----- Original Message ----- > From: "Stephen Smalley" <sds@xxxxxxxxxxxxx> > To: "Jan Stancek" <jstancek@xxxxxxxxxx> > Cc: selinux@xxxxxxxxxxxxx, serge@xxxxxxxxxx, jburke@xxxxxxxxxx > Sent: Wednesday, 23 January, 2013 4:41:38 PM > Subject: Re: [selinux-testsuite][PATCH] use extfs ioctls only when running on ext[234] > > On 01/23/2013 08:27 AM, Jan Stancek wrote: > > EXT2_* ioctls are likely to fail on other filesystems, > > for example: xfs. > > > > Signed-off-by: Jan Stancek <jstancek@xxxxxxxxxx> > > --- > > tests/ioctl/Makefile | 6 ++++++ > > tests/ioctl/test_ioctl.c | 2 ++ > > tests/ioctl/test_noioctl.c | 2 ++ > > 3 files changed, 10 insertions(+), 0 deletions(-) > > Shouldn't really be a compile-time flag but rather a runtime test of > the > filesystem type, right? Otherwise it presumes you built it on the > same > system and in the same filesystem where you will run it. Correct, I assumed everyone is running it with 'make test'. > > I thought however that your earlier patch to switch it to using > FS_IOC_ > command values would have > addressed the issue of other filesystem types already? What error > are > you getting from xfs? ioctl(3, FS_IOC32_SETVERSION or FS_IOC_SETVERSION or SONYPI_IOCGBAT1CAP, 0x7fffb5eeac24) = -1 ENOTTY (Inappropriate ioctl for device) > The code already checks for one errno value to > distinguish the not-supported case. I can't see any check for errno in this test: $ grep "errno" -r tests/ioctl/ $ Regards, Jan -- This message was distributed to subscribers of the selinux mailing list. If you no longer wish to subscribe, send mail to majordomo@xxxxxxxxxxxxx with the words "unsubscribe selinux" without quotes as the message.
