linux-kernel - Re: [PATCH 1/2] kernfs: dont take i

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 23 Dec 2022 07:11:38 +0800
From:   Ian Kent <raven@...maw.net>
To:     Anders Roxell <anders.roxell@...aro.org>, Tejun Heo <tj@...nel.org>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Minchan Kim <minchan@...nel.org>,
        Eric Sandeen <sandeen@...deen.net>,
        Al Viro <viro@...iv.linux.org.uk>,
        Rick Lindsley <ricklind@...ux.vnet.ibm.com>,
        David Howells <dhowells@...hat.com>,
        Miklos Szeredi <miklos@...redi.hu>,
        Carlos Maiolino <cmaiolino@...hat.com>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        Kernel Mailing List <linux-kernel@...r.kernel.org>,
        elver@...gle.com, arnd@...db.de
Subject: Re: [PATCH 1/2] kernfs: dont take i_lock on inode attr read


On 21/12/22 21:34, Anders Roxell wrote:
> On 2022-10-31 12:30, Tejun Heo wrote:
>> On Tue, Oct 18, 2022 at 10:32:42AM +0800, Ian Kent wrote:
>>> The kernfs write lock is held when the kernfs node inode attributes
>>> are updated. Therefore, when either kernfs_iop_getattr() or
>>> kernfs_iop_permission() are called the kernfs node inode attributes
>>> won't change.
>>>
>>> Consequently concurrent kernfs_refresh_inode() calls always copy the
>>> same values from the kernfs node.
>>>
>>> So there's no need to take the inode i_lock to get consistent values
>>> for generic_fillattr() and generic_permission(), the kernfs read lock
>>> is sufficient.
>>>
>>> Signed-off-by: Ian Kent <raven@...maw.net>
>> Acked-by: Tejun Heo <tj@...nel.org>
> Hi,
>
> Building an allmodconfig arm64 kernel on yesterdays next-20221220 and
> booting that in qemu I see the following "BUG: KCSAN: data-race in
> set_nlink / set_nlink".


I'll check if I missed any places where set_link() could be

called where the link count could be different.


If there aren't any the question will then be can writing the

same value to this location in multiple concurrent threads

corrupt it?


Ian

>
>
> ==================================================================
> [ 1540.388669][  T123] BUG: KCSAN: data-race in set_nlink / set_nlink
> [ 1540.392779][  T123]
> [ 1540.394302][  T123] write to 0xffff00000adcc3e4 of 4 bytes by task 126 on cpu 0:
> [ 1540.398828][ T123] set_nlink (/home/anders/src/kernel/next/fs/inode.c:371)
> [ 1540.401609][ T123] kernfs_refresh_inode (/home/anders/src/kernel/next/fs/kernfs/inode.c:181)
> [ 1540.404925][ T123] kernfs_iop_getattr (/home/anders/src/kernel/next/fs/kernfs/inode.c:194)
> [ 1540.408088][ T123] vfs_getattr_nosec (/home/anders/src/kernel/next/fs/stat.c:133)
> [ 1540.411334][ T123] vfs_statx (/home/anders/src/kernel/next/fs/stat.c:170)
> [ 1540.414224][ T123] vfs_fstatat (/home/anders/src/kernel/next/fs/stat.c:276)
> [ 1540.417166][ T123] __do_sys_newfstatat (/home/anders/src/kernel/next/fs/stat.c:446)
> [ 1540.420539][ T123] __arm64_sys_newfstatat (/home/anders/src/kernel/next/fs/stat.c:440 /home/anders/src/kernel/next/fs/stat.c:440)
> [ 1540.424003][ T123] el0_svc_common.constprop.0 (/home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:38 /home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:52 /home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:142)
> [ 1540.427648][ T123] do_el0_svc (/home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:197)
> [ 1540.430378][ T123] el0_svc (/home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:133 /home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:142 /home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:638)
> [ 1540.433053][ T123] el0t_64_sync_handler (/home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:656)
> [ 1540.436421][ T123] el0t_64_sync (/home/anders/src/kernel/next/arch/arm64/kernel/entry.S:584)
> [ 1540.439303][  T123]
> [ 1540.440828][  T123] 1 lock held by systemd-udevd/126:
> [ 1540.444055][ T123] #0: ffff00000609b960 (&root->kernfs_rwsem){++++}-{3:3}, at: kernfs_iop_getattr (/home/anders/src/kernel/next/fs/kernfs/inode.c:193)
> [ 1540.450699][  T123] irq event stamp: 185034
> [ 1540.453373][ T123] hardirqs last enabled at (185034): seqcount_lockdep_reader_access (/home/anders/src/kernel/next/mm/page_alloc.c:5302)
> [ 1540.460087][ T123] hardirqs last disabled at (185033): seqcount_lockdep_reader_access (/home/anders/src/kernel/next/include/linux/seqlock.h:103 (discriminator 4))
> [ 1540.466686][ T123] softirqs last enabled at (185001): fpsimd_restore_current_state (/home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:264 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:1780)
> [ 1540.473310][ T123] softirqs last disabled at (184999): fpsimd_restore_current_state (/home/anders/src/kernel/next/include/linux/bottom_half.h:20 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:242 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:1773)
> [ 1540.479872][  T123]
> [ 1540.481406][  T123] read to 0xffff00000adcc3e4 of 4 bytes by task 123 on cpu 0:
> [ 1540.485893][ T123] set_nlink (/home/anders/src/kernel/next/fs/inode.c:368)
> [ 1540.488663][ T123] kernfs_refresh_inode (/home/anders/src/kernel/next/fs/kernfs/inode.c:181)
> [ 1540.491961][ T123] kernfs_iop_permission (/home/anders/src/kernel/next/fs/kernfs/inode.c:290)
> [ 1540.495260][ T123] inode_permission (/home/anders/src/kernel/next/fs/namei.c:458 /home/anders/src/kernel/next/fs/namei.c:525)
> [ 1540.498450][ T123] link_path_walk (/home/anders/src/kernel/next/fs/namei.c:1715 /home/anders/src/kernel/next/fs/namei.c:2262)
> [ 1540.501552][ T123] path_lookupat (/home/anders/src/kernel/next/fs/namei.c:2473 (discriminator 2))
> [ 1540.504592][ T123] filename_lookup (/home/anders/src/kernel/next/fs/namei.c:2503)
> [ 1540.507740][ T123] user_path_at_empty (/home/anders/src/kernel/next/fs/namei.c:2876)
> [ 1540.511010][ T123] do_readlinkat (/home/anders/src/kernel/next/fs/stat.c:477)
> [ 1540.514097][ T123] __arm64_sys_readlinkat (/home/anders/src/kernel/next/fs/stat.c:504 /home/anders/src/kernel/next/fs/stat.c:501 /home/anders/src/kernel/next/fs/stat.c:501)
> [ 1540.517598][ T123] el0_svc_common.constprop.0 (/home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:38 /home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:52 /home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:142)
> [ 1540.521319][ T123] do_el0_svc (/home/anders/src/kernel/next/arch/arm64/kernel/syscall.c:197)
> [ 1540.524125][ T123] el0_svc (/home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:133 /home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:142 /home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:638)
> [ 1540.526795][ T123] el0t_64_sync_handler (/home/anders/src/kernel/next/arch/arm64/kernel/entry-common.c:656)
> [ 1540.530224][ T123] el0t_64_sync (/home/anders/src/kernel/next/arch/arm64/kernel/entry.S:584)
> [ 1540.533176][  T123]
> [ 1540.534710][  T123] 1 lock held by systemd-udevd/123:
> [ 1540.537977][ T123] #0: ffff00000609b960 (&root->kernfs_rwsem){++++}-{3:3}, at: kernfs_iop_permission (/home/anders/src/kernel/next/fs/kernfs/inode.c:289)
> [ 1540.544892][  T123] irq event stamp: 216564
> [ 1540.547603][ T123] hardirqs last enabled at (216564): seqcount_lockdep_reader_access (/home/anders/src/kernel/next/mm/page_alloc.c:5302)
> [ 1540.554368][ T123] hardirqs last disabled at (216563): seqcount_lockdep_reader_access (/home/anders/src/kernel/next/include/linux/seqlock.h:103 (discriminator 4))
> [ 1540.561107][ T123] softirqs last enabled at (216533): fpsimd_restore_current_state (/home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:264 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:1780)
> [ 1540.567833][ T123] softirqs last disabled at (216531): fpsimd_restore_current_state (/home/anders/src/kernel/next/include/linux/bottom_half.h:20 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:242 /home/anders/src/kernel/next/arch/arm64/kernel/fpsimd.c:1773)
> [ 1540.574496][  T123]
> [ 1540.576050][  T123] Reported by Kernel Concurrency Sanitizer on:
> [ 1540.587925][  T123] Hardware name: linux,dummy-virt (DT)
> [ 1540.591282][  T123] ==================================================================
>
>
> Reverting this patch I don't see the data race anymore.
>
> Cheers,
> Anders