linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: "Russell King (Oracle)" <linux@armlinux.org.uk>
To: Al Viro <viro@zeniv.linux.org.uk>
Cc: Xie Yuanbin <xieyuanbin1@huawei.com>,
	brauner@kernel.org, jack@suse.cz, will@kernel.org,
	nico@fluxnic.net, akpm@linux-foundation.org, hch@lst.de,
	jack@suse.com, wozizhi@huaweicloud.com,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org, linux-mm@kvack.org,
	lilinjie8@huawei.com, liaohua4@huawei.com,
	wangkefeng.wang@huawei.com, pangliyuan1@huawei.com
Subject: Re: [RFC PATCH] vfs: Fix might sleep in load_unaligned_zeropad() with rcu read lock held
Date: Wed, 26 Nov 2025 19:51:54 +0000	[thread overview]
Message-ID: <aSdaWjgqP4IVivlN@shell.armlinux.org.uk> (raw)
In-Reply-To: <20251126192640.GD3538@ZenIV>

On Wed, Nov 26, 2025 at 07:26:40PM +0000, Al Viro wrote:
> On Wed, Nov 26, 2025 at 07:05:05PM +0000, Russell King (Oracle) wrote:
> > On Wed, Nov 26, 2025 at 06:48:20PM +0000, Al Viro wrote:
> > > It's been years since I looked at 32bit arm exception handling, so I'd need
> > > quite a bit of (re)RTF{S,M} before I'm comfortable with poking in
> > > arch/arm/mm/fault.c; better let ARM folks deal with that.  But arch/* is
> > > where it should be dealt with; as for papering over that in fs/*:
> > 
> > Don't expect that to happen. I've not looked at it for over a decade,
> > I do very little 32-bit ARM stuff anymore. Others have modified the
> > fault handling, the VM has changed, I basically no longer have the
> > knowledge. Effectively, 32-bit ARM is unmaintained now, although it
> > still has many users.
> 
> Joy...  For quick and dirty variant (on current tree), how about
> adding
> 	if (unlikely(addr > TASK_SIZE) && !user_mode(regs))
> 		goto no_context;
> 
> right after
> 
> 	if (!ttbr0_usermode_access_allowed(regs))
> 		goto no_context;
> 
> in do_page_fault() there?
> 
> NOTE: that might or might not break vdso; I don't think it would, but...

I don't understand how that helps. Wasn't the report that the filename
crosses a page boundary in userspace, but the following page is
inaccessible which causes a fault to be taken (as it always would do).
Thus, wouldn't "addr" be a userspace address (that the kernel is
accessing) and thus be below TASK_SIZE ?

I'm also confused - if we can't take a fault and handle it while
reading the filename from userspace, how are pages that have been
swapped out or evicted from the page cache read back in from storage
which invariably results in sleeping - which we can't do here because
of the RCU context (not that I've ever understood RCU, which is why
I've always referred those bugs to Paul.)

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!


  reply	other threads:[~2025-11-26 19:52 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-26  9:05 [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context Zizhi Wo
2025-11-26 10:19 ` [RFC PATCH] vfs: Fix might sleep in load_unaligned_zeropad() with rcu read lock held Xie Yuanbin
2025-11-26 18:10   ` Al Viro
2025-11-26 18:48     ` Al Viro
2025-11-26 19:05       ` Russell King (Oracle)
2025-11-26 19:26         ` Al Viro
2025-11-26 19:51           ` Russell King (Oracle) [this message]
2025-11-26 20:02             ` Al Viro
2025-11-26 22:25               ` david laight
2025-11-26 23:51                 ` Al Viro
2025-11-26 23:31               ` Russell King (Oracle)
2025-11-27  3:03                 ` Xie Yuanbin
2025-11-27  7:20                   ` Sebastian Andrzej Siewior
2025-11-27 11:20                     ` Xie Yuanbin
2025-11-28  1:39           ` Xie Yuanbin
2025-11-26 20:42   ` Al Viro
2025-11-26 10:27 ` [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context Zizhi Wo
2025-11-26 21:12   ` Linus Torvalds
2025-11-27 10:27     ` Will Deacon
2025-11-27 10:57     ` Russell King (Oracle)
2025-11-28 17:06       ` Linus Torvalds
2025-11-29  1:01         ` Zizhi Wo
2025-11-29  1:35           ` Linus Torvalds
2025-11-29  4:08             ` [Bug report] hash_name() may cross page boundary and trigger Xie Yuanbin
2025-11-29  9:08               ` Al Viro
2025-11-29  9:25                 ` Xie Yuanbin
2025-11-29  9:44                   ` Al Viro
2025-11-29 10:05                     ` Xie Yuanbin
2025-11-29 10:45                 ` david laight
2025-11-29  8:54             ` [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context Al Viro
2025-12-01  2:08             ` Zizhi Wo
2025-11-29  2:18         ` [Bug report] hash_name() may cross page boundary and trigger Xie Yuanbin
2025-12-01 13:28         ` [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context Will Deacon
2025-12-02 12:43         ` Russell King (Oracle)
2025-12-02 13:02           ` Xie Yuanbin
2025-12-02 22:07           ` Linus Torvalds
2025-12-03  1:48             ` Xie Yuanbin
2025-12-05 12:08               ` Russell King (Oracle)
2025-11-26 18:55 ` Al Viro
2025-11-27  2:24   ` Zizhi Wo
2025-11-29  3:37     ` Al Viro
2025-11-30  3:01       ` [RFC][alpha] saner vmalloc handling (was Re: [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context) Al Viro
2025-11-30 11:32         ` david laight
2025-11-30 16:43           ` Al Viro
2025-11-30 18:14             ` Magnus Lindholm
2025-11-30 19:03             ` david laight
2025-11-30 20:31               ` Al Viro
2025-11-30 20:32                 ` Al Viro
2025-11-30 22:16         ` Linus Torvalds
2025-11-30 23:37           ` Al Viro
2025-12-01  2:03       ` [Bug report] hash_name() may cross page boundary and trigger sleep in RCU context Zizhi Wo
2025-11-27 12:59 ` Will Deacon
2025-11-28  1:17   ` Zizhi Wo
2025-11-28  1:18     ` Zizhi Wo
2025-11-28  1:39       ` Zizhi Wo
2025-11-28 12:25         ` Will Deacon
2025-11-29  1:02           ` Zizhi Wo
2025-11-29  3:55             ` Al Viro
2025-12-01  2:38               ` Zizhi Wo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=aSdaWjgqP4IVivlN@shell.armlinux.org.uk \
    --to=linux@armlinux.org.uk \
    --cc=akpm@linux-foundation.org \
    --cc=brauner@kernel.org \
    --cc=hch@lst.de \
    --cc=jack@suse.com \
    --cc=jack@suse.cz \
    --cc=liaohua4@huawei.com \
    --cc=lilinjie8@huawei.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=nico@fluxnic.net \
    --cc=pangliyuan1@huawei.com \
    --cc=viro@zeniv.linux.org.uk \
    --cc=wangkefeng.wang@huawei.com \
    --cc=will@kernel.org \
    --cc=wozizhi@huaweicloud.com \
    --cc=xieyuanbin1@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox