linux-kernel - RE: [PATCH V2] PCI/DOE: Detect on stack work items automatically

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: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 22 Nov 2022 22:06:04 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Dan Williams' <dan.j.williams@...el.com>,
        Lukas Wunner <lukas@...ner.de>
CC:     "'ira.weiny@...el.com'" <ira.weiny@...el.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Bjorn Helgaas <helgaas@...nel.org>,
        Gregory Price <gregory.price@...verge.com>,
        Jonathan Cameron <Jonathan.Cameron@...wei.com>,
        Vishal Verma <vishal.l.verma@...el.com>,
        "linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-cxl@...r.kernel.org" <linux-cxl@...r.kernel.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: RE: [PATCH V2] PCI/DOE: Detect on stack work items automatically

From: Dan Williams
> Sent: 22 November 2022 20:23
...
> > > > > Lukas pointed out that object_is_on_stack() is available to detect this
> > > > > automatically.
> > > > >
> > > > > Use object_is_on_stack() to determine the correct init work function to
> > > > > call.
> > > >
> > > > This is all a bit strange.
> > > > The 'onstack' flag is needed for the diagnostic check:
> > > > 	is_on_stack = object_is_on_stack(addr);
> > > > 	if (is_on_stack == onstack)
> > > > 		return;
> > > > 	pr_warn(...);
> > > > 	WARN_ON(1);
> > > >
> > > > So setting the flag to the location of the buffer just subverts the check.
> > > > It that is sane there ought to be a proper way to do it.
> > >
> > > If object_is_on_stack() is sufficient to check whether a struct
> > > is on the stack or not, why doesn't __init_work() use it to
> > > auto-detect whether to call debug_object_init_on_stack() or
> > > debug_object_init()?
> > >
> > > Forcing developers to use a specific initializer for something
> > > that can be auto-detected is akin to treating them like kids
> > > and telling them "You didn't say the magic word."
> > >
> > > What's the point?
> 
> I had this initial reaction as well, but INIT_WORK_ONSTACK() documents
> an important detail of the object's lifetime. Here are 2 examples of
> functions that would become trickier to read if the kernel did a
> global s/INIT_WORK_ONSTACK()/INIT_WORK()/
> 
>     synchronize_rcu_expedited_queue_work()
>     insert_wq_barrier()
> 
> ...where those take arguments that are known to come from the stack and
> be used in async context.

I suspect the check was added in response to some code that added
on on-stack item and then slept after returning from that function.

One option would be to change the diagnostic check to:
	is_on_stack != !object_is_on_stack(addr)
and then pass in 2 so the test always succeeds.
But I suspect that won't be liked.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)