From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=209.85.221.49 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1795183979; b=qweFaWuMK+7kxmqP9gQLzl2+1VAuYPHtNv9kohF/J+rlPGihsOTRhsLnOUYB/dNcxoXK QqMdOeVVi4CGL3rs4KpoQ1FfLF9puUDv0UAs1HXv0XS0uXaAvTenPVKozJJhYux5iRx0K J6giFVCOLNkGrwwaXwkCQBd9EH31wFLXlNeiR+KtDZ3fsZFrk/U5uqeA3pbH0Zes7VVmh bASYHNpPgvNwm/Mz3Xe4w/Zsu/KsmuUzEH3vqqbSKrnbv+yWcE126j0VhMHEnR9f/KgVH 837pwHNAxtebX4VYHPWHmnZlFqyPdN+U+ibejmPahAQW8B6kQDNG+1sluf95Z4uOaAw== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1795183979; h=DMARC-Filter:Received:DKIM-Signature:X-Google-DKIM-Signature: X-Forwarded-Encrypted:X-Gm-Message-State:X-Gm-Gg:X-Received:Received: Mime-Version:Content-Type:Date:Message-Id:Cc:Subject:From:To:X-Mailer: References:In-Reply-To; bh=ob1FlA4nHTyzVacAeo+PvFthXWjQDamhK5368LJKCsA=; b=s08vlPZQ1NReQn3l6wWQmYt7YIkgFfssVlvvccfDDBGeAEc/GWCrUimejLKlLjRfUcMq yWoOslDVW+7n7fWxFD0ee7RLFyHJWz36o1EVKp0IO8lb8yGNMyyQN209jQKsnhxKKTaKz Zyy/p+TyL7SyvnMbIOpCEAgDt73nArApZNjfFPukkWBToAJMmhCtZVhSB0+vSUdEfbrqm GCTJ7YO2xurooDElFrmA++Eku8mAXWTAclG14GPvmXAf/j+IJqkNu4X6BJVOlPUzD7x6X 1A0gMfRg4MJihqFbEZrTcYGhrSqykFFBdeUCuzpnQ7XowQwoOv4RNv4f+qW64GVOtHQ== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20260707 header.b=rbdeZn9I; arc=none smtp.remote-ip=209.85.221.49 Received: from mail-wr1-f49.google.com (mail-wr1-f49.google.com [209.85.221.49]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 2BF4660A2603 for ; Fri, 20 Nov 2026 14:12:59 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org 2BF4660A2603 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.a=rsa-sha256 header.s=20260707 header.b=rbdeZn9I Received: by mail-wr1-f49.google.com with SMTP id ffacd0b85a97d-48e20ed3b9dso806896f8f.2 for ; Fri, 20 Nov 2026 06:12:59 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20260707; t=1795183978; x=1795788778; darn=vger.kernel.org; h=in-reply-to:references:to:from:subject:cc:message-id:date :content-type:mime-version:from:to:cc:subject:date:message-id :reply-to:content-type; bh=ob1FlA4nHTyzVacAeo+PvFthXWjQDamhK5368LJKCsA=; b=rbdeZn9ILcLAYFNo6TnnJ/A/KWVu0+5+Msk52h1KKBNcteqyesyrPXI/akOkn7Y5nZ IJyrZg0HBWpxlikd55uohYATUxLodDwD7+b0Es+Sv1nPKJ9z0TnzqWj2CjA739oyp/6S YFXswRBrBr/Roowpp01nTbbkrenfbJmQnMPa7Wvw4iNxateK7/eiauenrJfEXM37DDZh JQQGJufT9v88mGvYant1EA3p1ELjXQH8dS6GrkmHhnTOb/00rX33bFSULeUgKsQShHVD y459Q8ZJz1MEcjiZvRKUq6FM655W5Trml3NMjWgRKIK/DU3HXLHP9kP7MoIvCkq7hyDx oZfQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20260707; t=1795183978; x=1795788778; h=in-reply-to:references:to:from:subject:cc:message-id:date :content-type:mime-version:x-gm-gg:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to:content-type; bh=ob1FlA4nHTyzVacAeo+PvFthXWjQDamhK5368LJKCsA=; b=Qp5Da0cqwTexIdPyrcKLBGoUUnlMEQBZg2EYhxoOlUuLkmVRgzS1+ZTNvBBZA3KOKZ 8V0UO7ER7VaiVs4NePeDLJWIX5DXmVv96Gd8uzAfMv2DznYPRC620VaHI3YJxNgu7zEw Lf5qILiiCaNJY62/IRgkCXq5G/d9oWB653TkbZ+YvR0ZfmXm21VV5DKtrJE2lImN7a5N a2ubkbt9/F+8CBVoGfHwSDnu6nV15ymo7szRCfsO51PLuvj4qrwca/qAD6yui7pp86Vd fS4ANXVmvu4PvSL3iZHG2BxoBLIssVkAUM+sKnpSehyBHC4ytMjX+SGBRA/TkogXVuyf Cibg== X-Forwarded-Encrypted: i=1; AL+hErHXUJEljD8ffJBp3gQZddZYn5ikRtgYmCwaDgYPywpIC68bz9PqGeQevtUMtq6/gDxY23TRglEfXHDd7Hmb@vger.kernel.org X-Gm-Message-State: APuU4Z+d0kMFyOxVpeGjQ1odeP0Hs4Ox0MRD4zvMci12f5E/otbnUQ81 Blb+4pwfWOSzIeoXI/U+nLykyxPkjEDFXiWzDPQzK/rSSHOjD68Tz9NUwxr59+1B X-Gm-Gg: AYBFou1PEm2HUwJ500QMf4x3Nmdx1mahg+w+f56pbNOqZYQBdPoX80dt9N8zjgGKJct XvnKoZt5VAg6Yw/8IO+Dhury2sJsXt3D8SX7DAjkDFf3w3C3HW2QQBPTUdMTncpTQblFq0nNKIr 6MUE8omCaLhgnvQIQ70JiNnrDClSDihJgZ7gAHaBvCizaXkdg6yNzFmTlJUh01LNzt3n0houX10 sShoeLcqTMnHozXp9YMJ5DNXXz5gPnHfu/dM3lAkcmiOUBcgmESSHM8mu5ARLd/Z6N8yR4SviMA 5i1nrlLSFVHz9e9BrOfTnlcM8CI641UwLbfBwaDz4GAZ31srH0Lu1tiSytGfy2SAQJOk3QTm1F5 buJSoZezWZl4Uw8LrsK1+TCFKU0sM+PEy6uNJCoLetXTTI2xl48NZAN8zmeAbnx3hh8xh3tSoli FneL7Trrkv6y4Se79JM/uM64IL+jGuTt0ScuPTnixt46JJ89HnFxRiEET+4p+vYyVwsnfEGE79k /+SBR4Qt2BHQCQ8ZN6IouS02ZJaPeY6S0zSEzT3QBEMZ2dBRGzUbOptC8E6wImJ8lsrvJlxMPrC hU2ay9Tub9o6NhH4Vor/S0Jb+l5grEO7n7dLBPB4Rz6KGcb8 X-Received: by 2002:a05:600c:3d05:b0:4b8:d17e:da27 with SMTP id 5b1f17b1804b1-4bb88930872mr78210085e9.27.1795183977527; Fri, 20 Nov 2026 06:12:57 -0800 (PST) Received: from localhost ([130.192.98.149]) by smtp.gmail.com with ESMTPSA id 5b1f17b1804b1-4bb8538b09bsm189744955e9.4.2026.11.20.06.12.54 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Fri, 20 Nov 2026 06:12:55 -0800 (PST) X-Mailing-List: linux-fsdevel@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list Mime-Version: 1.0 Content-Type: multipart/signed; boundary=1c0f90a8e81abc35d40b9da9afa2a275d20c96cec9fed1a0bbc0e0120237; micalg=pgp-sha512; protocol="application/pgp-signature" Date: Fri, 20 Nov 2026 15:12:54 +0100 Message-Id: Cc: "Lauri Gates" , "Dave Chinner" , "Christian Brauner" , "Christoph Hellwig" , , "Carlos Maiolino" , "Christian Brauner" , "Darrick J. Wong" , , , , , , , "Theodore Tso" Subject: Re: Concurrent FIEMAP on reflinked files serialises on the AGF buffer lock From: "koraynilay" To: "Amir Goldstein" , "koraynilay" X-Mailer: aerc 0.22.0-0-gc2f86b7abde3 References: <20261119095943.85048-1-lauri.gates@gmail.com> In-Reply-To: --1c0f90a8e81abc35d40b9da9afa2a275d20c96cec9fed1a0bbc0e0120237 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=UTF-8 On Fri Nov 20, 2026 at 6:52 AM CET, Amir Goldstein wrote: > On Thu, Nov 19, 2026 at 11:21=E2=80=AFPM koraynilay = wrote: >> >> On Thu Nov 19, 2026 at 6:04 PM CET, Amir Goldstein wrote: >> > On Thu, Nov 19, 2026 at 12:00=E2=80=AFPM koraynilay wrote: >> >> >> >> On Thu Nov 19, 2026 at 10:59 AM CET, Lauri Gates wrote: >> >> > On Wed, Nov 18, 2026 at 10:16PM +0200, Dave Chinner wrote: >> >> >> I really don't care if it is exposed by statx or file_attr; checki= ng >> >> >> the flag through either interface will be orders of magnitude fast= er >> >> >> than using FIEMAP to iterate extents on every file to find shared >> >> >> extents. My concern with statx is coherency - it runs completely >> >> >> unlocked and so can race with operations that add/remove the refli= nk >> >> >> flag... >> >> > >> >> > Thanks, Dave and Christian. Given Christian's preference and the >> >> > coherency point, I'll write the RFC against file_attr, once the >> >> > question at the end is settled. xfs_fileattr_get() already reads th= e >> >> > flags under XFS_ILOCK_SHARED, so the new flag gets that without ext= ra >> >> > work. The answer can still go stale right after the call, as with a= ny >> >> > stat result: a clone that lands after the read is missed, just as o= ne >> >> > that lands after FIEMAP returns is missed today. >> >> > >> >> > As I suggested on the 6th, the flag would take the inverted meaning= , >> >> > so that an unset flag stays the safe answer: say FS_XFLAG_NOSHARED >> >> > ("no extents shared with another file"), read-only, in >> >> > FS_XFLAG_RDONLY_MASK next to the case-folding flags. >> >> >> >> Hello, I don't have any say in the FIEMAP issue, but I wanted to let = you >> >> know that, since also a mask was mentioned and your patch would be >> >> touching the fileattr API, there is work going[1] on for adding a >> >> fsx_xflags_mask that filesystem use[2][3] to advertise supported flag= s (on >> >> get, like stx_attributes_mask) and users use to say what they want to >> >> change (on set). >> >> It doesn't seem like an FS_XFLAG_NOSHARED flag would touch this new >> >> change, but I thought it was worth pointing out, especially because t= his >> >> patch would also add FS_XFLAG_PROJID and a new FS_FILE_ATTR_XFLAGS_MA= SK >> >> to have a way to include both legacy and newly added xflags, so >> >> FS_XFLAG_NOSHARED would maybe also go there (correct me if I'm wrong)= . >> >> >> >> > Christian, one question on the shape within file_attr: would you ta= ke >> >> > the inverted flag, or would you rather keep a plain "may share" fla= g >> >> > and add a mask to struct file_attr that says which xflags the >> >> > filesystem reports? >> >> >> >> And if Christian chooses the mask approach, then my patch[4] becomes >> >> relevant. >> >> (in [4] I put the patch that will be sent in v5 of [1], since in [1] >> >> it's merged with the ptime change) >> >> >> >> (For Amir who I just CC'd, [5] is the main link for this thread) >> > >> > Thanks for the CC and the link >> > >> >> >> >> Thanks, >> >> koraynilay >> >> >> >> [1]: https://lore.kernel.org/all/20261115180343.576773-1-koray.fra@gm= ail.com/T/#u >> >> [2]: https://lore.kernel.org/all/CAOQ4uxj82K4z+2OiY5W1g7NyWAFOFjqqWnZ= SS3prytW36j1LgA@mail.gmail.com/ >> >> [3]: https://lore.kernel.org/all/CAOQ4uxiuJ=3DeZu28FZAUz3rpqPKqCimPxJ= ucOnYYuBowvD09_Cw@mail.gmail.com/ >> >> [5]: https://lore.kernel.org/linux-xfs/20261102193533.93690-1-lauri.g= ates@gmail.com/T/#u >> >> >> >> [4]: >> >> commit 0efe0de3b5cc2b767aa6e2bb9174eb05b0d818f1 >> >> Author: koraynilay >> >> Date: Tue Nov 17 14:57:47 2026 +0100 >> >> >> >> vfs: add xflags mask to file_attr API >> >> >> >> Add fa_xflags_mask to struct file_attr and equivalent fsx_xflags_= mask >> >> to struct file_kattr. >> >> >> >> This lets filesystems advertise their support for specific xflags= and >> >> users specify only what they want to change, behaving like >> >> statx.stx_attributes_mask: >> >> - on GET (file_getattr): each filesystem fills the mask according= to >> >> its supported xflags, e.g. FS_XFLAG_COWEXTSIZE or FS_XFLAG_EXTS= IZE, >> >> and provides the values in the corresponding struct fields; >> >> - on SET (file_setattr): userspace fills the mask according to wh= ich >> >> xflags it wants to change, VFS validates this and preserves xfl= ags >> >> that were not specified in the mask; >> >> - if xflags_mask =3D=3D 0, use legacy full-struct replacement to = keep >> >> compatibility with older programs. >> >> >> >> Furthermore, add the missing FS_XFLAG_PROJID flag, to allow users= pace to >> >> set the projid using file_[sg]etattr(). >> > >> > Correction. usersapce can already set projid using file_[sg]etattr(). >> > With FS_XFLAG_PROJID in fsx_xflags_mask, chattr -p >> > could set the projid without having to read and set all the other xfla= gs. >> >> Oh right, thanks, I'll fix it. >> >> >> >> >> Finally, in order to prevent the legacy FS_IOC_FSSETXATTR from ac= cepting >> >> FS_XFLAG_PROJID, add an FS_FILE_ATTR_XFLAGS_MASK to support new x= flags >> >> and use it in fileattr_to_file_attr()/file_attr_to_fileattr() to >> >> validate that the xflags passed in fa_xflags and fa_xflags_mask f= ields >> >> are valid. >> >> >> >> Assisted-by: Gemini:3.8-flash-high antigravity-cli-1.5.7 >> >> Signed-off-by: koraynilay >> >> --- >> >> fs/file_attr.c | 46 ++++++++++++++++++++++++++++++++++++++= ++++---- >> >> include/linux/fileattr.h | 5 +++++ >> >> include/uapi/linux/fs.h | 3 +++ >> >> 3 files changed, 50 insertions(+), 4 deletions(-) >> >> >> >> diff --git a/fs/file_attr.c b/fs/file_attr.c >> >> index 833e03810b24..56ea65f790f1 100644 >> >> --- a/fs/file_attr.c >> >> +++ b/fs/file_attr.c >> >> @@ -102,10 +102,18 @@ EXPORT_SYMBOL(vfs_fileattr_get); >> >> static void fileattr_to_file_attr(const struct file_kattr *fa, >> >> struct file_attr *fattr) >> >> { >> >> - __u32 mask =3D FS_XFLAGS_MASK; >> >> + /* >> >> + * fsx_xflags_mask contains only flags added after the >> >> + * file_[gs]etattr() API, so we need to tell userspace the fi= lesystem >> >> + * actually supports all legacy (FS_XFLAGS_MASK and FS_XFLAG_= PROJID, >> >> + * which lacked its own bit) and new advertised flags. >> >> + */ >> >> + __u32 mask =3D (FS_XFLAGS_MASK | FS_XFLAG_PROJID | fa->fsx_xf= lags_mask) & >> >> + FS_FILE_ATTR_XFLAGS_MASK; >> > >> > This is not exactly how I envisioned exporting fa_xflags_mask. >> >> This is exactly why I CC'd you :D >> >> > fa_xflags_mask =3D=3D 0 means old kernel or unmodified fs. >> > >> > For a new kernel with fs that supports flags other than >> > FS_XFLAGS_MASK it's the fs responsibility to advertise its supported >> > flags. >> > >> > Something like: >> > >> > __u32 mask =3D (fa->fsx_xflags_mask ?: FS_XFLAGS_MASK) & >> > FS_FILE_ATTR_XFLAGS_MASK; >> > >> > memset(fattr, 0, sizeof(struct file_attr)); >> (I'm sorry, but why memset if we assign the fields right after?) >> > fattr->fa_xflags =3D fa->fsx_xflags & mask; >> > fattr->fa_xflags_mask =3D fa->fsx_xflags_mask ? mask : 0; >> > >> > Then you can add filesystems that support get/set of new >> > attributes without support for legacy attributes. >> > >> > And specific filesytems could opt-into the set of an individual >> > attribute like projid by advertising >> > fa->fsx_xflags_mask =3D FS_XFLAGS_MASK | FS_XFLAG_PROJID >> > or a subset thereof. >> >> I agree, this is definitely better. >> >> > And I am still on the fence wrt should API also advertise the >> > fsx_xflags_ro_mask in addition to the supported mask. >> > >> > The advantage being that it makes the interface symmetric. >> > User can alway setattr() with the buffer they got from getattr() >> > and instead of randomly ignoring flags in set, the fsx_ro_mask >> > along with fsx_mask defines the behavior exactly. >> >> Hmm, so ro_mask would be ignored by vfs on set so users don't have to >> manually remove the read only flags, correct? >> > > Correct. Ok. > With non-zero xflags_mask, internal contact masks > can be used for -EINVAL not for silent ignoring. What do you mean by "internal contact masks"? >> Also I think I'll try writing some basic userspace programs that use >> this API and see how I'd expect it to be. >> > > Don't write programs please contribute to chattr (e2fsprogs) and > xfs_io (xfsprogs). The latter already support the new syscalls, the > former not yet. Oh no, I didn't mean full fledged programs, just simple struct -> syscall like the set_ptime[1] I've been using for testing. >> >> Anyway, I think it would be better to split the _mask change into its >> own separate patch series, so it can more easily be used by this shared >> extent patch and it's easier to reason about the API when it's >> disconnected from the ptime/btime change. > > Agree. Nice. Thanks, koraynilay > Thanks, > Amir. [1]: #include "include/uapi/linux/fs.h" int main(int argc, char *argv[]) { char *path =3D argv[1]; long long sec =3D atoll(argv[2]); long long nsec =3D atoll(argv[3]); struct file_attr fa =3D { .fa_xflags =3D FS_XFLAG_PTIME, .fa_xflags_mask =3D FS_XFLAG_PTIME, .fa_ptime_sec =3D sec, .fa_ptime_nsec =3D nsec, }; return syscall(__NR_file_setattr, AT_FDCWD, path, &fa, sizeof(fa), AT_SYML= INK_NOFOLLOW); } --1c0f90a8e81abc35d40b9da9afa2a275d20c96cec9fed1a0bbc0e0120237 Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iHUEABYKAB0WIQRJ2XvJtXGwULYEerQ0bnTtJTDoOAUCawBVZgAKCRA0bnTtJTDo OMQNAP9TcBGO7jcWF1/PpewpEzZJSL/5N2sNKL8LH7OSI8wxDwD8Cp4yvccYDNFU CFFrvp1gJzGGHbalK20JCnIW5/YPGQ0= =XOHR -----END PGP SIGNATURE----- --1c0f90a8e81abc35d40b9da9afa2a275d20c96cec9fed1a0bbc0e0120237--