> -----Original Message----- > From: Garrett Cooper [mailto:yanegomi@gmail.com] > Sent: Friday, April 16, 2010 2:06 PM > To: Mitani > Cc: Randy Dunlap; ltp-list@lists.sourceforge.net > Subject: Re: [LTP] [PATCH] fix "hugetlb" several tests > > On Thu, Apr 15, 2010 at 7:42 PM, Mitani wrote: > > > >> -----Original Message----- > >> From: Randy Dunlap [mailto:rdunlap@xenotime.net] > >> Sent: Thursday, April 15, 2010 12:20 AM > >> To: Mitani > >> Cc: 'Garrett Cooper'; ltp-list@lists.sourceforge.net > >> Subject: Re: [LTP] [PATCH] fix "hugetlb" several tests > >> > >> On Wed, 14 Apr 2010 17:58:17 +0900 Mitani wrote: > >> > >> > Hi, > >> > > >> > I suggest new patch. > >> > > >> > Signed-off-by: Tomonori Mitani > >> > >> This one is OK with me also, but... > >> > >> Is it possible to determine the difference between a kernel that > does > >> not support hugepages at all and a kernel that does support hugepages > >> but just does not have enough of them allocated? > >> > >> or maybe we don't care about this difference.  Do we? > > > > I cannot give a good opinion because I'm not a creator of these test > > programs, but ... > > I think that the difference of test refusal reason isn't so important. > > How about following message? : > > ------------ > > Not support Hugepages or not enough available Hugepages > > ------------ > > > > If users read above message, they will examine a reason by themselves > > and will do appropriate measures (set an enough hugepages value or > > gave up this test), I think. > > It's fine as-is. `Zero [supported] hugepages' is the same as `not > enough hugepages'. If people can't read the message, and put together > the details based on what's in the test output versus the > documentation and the source code, they quite frankly shouldn't be > running the tests. > Thanks, > -Garrett I made patch for both Hugepages non-support message and "tst_exit()" calling. I attach the patch as a file. Signed-off-by: Tomonori Mitani Regards-- -Tomonori Mitani