Re: + selftests-mm-define-madv_pageout-to-fix-compilation-issues.patch added to mm-unstable branch

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

 



On 1/10/23 15:04, David Hildenbrand wrote:

Of course, we can easily test with "gcc -E" what exactly gets included.

We don't seem to have a uniform way of handling MADV_PAGEOUT and
MADV_POPULATE(READ|WRITE) even within different C sources within
tools/testing/selftests/vm itself, so it means that we spread the spectrum
of systems and compilations that we will break. :-/

To repeat what is obvious, I am not an authority on the Linux kernel,
but as I said, I have the benefit of a fresh perspective ...

There was an open source proverb that "more eyes see more" :)

The thing is: I don't lose sleeps about defines that will never change as long as it compiles. That's the most important part (for me) with these tests. Even if it would change, and I don't see how, we'd get test failures and could fix it up.

... and I raised that I ran into issues with what you propose. That's really all I am saying :)

And to repeat what I said: If we can find a cleanup for all these relevant "ifndef" in selftests/mm I'd be very happy. As long as it compiles I'm all for cleanups. It would also be acceptable to say "what David tested was wrong" -- but that needs clear justification.

Then I rest my case.

Just to assert:

Probably if MADV_PAGEOUT or MADV_POPULATE_(READ|WRITE) isn't defined in in-tree
includes, then also the kernel wouldn't support them.

#include <sys/mman.h>
#include <linux/mman.h>

#ifdef MADV_PAGEOUT
        if (swapout) {
                madvise(mem, pagesize, MADV_PAGEOUT);
                if (!pagemap_is_swapped(pagemap_fd, mem)) {
                        ksft_test_result_skip("MADV_PAGEOUT did not work, is swap enabled?\n");
                        goto munmap;
                }
        }
#endif /* MADV_PAGEOUT */

... the above seems like the silver bullet that would catch all cases.

I tend to sometimes OCD on detail and lose the big picture. That's why "the war is
made by multitude of counselors".

I am not territorial about any of my patches, so what the community thinks its the
best ... This is not my day job, but a pet project, and I like the way it makes my
brain tick. :-)

I hope the Author of my story will make up for the years the locust had eaten.

Regards,
Mirsad

--
Mirsad Goran Todorovac
Sistem inženjer
Grafički fakultet | Akademija likovnih umjetnosti
Sveučilište u Zagrebu

System engineer
Faculty of Graphic Arts | Academy of Fine Arts
University of Zagreb, Republic of Croatia



[Index of Archives]     [Kernel Archive]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]

  Powered by Linux