Re: [PATCH v2] selftests/vm: Only run 128TBswitch with 5-level paging

[Adam Sindelar <adam@wowsignal.io>]

On Mon, Jun 27, 2022 at 09:28:34AM -0700, David Vernet wrote:
> On Mon, Jun 27, 2022 at 06:13:32PM +0200, Adam Sindelar wrote:
> > The test va_128TBswitch.c expects to be able to pass mmap an address hint
> > and length that cross the address 1<<47. This is not possible without
> > 5-level page tables, so the test fails.
> > 
> > The test is already only run on 64-bit powerpc and x86 archs, but this
> > patch adds an additional check that skips the test if PG_TABLE_LEVELS < 5.
> > There is precedent for checking /proc/config.gz in selftests, e.g. in
> > selftests/firmware.
> > 
> > Running `sudo make -C tools/testing/selftests TARGETS=vm run_tests`
> > produces the desired output:
> > 
> > ```
> > ---------------------------
> > running ./va_128TBswitch.sh
> > ---------------------------
> > ./va_128TBswitch.sh: PG_TABLE_LEVELS=4, must be >= 5 to run this test
> > [SKIP]
> > -------------------------------
> > ```
> 
> Different subsystems have different workflows, but I don't recall having
> ever seen markdown used in any mm patches. Can you update the patch
> description to not use markdown?
> 

Ah, my bad - force of habit.

> > v2:
> >   - Function variables declared local
> >   - Fixed variable name typo in Makefile
> >   - Comment to explain gzip -dcfq
> 
> Can you move the version-change descriptions out of the commit summary and
> into the patch itself, directly after the ---? We don't want it to be
> included in the changelog that eventually gets merged. This is described in
> more detail in [0].
> 
> [0] https://www.kernel.org/doc/html/latest/process/submitting-patches.html#the-canonical-patch-format
> 

Fixed in v3. Thanks for the link!

> > Signed-off-by: Adam Sindelar <adam@wowsignal.io>
> > ---
> >  tools/testing/selftests/vm/Makefile          |  1 +
> >  tools/testing/selftests/vm/run_vmtests.sh    |  2 +-
> >  tools/testing/selftests/vm/va_128TBswitch.sh | 39 ++++++++++++++++++++
> >  3 files changed, 41 insertions(+), 1 deletion(-)
> >  create mode 100755 tools/testing/selftests/vm/va_128TBswitch.sh
> > 
> > diff --git a/tools/testing/selftests/vm/Makefile b/tools/testing/selftests/vm/Makefile
> > index c45e535ff4b8..7860d0969888 100644
> > --- a/tools/testing/selftests/vm/Makefile
> > +++ b/tools/testing/selftests/vm/Makefile
> > @@ -91,6 +91,7 @@ endif
> >  TEST_PROGS := run_vmtests.sh
> >  
> >  TEST_FILES := test_vmalloc.sh
> > +TEST_FILES += va_128TBswitch.sh
> >  
> >  KSFT_KHDR_INSTALL := 1
> >  include ../lib.mk
> > diff --git a/tools/testing/selftests/vm/run_vmtests.sh b/tools/testing/selftests/vm/run_vmtests.sh
> > index a2302b5faaf2..edda7350ccca 100755
> > --- a/tools/testing/selftests/vm/run_vmtests.sh
> > +++ b/tools/testing/selftests/vm/run_vmtests.sh
> > @@ -149,7 +149,7 @@ if [ $VADDR64 -ne 0 ]; then
> >  	run_test ./virtual_address_range
> >  
> >  	# virtual address 128TB switch test
> > -	run_test ./va_128TBswitch
> > +	run_test ./va_128TBswitch.sh
> >  fi # VADDR64
> >  
> >  # vmalloc stability smoke test
> > diff --git a/tools/testing/selftests/vm/va_128TBswitch.sh b/tools/testing/selftests/vm/va_128TBswitch.sh
> > new file mode 100755
> > index 000000000000..767a6465b5d2
> > --- /dev/null
> > +++ b/tools/testing/selftests/vm/va_128TBswitch.sh
> > @@ -0,0 +1,39 @@
> > +#!/bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +#
> > +# Copyright (C) 2022 Adam Sindelar (Meta) <adam@wowsignal.io>
> > +#
> > +# This is a test for mmap behavior with 5-level paging. This script wraps the
> > +# real test to check that the kernel is configured to support at least 5
> > +# pagetable levels.
> > +
> > +# 1 means the test failed
> > +exitcode=1
> 
> Agreed that it's preferable to follow the contours that already exist in
> the vm/ tests, but IMO this is pretty confusing and should be fixed for the
> whole suite in a follow-on patch. Both ksft_skip and exitcode are exit
> codes so this variable name doesn't really make sense.
> 
> > +# Kselftest framework requirement - SKIP code is 4.
> > +ksft_skip=4
> > +
> > +die()
> 
> Sorry, missed this on the first pass, but can we change the name of this
> function to fail() to show that we're failing the test rather than killing
> the testrun?
> 
> > +{
> > +    echo "$1"
> > +    exit $exitcode
> > +}
> > +
> > +check_test_requirements()
> > +{
> > +    local config="/proc/config.gz"
> > +    [[ -f "${config}" ]] || config="/boot/config-$(uname -r)"
> > +    [[ -f "${config}" ]] || die "Cannot find kernel config in /proc or /boot"
> > +
> > +    # gzip -dcfq automatically handles both compressed and plaintext input.
> > +    # See man 1 gzip under '-f'.
> > +    local pg_table_levels=$(gzip -dcfq "${config}" | grep PGTABLE_LEVELS | cut -d'=' -f 2)
> > +
> > +    if [[ "${pg_table_levels}" -lt 5 ]]; then
> > +        echo "$0: PG_TABLE_LEVELS=${pg_table_levels}, must be >= 5 to run this test"
> > +        exit $ksft_skip
> > +    fi
> > +}
> > +
> > +check_test_requirements
> > +./va_128TBswitch
> > -- 
> > 2.35.1