Re: [PATCH v3] ACPI/IORT: Remove erroneous id_count check in iort_node_get_rmr_info()

From: Lorenzo Pieralisi <lpieralisi@kernel.org>
To: "guanghui.fgh" <guanghuifeng@linux.alibaba.com>
Cc: guohanjun@huawei.com, sudeep.holla@arm.com, rafael@kernel.org,
	linux-acpi@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, baolin.wang@linux.alibaba.com,
	alikernel-developer@linux.alibaba.com, will@kernel.org,
	catalin.marinas@arm.com, shameerali.kolothum.thodi@huawei.com
Subject: Re: [PATCH v3] ACPI/IORT: Remove erroneous id_count check in iort_node_get_rmr_info()
Date: Tue, 18 Jul 2023 11:16:49 +0200	[thread overview]
Message-ID: <ZLZYgZMnNB0pdycr@lpieralisi> (raw)
In-Reply-To: <8e2cfde3-15fa-4e5a-02bf-0763abf10821@linux.alibaba.com>

On Tue, Jul 18, 2023 at 04:54:36PM +0800, guanghui.fgh wrote:
> 
> 
> 在 2023/7/18 15:52, Lorenzo Pieralisi 写道:
> > [+Catalin, Will, Shameer]
> > 
> > On Mon, Jul 17, 2023 at 07:33:45PM +0800, Guanghui Feng wrote:
> > > According to the ARM IORT specifications DEN 0049 issue E,
> > > the "Number of IDs" field in the ID mapping format reports
> > > the number of IDs in the mapping range minus one.
> > > 
> > > In iort_node_get_rmr_info(), we erroneously skip ID mappings
> > > whose "Number of IDs" equal to 0, resulting in valid mapping
> > > nodes with a single ID to map being skipped, which is wrong.
> > > 
> > > Fix iort_node_get_rmr_info() by removing the bogus id_count
> > > check.
> > > 
> > > Fixes: 491cf4a6735a ("ACPI/IORT: Add support to retrieve IORT RMR reserved regions")
> > > Signed-off-by: Guanghui Feng <guanghuifeng@linux.alibaba.com>
> > > ---
> > >   drivers/acpi/arm64/iort.c | 3 ---
> > >   1 file changed, 3 deletions(-)
> > > 
> > > diff --git a/drivers/acpi/arm64/iort.c b/drivers/acpi/arm64/iort.c
> > > index 3631230..56d8873 100644
> > > --- a/drivers/acpi/arm64/iort.c
> > > +++ b/drivers/acpi/arm64/iort.c
> > > @@ -1007,9 +1007,6 @@ static void iort_node_get_rmr_info(struct acpi_iort_node *node,
> > >   	for (i = 0; i < node->mapping_count; i++, map++) {
> > >   		struct acpi_iort_node *parent;
> > > -		if (!map->id_count)
> > > -			continue;
> > > -
> > >   		parent = ACPI_ADD_PTR(struct acpi_iort_node, iort_table,
> > >   				      map->output_reference);
> > >   		if (parent != iommu)
> > 
> > Shameer, I know this may look like overkill since the hunk we are
> > removing is buggy but can you please test this patch on platforms
> > with RMR to make sure we are not triggering regressions by removing
> > it (by the specs that's what should be done but current firmware
> > is always something to reckon with) ?
> > 
> > Thanks,
> > Lorenzo
> Thanks for your response.
> 
> Firstly, I have tested iort rmr with one ID mapping and works well.
> 
> Secondly, When there is more than two id mappings, the buggly id_count check
> has no effect(for the id_count is larger than zero).

Thank you, I wasn't asking you, it is comforting to know you tested the
patch before sending it though.

I asked Shameer to make sure the id_count check did not mask a bug
in the platforms on which the RMR IORT implementation was initially
tested - that's it, I don't want to merge a fix and revert it the
next day.

Thanks,
Lorenzo