lkml.org 
[lkml]   [2001]   [Jun]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [patch] truncate_inode_pages
    Daniel Phillips wrote:
    >
    > On Sunday 10 June 2001 03:31, Andrew Morton wrote:
    > > Daniel Phillips wrote:
    > > > This is easy, just set the list head to the page about to be truncated.
    > >
    > > Works for me.
    >
    > It looks good, but it's black magic

    No, it's wrong. I'm getting BUG()s in clear_inode():

    485 if (inode->i_data.nrpages)
    486 BUG();

    (gdb) p inode->i_data
    $5 = {clean_pages = {next = 0xd7423ae4, prev = 0xc1628280}, dirty_pages = {next = 0xd7423aec, prev = 0xd7423aec},
    locked_pages = {next = 0xd7423af4, prev = 0xd7423af4}, nrpages = 0x1, a_ops = 0xc02e4940, host = 0xd7423a40,
    i_mmap = 0x0, i_mmap_shared = 0x0, i_shared_lock = {lock = 0x1}, gfp_mask = 0x5}

    (gdb) p *(struct page *)0xd7423ae4
    $6 = {list = {next = 0xd7423ae4, prev = 0xc1628280}, mapping = 0xd7423aec, index = 0xd7423aec,
    next_hash = 0xd7423af4, count = {counter = 0xd7423af4}, flags = 0x1, lru = {next = 0xc02e4940,
    prev = 0xd7423a40}, age = 0x0, wait = {lock = {lock = 0x0}, task_list = {next = 0x1, prev = 0x5}},
    pprev_hash = 0x0, buffers = 0x0, virtual = 0x0, zone = 0x0}
    (gdb) p &inode->i_data

    The lists are mangled.

    I think what's happening is:

    curr = curr->next;
    offset = page->index;

    /* Is one of the pages to truncate? */
    if ((offset >= start) || (*partial && (offset + 1) == start)) {
    list_del(head);
    list_add_tail(head, curr);

    `curr' can point to `head' when we get inside the `if'.
    So we do, effectively,

    list_del(head);
    list_add_tail(head, head);

    Which turns mapping->clean_pages into an empty list,
    and the target page.list thinks it's on a list, but
    really isn't. A subsequent list_del() on the page
    screws things up somehow.

    Ho-hum. I'll have another go tomorrow. Probably a
    simple `if (curr != head)' will cure it.

    -
    -
    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/

    \
     
     \ /
      Last update: 2005-03-22 12:55    [W:0.022 / U:0.316 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site