linux-kernel - Re: [PATCH 3/3 v2] move hugepage test examples to tools/testing/selftests/vm

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 9 Feb 2012 15:03:16 -0800
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Dave Young <dyoung@...hat.com>
Cc:	linux-kernel@...r.kernel.org, linux-mm@...ck.org,
	xiyou.wangcong@...il.com, penberg@...nel.org,
	fengguang.wu@...el.com, cl@...ux.com,
	Frederic Weisbecker <fweisbec@...il.com>
Subject: Re: [PATCH 3/3 v2] move hugepage test examples to
 tools/testing/selftests/vm

On Thu, 9 Feb 2012 09:46:22 +0800
Dave Young <dyoung@...hat.com> wrote:

> Andrew, updated the patch as below, is it ok to you?
> ---
> 
> hugepage-mmap.c, hugepage-shm.c and map_hugetlb.c in Documentation/vm are
> simple pass/fail tests, It's better to promote them to tools/testing/selftests
> 
> Thanks suggestion of Andrew Morton about this. They all need firstly setting up
> proper nr_hugepages and hugepage-mmap need to mount hugetlbfs. So I add a shell
> script run_vmtests to do such work which will call the three test programs and
> check the return value of them.
> 
> Changes to original code including below:
> a. add run_vmtests script
> b. return error when read_bytes mismatch with writed bytes.
> c. coding style fixes: do not use assignment in if condition
> 
> [v1 -> v2]:
> 1. [akpm:] rebased on runing make run_tests from Makefile
> 2. [akpm:] rename test script from run_test ro run_vmtests
> 2. fix a bug about shell exit code checking 
> 

So I tried to run this, from tools/testing/selftests.

a) The testing failed because ./vm's run_test target requires root. 

   We need to make a policy decision here.  Do we require that
   selftests run as root?  If not then the root-requiring selftests
   should warn and bail out without declaring a failure, so that those
   tests which can be run without root permissions can be successfully
   used.

b) When I ran the vm test, my machine went paralytically comatose
   for half a minute.  That's a bit rude - if all the selftests do this
   then the selftests become kinda useless.

c) I can run "make run_tests" in the top-lvel directory and all is
   well: the tools in ./vm get compiled first.  But when I do "make
   clean ; cd vm ; make run-tests" it fails, because vm/Makefile
   doesn't build the targets before trying to run them.

   This can be fixed with

--- a/tools/testing/selftests/vm/Makefile~a
+++ a/tools/testing/selftests/vm/Makefile
@@ -7,7 +7,7 @@ all: hugepage-mmap hugepage-shm  map_hug
 %: %.c
 	$(CC) $(CFLAGS) -o $@ $^

-run_tests:
+run_tests: all
 	/bin/sh ./run_vmtests

 clean:

   But this is unpleasing: a top-level "make run_tests" will end up
   trying to compile the targets twice.

   We could change the top-level Makefile to a single-pass thing
   which just descends into the subdirectories and runs "make
   run_tests".  But that gives us no way of compiling everything
   without also running everything.  That's a huge PITA if running
   everything sends your machine comatose for half a minute!

   So I think I'll go with the above patch.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/