Re: wrong madvise(MADV_DONTNEED) semantic

From: Samuel Thibault (samuel.thibault_at_ens-lyon.org)
Date: 06/28/05

  • Next message: Valdis.Kletnieks_at_vt.edu: "Re: Memory Management during Program Loading" 
    Date:	Tue, 28 Jun 2005 20:54:47 +0200
To: Andy Isaacson <adi@hexapodia.org>

    Andy Isaacson, le Tue 28 Jun 2005 11:16:20 -0700, a écrit :
    > If your interpretation of the problem is correct, then it should be
    > trivial to write a test program demonstrating the problem. Did you
    > write the simple test program and run it?

    I indeed didn't, trusting both the man page, the source code comments,
    and my knowledge of zap_page_range().

    > MADV_DONTNEED
    > Do not expect access in the near future. (For the time
    > being, the application is finished with the given range,
    > so the kernel can free resources associated with it.)
    > Subsequent accesses of pages in this range will succeed,
    > but will result either in reloading of the memory contents
    > from the underlying mapped file (see mmap) or
    > zero-fill-on-demand pages for mappings without an
    > underlying file.
    >
    > You seem to think that "reloading ... from the underlying mapped file"
    > means that changes are lost, but that's not implied.

    I didn't say anything precise. What mostly feared me was the
    "zero-fill-on-demand pages for mappings without an underlying file."

    > Below is the test program I used.

    It does indeed work, but this is no proof. It your testcase it does
    indeed work, since the page still remains in the page cache (it's a
    shared mapping of the file). But now try this one. It uses private
    mappings, and fails as expected, getting 0 in the ANONYMOUS case, and
    the original file value in the file mapping case.

    Regards,
    Samuel Thibault

    #include <stdio.h>
    #include <stdlib.h>
    #include <stdarg.h>
    #include <string.h>
    #include <errno.h>

    #include <signal.h>
    #include <unistd.h>
    #include <sys/mman.h>
    #include <fcntl.h>
    #include <sys/stat.h>

    typedef unsigned int u32;

    void die(char *fmt, ...)
    {
        va_list ap;

        va_start(ap, fmt);
        vfprintf(stderr, fmt, ap);
        va_end(ap);
        exit(1);
    }

    int check_cookie(char *file, u32 cookie)
    {
        u32 buf;
        int fd;

        if((fd = open(file, O_RDONLY, 0)) == -1)
            die("%s: %s\n", file, strerror(errno));
        if(read(fd, &buf, sizeof(buf)) == -1)
            die("read: %s\n", strerror(errno));
        close(fd);
        return buf == cookie;
    }

    int dotest(char *file, u32 cookie, int do_anonymous, int do_msync, int do_madvise)
    {
        void *p;
        int fd, len = 16 * 1024, prot = PROT_READ|PROT_WRITE;
        u32 newcookie;

        if (!do_anonymous) {
            if((fd = open(file, O_RDWR|O_CREAT, 0666)) == -1)
                die("%s: %s\n", file, strerror(errno));

            if(ftruncate(fd, len) == -1)
                die("ftruncate: %s\n", strerror(errno));
            if(write(fd, "", 1) == -1)
                die("write: %s\n", strerror(errno));
        }

        if((p = mmap(0, len, prot, (do_anonymous ? MAP_ANONYMOUS : 0) | MAP_PRIVATE, do_anonymous ? -1 : fd, 0)) == MAP_FAILED)
            die("mmap: %s\n", strerror(errno));

        *(u32 *)p = cookie;

        if(do_msync)
            if(msync(p, len, MS_SYNC) == -1)
                die("msync: %s\n", strerror(errno));
        if(do_madvise)
            if(madvise(p, len, MADV_DONTNEED) == -1)
                die("madvise: %s\n", strerror(errno));

        newcookie = *(u32 *)p;

        printf("c = %08x msync: %s madvise: %s %s\n",
            cookie, do_msync ? "YES" : " NO", do_madvise ? "YES" : " NO",
            newcookie == cookie ? "ok" : "FAILED");
    }

    int main(int argc, char **argv)
    {
        if(argc != 2) die("usage: %s file\n", argv[0]);

        lrand48();
        dotest(argv[1], lrand48(), 0, 0, 0);
        dotest(argv[1], lrand48(), 0, 1, 0);
        dotest(argv[1], lrand48(), 0, 0, 1);
        dotest(argv[1], lrand48(), 0, 1, 1);
        dotest(argv[1], lrand48(), 1, 0, 0);
        dotest(argv[1], lrand48(), 1, 1, 0);
        dotest(argv[1], lrand48(), 1, 0, 1);
        dotest(argv[1], lrand48(), 1, 1, 1);
    }

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

  • Next message: Valdis.Kletnieks_at_vt.edu: "Re: Memory Management during Program Loading"

    Relevant Pages