gfs2: fix infinite loop when checking ail item count before go_inval

Before this patch, the rgrp_go_inval and inode_go_inval functions each
checked if there were any items left on the ail count (by way of a
count), and if so, did a withdraw. But the withdraw code now uses
glocks when changing the file system to read-only status. So we can
not have glock functions withdrawing or a hang will likely result:
The glocks can't be serviced by the work_func if the work_func is
busy doing its own withdraw.

This patch removes the checks from the go_inval functions and adds
a centralized check in do_xmote to warn about the problem and not
withdraw, but flag the error so it's eventually caught when the logd
daemon eventually runs.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 7602d0e2492ce12e22242f3f4eaf608765045090..5afaf92057c04fc1bae2fbe628c7daf817d240e5 100644 (file)
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -604,9 +604,22 @@ __acquires(&gl->gl_lockref.lock)
        spin_unlock(&gl->gl_lockref.lock);
        if (glops->go_sync)
                glops->go_sync(gl);
-       if (test_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags))
+       if (test_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags)) {
+               /*
+                * The call to go_sync should have cleared out the ail list.
+                * If there are still items, we have a problem. We ought to
+                * withdraw, but we can't because the withdraw code also uses
+                * glocks. Warn about the error, dump the glock, then fall
+                * through and wait for logd to do the withdraw for us.
+                */
+               if ((atomic_read(&gl->gl_ail_count) != 0) &&
+                   (!cmpxchg(&sdp->sd_log_error, 0, -EIO))) {
+                       gfs2_assert_warn(sdp, !atomic_read(&gl->gl_ail_count));
+                       gfs2_dump_glock(NULL, gl, true);
+               }
                glops->go_inval(gl, target == LM_ST_DEFERRED ? 0 : DIO_METADATA);
-       clear_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags);
+               clear_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags);
+       }
 
        gfs2_glock_hold(gl);
        if (sdp->sd_lockstruct.ls_ops->lm_lock) {

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 7cfacbe35e592d76cd5366763e83fdb478d9977c..b58924482d9abadd4a57afc0a7fd54125eef0cf2 100644 (file)
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -188,7 +188,6 @@ static void rgrp_go_inval(struct gfs2_glock *gl, int flags)
                gfs2_rgrp_brelse(rgd);
 
        WARN_ON_ONCE(!(flags & DIO_METADATA));
-       gfs2_assert_withdraw(sdp, !atomic_read(&gl->gl_ail_count));
        truncate_inode_pages_range(mapping, gl->gl_vm.start, gl->gl_vm.end);
 
        if (rgd)
@@ -288,8 +287,6 @@ static void inode_go_inval(struct gfs2_glock *gl, int flags)
 {
        struct gfs2_inode *ip = gfs2_glock2inode(gl);
 
-       gfs2_assert_withdraw(gl->gl_name.ln_sbd, !atomic_read(&gl->gl_ail_count));
-
        if (flags & DIO_METADATA) {
                struct address_space *mapping = gfs2_glock2aspace(gl);
                truncate_inode_pages(mapping, 0);