Received-SPF: pass (google.com: domain of linux-kernel+bounces-37071-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Wed, 24 Jan 2024 08:39:33 -0500
From: Audra Mitchell <audra@redhat.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: shuah@kernel.org, linux-kselftest@vger.kernel.org, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, raquini@redhat.com,
	Adam Sindelar <adam@wowsignal.io>
Subject: Re: [PATCH] selftests/mm: Update va_high_addr_switch.sh to check CPU
 for la57 flag
Message-ID: <ZbETFZX33L2ELujO@fedora>
References: <20240119205801.62769-1-audra@redhat.com>
 <20240121143153.5dcfe26dea53ba0e896dbf85@linux-foundation.org>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240121143153.5dcfe26dea53ba0e896dbf85@linux-foundation.org>

On Sun, Jan 21, 2024 at 02:31:53PM -0800, Andrew Morton wrote:
> On Fri, 19 Jan 2024 15:58:01 -0500 Audra Mitchell <audra@redhat.com> wrote:
> 
> > In order for the page table level 5 to be in use, the CPU must have the
> > setting enabled in addition to the CONFIG option. Check for the flag to be
> > set to avoid false test failures on systems that do not have this cpu flag
> > set.
> > 
> 
> Thanks.
> 
> I'll add
> 
> Cc: Adam Sindelar <adam@wowsignal.io>
> Fixes: 4f2930c6718a ("selftests/vm: only run 128TBswitch with 5-level paging")
> Cc: <stable@vger.kernel.org>
> 
> A more complete description of these "test failures" would be helpful
> please.
> 

Hey, sorry for the incomplete description. The test does a series of mmap
calls including three using the MAP_FIXED flag and specifying an address that
is 1<<47 or 1<<48. These addresses are only available if you are using level 5
page tables, which requires both the CPU to have the capabiltiy (la57 flag)
and the kernel to be configured. Currently the test only checks for the kernel
configuration option, so this test can still report a false positive. Here are
the three failing lines:

$ ./va_high_addr_switch | grep FAILED
mmap(ADDR_SWITCH_HINT, 2 * PAGE_SIZE, MAP_FIXED): 0xffffffffffffffff - FAILED
mmap(HIGH_ADDR, MAP_FIXED): 0xffffffffffffffff - FAILED
mmap(ADDR_SWITCH_HINT, 2 * PAGE_SIZE, MAP_FIXED): 0xffffffffffffffff - FAILED

I thought (for about a second) refactoring the test so that these three mmap
calls will only be run on systems with the level 5 page tables available, but
the whole point of the test is to check the level 5 feature...

If you would like me to also update the description on the patch let me know
and I'll update it and resubmit. Thanks!


> > --- a/tools/testing/selftests/mm/va_high_addr_switch.sh
> > +++ b/tools/testing/selftests/mm/va_high_addr_switch.sh
> > @@ -29,9 +29,15 @@ check_supported_x86_64()
> >  	# See man 1 gzip under '-f'.
> >  	local pg_table_levels=$(gzip -dcfq "${config}" | grep PGTABLE_LEVELS | cut -d'=' -f 2)
> >  
> > +	local cpu_supports_pl5=$(awk '/^flags/ {if (/la57/) {print 0;}
> > +		else {print 1}; exit}' /proc/cpuinfo 2>/dev/null)
> > +
> >  	if [[ "${pg_table_levels}" -lt 5 ]]; then
> >  		echo "$0: PGTABLE_LEVELS=${pg_table_levels}, must be >= 5 to run this test"
> >  		exit $ksft_skip
> > +	elif [[ "${cpu_supports_pl5}" -ne 0 ]]; then
> > +		echo "$0: CPU does not have the necessary la57 flag to support page table level 5"
> > +		exit $ksft_skip
> >  	fi
> >  }
> >  
> > -- 
> > 2.43.0
>