a darwin equivalent to Linux readahead(2) function. <http://www.die.net/doc/linux/man/man2/readahead.2.html> This is used to "populates the page cache with data from a file so that subsequent reads from that file will not block on disk I/O". I've been googling for readahead darwin, readahead bsd etc. but found nothing relevant :-( Any pointer to a function or even more complicated method to achieve this goal is welcome. See fcntl(2): struct radvisory { off_t ra_offset; /* offset into the file */ int ra_count; /* size of the read */ }; = Mike _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… F_RDADVISE Issue an advisory read async with no copy to user. ... The F_RDADVISE command operates on the following structure which holds information passed from the user to the system: This email sent to site_archiver(a)lists.apple.com

1 0

problem with IOGraphics-191.5/tools/setdparam.c
by Dirk Schelfhout 26 Dec '06

26 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: rsync ACLs and unix permissions (with/without rsrc)
by Corey O'Connor 25 Dec '06

25 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=qIGFBGDtAJYPhEEVlQlzlpgIL8cSHrZkD3sKnyYQ8/WE7zGxICv8vIx9ImfA5nW3ZN71zJRkWg7IDnNnXB2hiIVK9rxlwRKplWR6SgdX32GSh0XSrzlMSBODaE4sO0bT+8vn3rcIu9ucspAOi5KpsmyWvvPU1uhU3wwEmY6FLq4= I've never had any luck with the rsync that comes with 10.4. I use the rsync build from here: http://www.onthenet.com.au/~q/rsync/ Which has worked fine in my limited testing. You might have better luck with it. -Corey On 24 Dec 2006, at 01:41, Giuliano Gavazzi wrote: > a file with an ACL, copied with rsync local to local (from one > volume to another), does not get its unix permissions copied. For > instance a 755 file is copied with 600 permissions. sorry, the permission of the copied file are 400. This bug seems related to the permissions assigned to the extended attributes (the ._filename). Indeed, in another test (the source file is a_file): [sender] i=0 /var/tmp/testsource <NULL> . mode=040755 len=102 uid=0 gid=0 [sender] i=1 /var/tmp/testsource <NULL> ._a_file mode=0100400 len=1 uid=0 gid=0 [sender] i=2 /var/tmp/testsource <NULL> a_file mode=0100644 len=10 uid=0 gid=0 presumably the problem is in the order they are necessarily transferred; from rsync/generator.c: /* APPLE: extended attribute files (._foo) need to be transferred * after the corresponding file (foo). This creates a map the size * of flist with the number of the file that preempts the current * file's delivery. Set to -1 if there's nothing to do. */ g _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/coreyoconnor%40gmail.com This email sent to coreyoconnor(a)gmail.com -- -Corey O'Connor _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… On 12/24/06, Giuliano Gavazzi <dev+lists(a)humph.com> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Re: rsync ACLs and unix permissions (with/without rsrc)
by Giuliano Gavazzi 24 Dec '06

24 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com g _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… On 24 Dec 2006, at 01:41, Giuliano Gavazzi wrote: a file with an ACL, copied with rsync local to local (from one volume to another), does not get its unix permissions copied. For instance a 755 file is copied with 600 permissions. sorry, the permission of the copied file are 400. This bug seems related to the permissions assigned to the extended attributes (the ._filename). Indeed, in another test (the source file is a_file): [sender] i=0 /var/tmp/testsource <NULL> . mode=040755 len=102 uid=0 gid=0 [sender] i=1 /var/tmp/testsource <NULL> ._a_file mode=0100400 len=1 uid=0 gid=0 [sender] i=2 /var/tmp/testsource <NULL> a_file mode=0100644 len=10 uid=0 gid=0 presumably the problem is in the order they are necessarily transferred; from rsync/generator.c: /* APPLE: extended attribute files (._foo) need to be transferred * after the corresponding file (foo). This creates a map the size * of flist with the number of the file that preempts the current * file's delivery. Set to -1 if there's nothing to do. */ This email sent to site_archiver(a)lists.apple.com

1 0

re: my previous message [Help with debugging a trace]
by Vince Loden 24 Dec '06

24 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thanks, Vince _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Does anyone know about the asl.log? fsck reported it as damaged. Something to do with my previous report on crashes. This email sent to site_archiver(a)lists.apple.com

1 0

rsync ACLs and unix permissions (with/without rsrc)
by Giuliano Gavazzi 24 Dec '06

24 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com rsync2.6.3 from 10.4.8: Giuliano _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… a file with an ACL, copied with rsync local to local (from one volume to another), does not get its unix permissions copied. For instance a 755 file is copied with 600 permissions. Conversely rsync 2.6.9, that does not support resource forks yet, does behave properly. I have not tested with different ACLs, but it would be strange if the result was dependent on their content. If this issue has not been resolved in recent developments I will file a bug. [Since it is known that rsync 2.6.3 crashes when both ACLs and rsrc exist for a file, at the current state of affairs I would say that it is better to ignore resource forks and use rsync 2.6.9...] This email sent to site_archiver(a)lists.apple.com

1 0

Re: Help with debugging a trace
by Vince Loden 23 Dec '06

23 Dec '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 23 Dec 2006, at 3:18 PM, Paul Ripke wrote: Thanks for replying, Paul. It panics with or without the 200 GB attached. Yes I'd thought maybe a FireWire problem too. I just had a quick look at the source - that panic looks like it is there to protect against a condition that shouldn't happen... I'm beat - but running fsck over both disks wouldn't hurt. On 23 Dec 2006, at 2:47 PM, Paul Ripke wrote: Without studying the info provided too deeply: Does it only panic when the 200 GB drive is mounted? I'd guess that the HFS+ (I assume) metadata structures are corrupted on your 200 GB drive - umount and run fsck (Darwin method) or Disk Utility (Mac OS X) over the filesystem. Second look at the panic string and backtrace, I'm guessing the later. Cheers, -- Paul Ripke Unix/OpenVMS/TSM/DBA I love deadlines. I like the whooshing sound they make as they fly by. -- Douglas Adams _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I've used fsck, applejack, onyx countless times. I'd really like to re-install without reformatting though. On Sat, Dec 23, 2006 at 02:52:16PM +0800, Vince Loden wrote: Thanks. I hope I've given you enough info. On Sat, Dec 23, 2006 at 01:45:23PM +0800, Vince Loden wrote: panic(cpu 0 caller 0x000E3DDC): vnode_writedone: numoutput < 0 Either that, or there's unrelocatable bad blocks in the HFS+ metadata structures that's causing problems. I haven't checked the source - what does Darwin do on metadata write failures? Panic? This email sent to site_archiver(a)lists.apple.com

1 0