Re: [PATCH] generic/395: test GETNEXTQUOTA near INT_MAX

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]



On Mon, Dec 19, 2016 at 09:23:54AM -0600, Eric Sandeen wrote:
> XFS kernel code had a bug where GETNEXTQUOTA-type
> quotactls requesting an ID near UINT_MAX could overflow
> and return 0 as the "next" active ID.
> 
> This test checks that by creating an active quota near
> UINT_MAX, then asking for the next one after it.
> 
> The proper answer is ENOENT, but if we wrap we'll return
> ID 0.
> 
> This also changes test-nextquota.c so that it checks
> both GETNEXTQUOTA and XGETNEXTQUOTA even if one fails;
> it stores the failure conditions and returns 1 if either
> of them fails.
> 
> Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx>
> ---
> 
> diff --git a/src/test-nextquota.c b/src/test-nextquota.c
> index ba4de27..73c63d8 100644
> --- a/src/test-nextquota.c
> +++ b/src/test-nextquota.c
> @@ -73,6 +73,7 @@ int main(int argc, char *argv[])
>  	int cmd;
>  	int type = -1, typeflag = 0;
>  	int verbose = 0;
> +	int retval = 0;
>  	uint id = 0, idflag = 0;
>  	char *device = NULL;
>  	char *tmp;
> @@ -140,30 +141,32 @@ int main(int argc, char *argv[])
>  	cmd = QCMD(Q_GETNEXTQUOTA, type);
>  	if (quotactl(cmd, device, id, (void *)&dqb) < 0) {
>  		perror("Q_GETNEXTQUOTA");
> -		return 1;
> +		retval = 1;
> +	} else {
> +		/*
> +		 * We only print id and inode limits because
> +		 * block count varies depending on fs block size, etc;
> +		 * this is just a sanity test that we can retrieve the quota,
> +		 * and inode limits have the same units across both calls.
> +		 */
> +		printf("id        %u\n", dqb.dqb_id);
> +		printf("ihard     %llu\n",
> +				  (unsigned long long)dqb.dqb_ihardlimit);
> +		printf("isoft     %llu\n",
> +				  (unsigned long long)dqb.dqb_isoftlimit);
>  	}
>  
> -	/*
> -	 * We only print id and inode limits because
> -	 * block count varies depending on fs block size, etc;
> -	 * this is just a sanity test that we can retrieve the quota,
> -	 * and inode limits have the same units across both calls.
> -	 */
> -	printf("id        %u\n", dqb.dqb_id);
> -	printf("ihard     %llu\n", (unsigned long long)dqb.dqb_ihardlimit);
> -	printf("isoft     %llu\n", (unsigned long long)dqb.dqb_isoftlimit);
> -
>  	if (verbose)
>  		printf("====Q_XGETNEXTQUOTA====\n");
>  	cmd = QCMD(Q_XGETNEXTQUOTA, USRQUOTA);
>  	if (quotactl(cmd, device, id, (void *)&xqb) < 0) {
>  		perror("Q_XGETNEXTQUOTA");
> -		return 1;
> +		retval = 1;
> +	} else {
> +		printf("id        %u\n", xqb.d_id);
> +		printf("ihard     %llu\n", xqb.d_ino_hardlimit);
> +		printf("isoft     %llu\n", xqb.d_ino_softlimit);
>  	}
>  
> -	printf("id        %u\n", xqb.d_id);
> -	printf("ihard     %llu\n", xqb.d_ino_hardlimit);
> -	printf("isoft     %llu\n", xqb.d_ino_softlimit);
> -
> -	return 0;
> +	return retval;
>  }
> diff --git a/tests/generic/395 b/tests/generic/395
> new file mode 100755
> index 0000000..c0cfc31
> --- /dev/null
> +++ b/tests/generic/395
> @@ -0,0 +1,99 @@
> +#! /bin/bash
> +# FS QA Test 394
> +#
> +# test out high quota ids retrieved by Q_GETNEXTQUOTA
> +# Request for next ID near 2^32 should not wrap to 0
> +#
> +# Designed to use the new Q_GETNEXTQUOTA quotactl
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2016 Red Hat, Inc.  All Rights Reserved.
> +#
> +# This program is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU General Public License as
> +# published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope that it would be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program; if not, write the Free Software Foundation,
> +# Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> +#-----------------------------------------------------------------------
> +#
> +
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +status=1	# failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> +	cd /
> +}

We still need to define and cleanup $tmp even if it's not used
explicitly in the test, it might be used by helper functions.

> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/quota
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +_supported_fs generic
> +_supported_os Linux
> +_require_quota
> +_require_scratch
> +
> +scratch_unmount 2>/dev/null

_require_scratch has done this for you :)

> +_scratch_mkfs >> $seqres.full 2>&1
> +_scratch_mount "-o usrquota,grpquota"
> +quotacheck -u -g $SCRATCH_MNT 2>/dev/null
> +quotaon $SCRATCH_MNT 2>/dev/null
> +_scratch_unmount

I don't see why above steps are necessary (mount, quotacheck and
umount), we override MOUNT_OPTIONS and do _qmount anyway. Did I miss
anything?

> +
> +TYPES="u g"

Defined but not used.

> +MOUNT_OPTIONS="-o usrquota,grpquota"
> +
> +_qmount
> +quotaon $SCRATCH_MNT 2>/dev/null

_qmount does quotacheck & quotaon for non-XFS already, seems we don't
need it too.

> +
> +# Ok, do we even have GETNEXTQUOTA?  Querying ID 0 should work.
> +$here/src/test-nextquota -i 0 -u -d $SCRATCH_DEV &> $seqres.full || \
> +	_notrun "No GETNEXTQUOTA support"

Introduce a new _require rule? And generic/244 could use it too. Perhaps
we need to put '_require_test_program "test-nextquota"' in the new
require rule too.

Thanks,
Eryu

> +
> +echo "Launch all quotas"
> +
> +# We want to create a block of quotas for an id very near
> +# 2^32, then ask for the next quota after it.  The returned
> +# ID should not overflow to 0.
> +
> +# Populate  with 2^32-4
> +ID=4294967292
> +setquota -u $ID $ID $ID $ID $ID $SCRATCH_MNT
> +touch ${SCRATCH_MNT}/${ID}
> +chown ${ID} ${SCRATCH_MNT}/${ID}
> +
> +# remount just for kicks, make sure we get it off disk
> +_scratch_unmount
> +_qmount
> +quotaon $SCRATCH_MNT 2>/dev/null
> +
> +# Ask for the next quota after $ID; should get nothing back
> +# If kernelspace wraps, we'll get 0 back.
> +for TYPE in u g; do
> +	let NEXT=ID+1
> +	echo "Ask for ID after $NEXT expecting nothing"
> +	$here/src/test-nextquota -i $NEXT -${TYPE} -d $SCRATCH_DEV
> +done
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/395.out b/tests/generic/395.out
> new file mode 100644
> index 0000000..bcd87ec
> --- /dev/null
> +++ b/tests/generic/395.out
> @@ -0,0 +1,8 @@
> +QA output created by 395
> +Launch all quotas
> +Ask for ID after 4294967293 expecting nothing
> +Q_GETNEXTQUOTA: No such file or directory
> +Q_XGETNEXTQUOTA: No such file or directory
> +Ask for ID after 4294967293 expecting nothing
> +Q_GETNEXTQUOTA: No such file or directory
> +Q_XGETNEXTQUOTA: No such file or directory
> diff --git a/tests/generic/group b/tests/generic/group
> index 20b31ef..e9d3e4a 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -397,3 +397,4 @@
>  392 auto quick metadata
>  393 auto quick rw
>  394 auto quick
> +395 auto quick quota
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Linux Filesystems Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux