linux-kernel - Re: [PATCH rc v1 1/4] iommu/arm-smmu-v3: Add ignored bits to fix STE update sequence

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]

Message-ID: <aTXfdr5iWS7vNBJT@nvidia.com>
Date: Sun, 7 Dec 2025 12:11:34 -0800
From: Nicolin Chen <nicolinc@...dia.com>
To: Jason Gunthorpe <jgg@...dia.com>
CC: <will@...nel.org>, <robin.murphy@....com>, <joro@...tes.org>,
	<linux-arm-kernel@...ts.infradead.org>, <iommu@...ts.linux.dev>,
	<linux-kernel@...r.kernel.org>, <skolothumtho@...dia.com>, <praan@...gle.com>
Subject: Re: [PATCH rc v1 1/4] iommu/arm-smmu-v3: Add ignored bits to fix STE
 update sequence

On Sun, Dec 07, 2025 at 11:35:06AM -0800, Nicolin Chen wrote:
> On Sun, Dec 07, 2025 at 12:09:10PM -0400, Jason Gunthorpe wrote:
> > On Sat, Dec 06, 2025 at 08:37:30PM -0800, Nicolin Chen wrote:
> > > One more change that we need is at the last equation:
> > > -		if ((unused_update[i] & target_used[i]) != target[i])
> > > +		if ((unused_update[i] & target_used[i] & ~ignored[i]) !=
> > > +		    (target[i] & ~ignored[i]))
> > > 
> > > Either side might have the ignored bits, so we have to suppress
> > > ignored on both sides, which is required in the similar routine
> > > in arm_smmu_entry_differs_in_used_bits() of the kunit code.
> > 
> > The only way ignored is set is if both sides have it set and then we
> > update the bit in the firsy cycle meaning unused_update must have the
> > final value. There is no need to mask target since it will match. Not
> > changing this line is a big part of what makes this appealing because
> > it keeps the logic straightforward, in case ignored is used we shift
> > the update always to the first cycle then everything else is the same.
> 
> The reason that I changed this is because the kunit tests failed
> in arm_smmu_entry_differs_in_used_bits() when running the nested
> cases:
...
>     # arm_smmu_v3_write_ste_test_nested_s1dssbypass_to_s1bypass: EXPECTATION FAILED at drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-test.c:92
>                    Expected arm_smmu_entry_differs_in_used_bits( test_writer->entry, entry_used_bits, test_writer->init_entry, ignored, 8) && arm_smmu_entry_differs_in_used_bits( test_writer->entry, entry_used_bits, test_writer->target_entry, ignored, 8) to be false, but is true

I think I figured it out. The driver one has included ignored bits
from target in unused_update[]. But the kunit one is comparing the
raw value.

So, we don't need to change the driver one as you remarked, but we
do need to mask the target[] in the kunit one:

@@ -46,7 +46,9 @@ static bool arm_smmu_entry_differs_in_used_bits(const __le64 *entry,
        unsigned int i;

        for (i = 0; i < length; i++) {
-               if ((entry[i] & used_bits[i]) != target[i])
+               __le64 used = used_bits[i] & ~ignored[i];
+
+               if ((entry[i] & used) != (target[i] & used))
                        differs = true;
        }
        return differs;

I will wrap up v2 and send it today.

Thanks
Nicolin