linux-kernel - Re: [PATCH v4 2/7] iommufd: Add iommufd_object_tombstone

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

Message-ID: <20250711213323.GH1951027@nvidia.com>
Date: Fri, 11 Jul 2025 18:33:23 -0300
From: Jason Gunthorpe <jgg@...dia.com>
To: Xu Yilun <yilun.xu@...ux.intel.com>
Cc: Nicolin Chen <nicolinc@...dia.com>, kevin.tian@...el.com,
	will@...nel.org, aneesh.kumar@...nel.org, iommu@...ts.linux.dev,
	linux-kernel@...r.kernel.org, joro@...tes.org, robin.murphy@....com,
	shuah@...nel.org, aik@....com, dan.j.williams@...el.com,
	baolu.lu@...ux.intel.com, yilun.xu@...el.com
Subject: Re: [PATCH v4 2/7] iommufd: Add iommufd_object_tombstone_user()
 helper

On Fri, Jul 11, 2025 at 11:02:53PM +0800, Xu Yilun wrote:
> > With one nit:
> > 
> > > -	while (!xa_empty(&ictx->objects)) {
> > > +	for (;;) {

Actually just leave this, if xa_empty does work out then the loop
exits quickly, otherwise the break will deal with the tombstones
 it.

> > >  		unsigned int destroyed = 0;
> > >  		unsigned long index;
> > > +		bool empty = true;
> > >  
> > > +		/*
> > > +		 * xa_for_each() will not return tomestones (zeroed entries),
> > > +		 * which prevent the xarray being empty. So use an empty flags
> > 
> > Since the first "empty" and the second "empty" are different things,
> > 
> > > +		 * instead of xa_empty() to indicate all entries are either
> > > +		 * NULLed or tomestoned.
> > > +		 */
> > 
> > let's write something like this (correcting typos too):
> > 
> > 		/*
> > 		 * We can't use xa_empty(), as a tombstone (NULLed entry) would
>                                                             ^
> > 		 * prevent it returning true, unlike xa_for_each() ignoring the
> > 		 * NULLed entries. So use an empty flag instead of xa_empty() to
>                    ^
> s/NULLed/zeroed, are they?

Maybe:

 We can't use xa_empty() to end the loop as the tombstones are stored
 as XA_ZERO_ENTRY in the xarray. However xa_for_each() automatically
 converts them to NULL and skips them causing xa_empty() to be kept
 false. Thus once xa_for_each() finds no further !NULL entries the
 loop is done.

?

Jason