From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-yx2-f39.google.com (mail-yx2-f39.google.com [74.125.224.167]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A4A5F4C8C61 for ; Tue, 15 Sep 2026 18:49:28 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=74.125.224.167 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1789498170; cv=none; b=DQDLVmvB8FpvJcFhwAo7SaIgR4B3EPjxX7Vqy27z+uUTqBSPNW4p/OYKhDD1pvIIEUVU/y5BjlkscmWVTUotznSXLQ4NXITsKNbqgT4qeazpI3Lp4ZRFG5w9p/8p0dB924YlWJusTm6QlS/KRZZdONCm1biqMa2OX+9GRkUQzjg= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1789498170; c=relaxed/simple; bh=dn0nYsAAmXuOK3R+wRJqX7TPcEmhQYGi/QPpsB0ZaEQ=; h=Message-ID:Subject:From:To:Cc:Date:In-Reply-To:References: Content-Type:MIME-Version; b=T9qu62SJwF5AZnCrCpRQ9GvXqdpRtZ72L89GBD5ColUE8sLd005lYQMbWz/EMe7EF2kMh4qaXvPWabxS2Xxg3yNC9o0CREf7iifOqld5ocyqjelPnRT3b/ns0PuJzOOd4lLZA5g70lN7b/kcoeSGo3Lte0/gO6z4KyN/s5s7n+U= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=dubeyko.com; spf=pass smtp.mailfrom=dubeyko.com; dkim=pass (2048-bit key) header.d=dubeyko-com.20251104.gappssmtp.com header.i=@dubeyko-com.20251104.gappssmtp.com header.b=MSrt+0bs; arc=none smtp.client-ip=74.125.224.167 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=dubeyko.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=dubeyko.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=dubeyko-com.20251104.gappssmtp.com header.i=@dubeyko-com.20251104.gappssmtp.com header.b="MSrt+0bs" Received: by mail-yx2-f39.google.com with SMTP id 956f58d0204a3-66f7a9afe43so122654d50.1 for ; Tue, 15 Sep 2026 11:49:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=dubeyko-com.20251104.gappssmtp.com; s=20251104; t=1789498167; x=1790102967; darn=vger.kernel.org; h=mime-version:user-agent:content-transfer-encoding:content-type :autocrypt:references:in-reply-to:date:cc:to:from:subject:message-id :from:to:cc:subject:date:message-id:reply-to:content-type; bh=8sOdqh+xMVjOPfQTmb4GlIFWW+djME37k9ELcXIHIfs=; b=MSrt+0bsGkmYdM+x4/yC1LLmriZQ3N7OkDhwkkO+SHphTgw41hTRGswJvMaprNluLh w1LoN3rEiv87JPHt4pUfbaPBIm9UabC/Q7p+2ep2YJ+AFXqhBvXDXTSZBmc/tu+4azkw M1YzAagkIyFNMtfplLR5vfXDA2ejRqM7X/JTVD1b4ZbBh9bB9CSZDlLRzSoPm97xi2AQ 7MW0VS4Ooo3H9YzTp9ozdD6jLR0/N5wvOzJ/2HhVpnHVcb7Te88lgc8ZcMW+GWIybDqr bvdQiFkYKxNlD+YkTuzWF+amwWcQLA9WSsjAcz1uvV7KMi60Z06O538GfOnBSyhA19Ml 7aCw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20260707; t=1789498167; x=1790102967; h=mime-version:user-agent:content-transfer-encoding:content-type :autocrypt:references:in-reply-to:date:cc:to:from:subject:message-id :x-gm-gg:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to:content-type; bh=8sOdqh+xMVjOPfQTmb4GlIFWW+djME37k9ELcXIHIfs=; b=Usdc1YoXaXrAsdTP5I+XC/vVAeoraVwhT/gFieNTrkE725jg322JdIicJ1nW/G4QmO o1H9QRBiRFEwrgbofBzaccWMDyjTPoUhYQXhxRAl7Bvqy+raLD3GqR4PMBTWRGKrVj7p q0Ys5Xsok/GADTDwVuJXi6iFI3tKTVTGrE6SVpXmFRFOdi53kIWu6nosacUTP+i3/aP0 ylQUWy0Q07l3iDsDTuG3+/1sdS88A1D/J6oeF9V8DDzrWQKovFpjAdkTWFfVxc3L4TWi 2+LPPh43IRjQ0SaRPxkdrCGKXl08wugGY58KIrB6WRHfLn76dxwBClVFubvHAvZOFQLE qV4g== X-Forwarded-Encrypted: i=1; AKwUvBxWDq2gbN9B4BxooeQc7N7oMe/lKVn5FK5KpAnmwtEGosGZoUO7ttFyNd85Ua2SIrKO833gFCoqUku2XzM=@vger.kernel.org X-Gm-Message-State: AFuF++nqupp9hZA+FnIlcPPbxmtLx52DgQoYAhyRYH5CRGI1IJI8A5FL NPwwYepazyMnI2XQ+f+7BlX3rmIO+dMEYfWkI6xsl8m1Ocy5F1TM/H3zHH9rOgu303s= X-Gm-Gg: AYBFou0YPOQMj1zu3V1+Ok6WOBS41+Vsuok7hx0CV59uvX8lYE9HoZuc5AK7NCHhTcg avHuZ5WeeBgroQs2FytHUig8JiZqIwTDz5R6IZq7CqkHXdvY6aUIqwjvqFLj8RBwW+0MnFAZPDm 2lzwtuIqaG6se+Ij864yi6yDavmaEvkQsMCmuapdvlosnU+rAxx1iwdAsqMLEqDsYEkOMgCT+gN W8DiStbyrGXqPv0yx54CuYaJalrtuQvrec/gBxM6TdviAvYJ3efgs9liumaXpAqynyxrbKTE9cD HvVG+v3ZJ1Zt8i7Lln7YjGY9SM8Ldj7pwYWwCGgvjICjHnVSloF4hjiA0cP+aXnYPXoG13bY8P7 ZfxohWmd9CSNsPaYms9u5CMOzMEJOMKgL8zga5DaUTaxsoeeWYep9oCWVTGP9TsaQ6yAPrwdp6x P34zA4zWwy3y/1mM5/FDCHQoi6NnnADJ97x7kEs4wSgLROMUkKoi8Ogoc3jIAA50wPxktetkJQ6 69o7bYNLf9xAuSLsWxgJDs83GuvL78SxjtfgkXShzeeA1MGMl6kGDr6YNLn185xPTdd0SS3fUaF A5282ZQDt1pNHooNhP+aC0oKbx5BRZE3PoqnAlkAqpf4ztOtey4v+yT/dvvDgeAU X-Received: by 2002:a05:690e:1208:b0:671:2d6e:b7ec with SMTP id 956f58d0204a3-6715c93e52cmr1270970d50.0.1789498167258; Tue, 15 Sep 2026 11:49:27 -0700 (PDT) Received: from ?IPv6:2600:1700:6476:1430:e9c6:4a77:70a9:9f1d? ([2600:1700:6476:1430:e9c6:4a77:70a9:9f1d]) by smtp.gmail.com with ESMTPSA id 00721157ae682-891ee4a536asm509227b3.10.2026.09.15.11.49.26 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 15 Sep 2026 11:49:26 -0700 (PDT) Message-ID: Subject: Re: [PATCH] nilfs2: fix checkpoint root lifetime on sysfs errors From: Viacheslav Dubeyko To: Aldo Ariel Panzardo , Ryusuke Konishi Cc: linux-nilfs@vger.kernel.org, linux-kernel@vger.kernel.org, stable@vger.kernel.org Date: Tue, 15 Sep 2026 11:49:25 -0700 In-Reply-To: <20260915041206.2430937-1-qwe.aldo@gmail.com> References: <20260915041206.2430937-1-qwe.aldo@gmail.com> Autocrypt: addr=slava@dubeyko.com; prefer-encrypt=mutual; keydata=mQINBGgaTLYBEADaJc/WqWTeunGetXyyGJ5Za7b23M/ozuDCWCp+yWUa2GqQKH40dxRIR zshgOmAue7t9RQJU9lxZ4ZHWbi1Hzz85+0omefEdAKFmxTO6+CYV0g/sapU0wPJws3sC2Pbda9/eJ ZcvScAX2n/PlhpTnzJKf3JkHh3nM1ACO3jzSe2/muSQJvqMLG2D71ccekr1RyUh8V+OZdrPtfkDam V6GOT6IvyE+d+55fzmo20nJKecvbyvdikWwZvjjCENsG9qOf3TcCJ9DDYwjyYe1To8b+mQM9nHcxp jUsUuH074BhISFwt99/htZdSgp4csiGeXr8f9BEotRB6+kjMBHaiJ6B7BIlDmlffyR4f3oR/5hxgy dvIxMocqyc03xVyM6tA4ZrshKkwDgZIFEKkx37ec22ZJczNwGywKQW2TGXUTZVbdooiG4tXbRBLxe ga/NTZ52ZdEkSxAUGw/l0y0InTtdDIWvfUT+WXtQcEPRBE6HHhoeFehLzWL/o7w5Hog+0hXhNjqte fzKpI2fWmYzoIb6ueNmE/8sP9fWXo6Av9m8B5hRvF/hVWfEysr/2LSqN+xjt9NEbg8WNRMLy/Y0MS p5fgf9pmGF78waFiBvgZIQNuQnHrM+0BmYOhR0JKoHjt7r5wLyNiKFc8b7xXndyCDYfniO3ljbr0j tXWRGxx4to6FwARAQABtCZWaWFjaGVzbGF2IER1YmV5a28gPHNsYXZhQGR1YmV5a28uY29tPokCVw QTAQoAQQIbAQUJA8JnAAULCQgHAgYVCgkICwIEFgIDAQIeAQIXgBYhBFXDC2tnzsoLQtrbBDlc2cL fhEB1BQJoGl5PAhkBAAoJEDlc2cLfhEB17DsP/jy/Dx19MtxWOniPqpQf2s65enkDZuMIQ94jSg7B F2qTKIbNR9SmsczjyjC+/J7m7WZRmcqnwFYMOyNfh12aF2WhjT7p5xEAbvfGVYwUpUrg/lcacdT0D Yk61GGc5ZB89OAWHLr0FJjI54bd7kn7E/JRQF4dqNsxU8qcPXQ0wLHxTHUPZu/w5Zu/cO+lQ3H0Pj pSEGaTAh+tBYGSvQ4YPYBcV8+qjTxzeNwkw4ARza8EjTwWKP2jWAfA/ay4VobRfqNQ2zLoo84qDtN Uxe0zPE2wobIXELWkbuW/6hoQFPpMlJWz+mbvVms57NAA1HO8F5c1SLFaJ6dN0AQbxrHi45/cQXla 9hSEOJjxcEnJG/ZmcomYHFneM9K1p1K6HcGajiY2BFWkVet9vuHygkLWXVYZ0lr1paLFR52S7T+cf 6dkxOqu1ZiRegvFoyzBUzlLh/elgp3tWUfG2VmJD3lGpB3m5ZhwQ3rFpK8A7cKzgKjwPp61Me0o9z HX53THoG+QG+o0nnIKK7M8+coToTSyznYoq9C3eKeM/J97x9+h9tbizaeUQvWzQOgG8myUJ5u5Dr4 6tv9KXrOJy0iy/dcyreMYV5lwODaFfOeA4Lbnn5vRn9OjuMg1PFhCi3yMI4lA4umXFw0V2/OI5rgW BQELhfvW6mxkihkl6KLZX8m1zcHitCpWaWFjaGVzbGF2IER1YmV5a28gPFNsYXZhLkR1YmV5a29Aa WJtLmNvbT6JAlQEEwEKAD4WIQRVwwtrZ87KC0La2wQ5XNnC34RAdQUCaBpd7AIbAQUJA8JnAAULCQ gHAgYVCgkICwIEFgIDAQIeAQIXgAAKCRA5XNnC34RAdYjFEACiWBEybMt1xjRbEgaZ3UP5i2bSway DwYDvgWW5EbRP7JcqOcZ2vkJwrK3gsqC3FKpjOPh7ecE0I4vrabH1Qobe2N8B2Y396z24mGnkTBbb 16Uz3PC93nFN1BA0wuOjlr1/oOTy5gBY563vybhnXPfSEUcXRd28jI7z8tRyzXh2tL8ZLdv1u4vQ8 E0O7lVJ55p9yGxbwgb5vXU4T2irqRKLxRvU80rZIXoEM7zLf5r7RaRxgwjTKdu6rYMUOfoyEQQZTD 4Xg9YE/X8pZzcbYFs4IlscyK6cXU0pjwr2ssjearOLLDJ7ygvfOiOuCZL+6zHRunLwq2JH/RmwuLV mWWSbgosZD6c5+wu6DxV15y7zZaR3NFPOR5ErpCFUorKzBO1nA4dwOAbNym9OGkhRgLAyxwpea0V0 ZlStfp0kfVaSZYo7PXd8Bbtyjali0niBjPpEVZdgtVUpBlPr97jBYZ+L5GF3hd6WJFbEYgj+5Af7C UjbX9DHweGQ/tdXWRnJHRzorxzjOS3003ddRnPtQDDN3Z/XzdAZwQAs0RqqXrTeeJrLppFUbAP+HZ TyOLVJcAAlVQROoq8PbM3ZKIaOygjj6Yw0emJi1D9OsN2UKjoe4W185vamFWX4Ba41jmCPrYJWAWH fAMjjkInIPg7RLGs8FiwxfcpkILP0YbVWHiNAabQoVmlhY2hlc2xhdiBEdWJleWtvIDx2ZHViZXlr b0BrZXJuZWwub3JnPokCVAQTAQoAPhYhBFXDC2tnzsoLQtrbBDlc2cLfhEB1BQJoVemuAhsBBQkDw mcABQsJCAcCBhUKCQgLAgQWAgMBAh4BAheAAAoJEDlc2cLfhEB1GRwP/1scX5HO9Sk7dRicLD/fxo ipwEs+UbeA0/TM8OQfdRI4C/tFBYbQCR7lD05dfq8VsYLEyrgeLqP/iRhabLky8LTaEdwoAqPDc/O 9HRffx/faJZqkKc1dZryjqS6b8NExhKOVWmDqN357+Cl/H4hT9wnvjCj1YEqXIxSd/2Pc8+yw/KRC AP7jtRzXHcc/49Lpz/NU5irScusxy2GLKa5o/13jFK3F1fWX1wsOJF8NlTx3rLtBy4GWHITwkBmu8 zI4qcJGp7eudI0l4xmIKKQWanEhVdzBm5UnfyLIa7gQ2T48UbxJlWnMhLxMPrxgtC4Kos1G3zovEy Ep+fJN7D1pwN9aR36jVKvRsX7V4leIDWGzCdfw1FGWkMUfrRwgIl6i3wgqcCP6r9YSWVQYXdmwdMu 1RFLC44iF9340S0hw9+30yGP8TWwd1mm8V/+zsdDAFAoAwisi5QLLkQnEsJSgLzJ9daAsE8KjMthv hUWHdpiUSjyCpigT+KPl9YunZhyrC1jZXERCDPCQVYgaPt+Xbhdjcem/ykv8UVIDAGVXjuk4OW8la nf8SP+uxkTTDKcPHOa5rYRaeNj7T/NClRSd4z6aV3F6pKEJnEGvv/DFMXtSHlbylhyiGKN2Amd0b4 9jg+DW85oNN7q2UYzYuPwkHsFFq5iyF1QggiwYYTpoVXsw Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.60.1 (by Flathub.org) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 On Tue, 2026-09-15 at 01:12 -0300, Aldo Ariel Panzardo wrote: > nilfs_find_or_create_root() publishes a new root in the checkpoint > tree before creating its sysfs object.=C2=A0 If sysfs registration fails, > the root is freed while it is still reachable from the tree.=C2=A0 Merely > erasing it in the error path is insufficient because a concurrent > lookup may already hold a reference. >=20 > Serialize root creation and removal, finish sysfs registration before > publishing the root, and wait for the embedded kobject release before > freeing its container.=C2=A0 The wait also makes normal root removal safe > when kobject release is delayed. >=20 > Fixes: dd70edbde262 ("nilfs2: integrate sysfs support into driver") > Cc: stable@vger.kernel.org > Signed-off-by: Aldo Ariel Panzardo > --- > =C2=A0fs/nilfs2/the_nilfs.c | 38 +++++++++++++++++++++++++------------- > =C2=A0fs/nilfs2/the_nilfs.h |=C2=A0 3 +++ > =C2=A02 files changed, 28 insertions(+), 13 deletions(-) >=20 > diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c > index 7b23e373a..01389dba2 100644 > --- a/fs/nilfs2/the_nilfs.c > +++ b/fs/nilfs2/the_nilfs.c > @@ -71,6 +71,7 @@ struct the_nilfs *alloc_nilfs(struct super_block > *sb) > =C2=A0 spin_lock_init(&nilfs->ns_last_segment_lock); > =C2=A0 nilfs->ns_cptree =3D RB_ROOT; > =C2=A0 spin_lock_init(&nilfs->ns_cptree_lock); > + mutex_init(&nilfs->ns_cptree_mutex); > =C2=A0 init_rwsem(&nilfs->ns_segctor_sem); > =C2=A0 nilfs->ns_sb_update_freq =3D NILFS_SB_FREQ; > =C2=A0 > @@ -881,8 +882,15 @@ nilfs_find_or_create_root(struct the_nilfs > *nilfs, __u64 cno) > =C2=A0 if (!new) > =C2=A0 return NULL; > =C2=A0 > - spin_lock(&nilfs->ns_cptree_lock); > + new->cno =3D cno; > + new->ifile =3D NULL; > + new->nilfs =3D nilfs; > + refcount_set(&new->count, 1); > + atomic64_set(&new->inodes_count, 0); > + atomic64_set(&new->blocks_count, 0); > =C2=A0 > + mutex_lock(&nilfs->ns_cptree_mutex); > + spin_lock(&nilfs->ns_cptree_lock); > =C2=A0 p =3D &nilfs->ns_cptree.rb_node; > =C2=A0 parent =3D NULL; > =C2=A0 > @@ -897,29 +905,28 @@ nilfs_find_or_create_root(struct the_nilfs > *nilfs, __u64 cno) > =C2=A0 } else { > =C2=A0 refcount_inc(&root->count); > =C2=A0 spin_unlock(&nilfs->ns_cptree_lock); > + mutex_unlock(&nilfs->ns_cptree_mutex); > =C2=A0 kfree(new); > =C2=A0 return root; > =C2=A0 } > =C2=A0 } > - > - new->cno =3D cno; > - new->ifile =3D NULL; > - new->nilfs =3D nilfs; > - refcount_set(&new->count, 1); > - atomic64_set(&new->inodes_count, 0); > - atomic64_set(&new->blocks_count, 0); > - > - rb_link_node(&new->rb_node, parent, p); > - rb_insert_color(&new->rb_node, &nilfs->ns_cptree); > - > =C2=A0 spin_unlock(&nilfs->ns_cptree_lock); > =C2=A0 > =C2=A0 err =3D nilfs_sysfs_create_snapshot_group(new); > =C2=A0 if (err) { > + mutex_unlock(&nilfs->ns_cptree_mutex); > + wait_for_completion(&new->snapshot_kobj_unregister); > =C2=A0 kfree(new); > - new =3D NULL; > + return NULL; > =C2=A0 } > =C2=A0 > + spin_lock(&nilfs->ns_cptree_lock); > + rb_link_node(&new->rb_node, parent, p); > + rb_insert_color(&new->rb_node, &nilfs->ns_cptree); > + > + spin_unlock(&nilfs->ns_cptree_lock); > + mutex_unlock(&nilfs->ns_cptree_mutex); > + > =C2=A0 return new; > =C2=A0} > =C2=A0 > @@ -927,13 +934,18 @@ void nilfs_put_root(struct nilfs_root *root) > =C2=A0{ > =C2=A0 struct the_nilfs *nilfs =3D root->nilfs; > =C2=A0 > + mutex_lock(&nilfs->ns_cptree_mutex); > =C2=A0 if (refcount_dec_and_lock(&root->count, &nilfs- > >ns_cptree_lock)) { > =C2=A0 rb_erase(&root->rb_node, &nilfs->ns_cptree); > =C2=A0 spin_unlock(&nilfs->ns_cptree_lock); > =C2=A0 > =C2=A0 nilfs_sysfs_delete_snapshot_group(root); > + mutex_unlock(&nilfs->ns_cptree_mutex); > + wait_for_completion(&root- > >snapshot_kobj_unregister); > =C2=A0 iput(root->ifile); > =C2=A0 > =C2=A0 kfree(root); > + } else { > + mutex_unlock(&nilfs->ns_cptree_mutex); > =C2=A0 } > =C2=A0} > diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h > index 4776a70f0..72affad9e 100644 > --- a/fs/nilfs2/the_nilfs.h > +++ b/fs/nilfs2/the_nilfs.h > @@ -68,6 +68,7 @@ enum { > =C2=A0 * @ns_sufile: segusage file inode > =C2=A0 * @ns_cptree: rb-tree of all mounted checkpoints (nilfs_root) > =C2=A0 * @ns_cptree_lock: lock protecting @ns_cptree > + * @ns_cptree_mutex: mutex serializing checkpoint root creation and > removal > =C2=A0 * @ns_dirty_files: list of dirty files > =C2=A0 * @ns_inode_lock: lock protecting @ns_dirty_files > =C2=A0 * @ns_gc_inodes: dummy inodes to keep live blocks > @@ -151,6 +152,8 @@ struct the_nilfs { > =C2=A0 /* Checkpoint tree */ > =C2=A0 struct rb_root ns_cptree; > =C2=A0 spinlock_t ns_cptree_lock; > + /* Serialize root creation and removal. */ > + struct mutex ns_cptree_mutex; Do we really need another lock here? If the ns_cptree_lock is not good enough, then, maybe, we need to change the lock type? But adding the another lock for ns_cptree looks strange for my taste. Thanks, Slava. > =C2=A0 > =C2=A0 /* Dirty inode list */ > =C2=A0 struct list_head ns_dirty_files;