On Wed, Jun 21, 2017 at 11:46:12AM -0700, Darrick J. Wong wrote:
> On Wed, Jun 21, 2017 at 02:18:48PM -0400, Brian Foster wrote:
> > On Tue, Jun 20, 2017 at 06:11:18PM -0700, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> > >
> > > Remove the xfs_etest structure in favor of a per-mountpoint structure.
> > > This will give us the flexibility to set as many error injection points
> > > as we want, and later enable us to set up sysfs knobs to set the trigger
> > > frequency as we wish. This comes at a cost of higher memory use, but
> > > unti we hit 1024 injection points (we're at 29) or a lot of mounts this
> > > shouldn't be a huge issue.
> > >
> > > Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> > > ---
> > > fs/xfs/xfs_error.c | 161 +++++++++++++++++++++++++++++++---------------------
> > > fs/xfs/xfs_error.h | 23 ++++---
> > > fs/xfs/xfs_ioctl.c | 4 +
> > > fs/xfs/xfs_mount.c | 10 +++
> > > fs/xfs/xfs_mount.h | 7 ++
> > > 5 files changed, 125 insertions(+), 80 deletions(-)
> > >
> > >
> > > diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
> > > index ed7ee4e..ef16ac2 100644
> > > --- a/fs/xfs/xfs_error.c
> > > +++ b/fs/xfs/xfs_error.c
> > > @@ -25,97 +25,126 @@
> > >
> > > #ifdef DEBUG
> > >
> > > -int xfs_etest[XFS_NUM_INJECT_ERROR];
> > > -int64_t xfs_etest_fsid[XFS_NUM_INJECT_ERROR];
> > > -char * xfs_etest_fsname[XFS_NUM_INJECT_ERROR];
> > > -int xfs_error_test_active;
> > > +static unsigned int xfs_errortag_random_default[] = {
> > > + XFS_RANDOM_DEFAULT,
> > > + XFS_RANDOM_IFLUSH_1,
> > > + XFS_RANDOM_IFLUSH_2,
> > > + XFS_RANDOM_IFLUSH_3,
> > > + XFS_RANDOM_IFLUSH_4,
> > > + XFS_RANDOM_IFLUSH_5,
> > > + XFS_RANDOM_IFLUSH_6,
> > > + XFS_RANDOM_DA_READ_BUF,
> > > + XFS_RANDOM_BTREE_CHECK_LBLOCK,
> > > + XFS_RANDOM_BTREE_CHECK_SBLOCK,
> > > + XFS_RANDOM_ALLOC_READ_AGF,
> > > + XFS_RANDOM_IALLOC_READ_AGI,
> > > + XFS_RANDOM_ITOBP_INOTOBP,
> > > + XFS_RANDOM_IUNLINK,
> > > + XFS_RANDOM_IUNLINK_REMOVE,
> > > + XFS_RANDOM_DIR_INO_VALIDATE,
> > > + XFS_RANDOM_BULKSTAT_READ_CHUNK,
> > > + XFS_RANDOM_IODONE_IOERR,
> > > + XFS_RANDOM_STRATREAD_IOERR,
> > > + XFS_RANDOM_STRATCMPL_IOERR,
> > > + XFS_RANDOM_DIOWRITE_IOERR,
> > > + XFS_RANDOM_BMAPIFORMAT,
> > > + XFS_RANDOM_FREE_EXTENT,
> > > + XFS_RANDOM_RMAP_FINISH_ONE,
> > > + XFS_RANDOM_REFCOUNT_CONTINUE_UPDATE,
> > > + XFS_RANDOM_REFCOUNT_FINISH_ONE,
> > > + XFS_RANDOM_BMAP_FINISH_ONE,
> > > + XFS_RANDOM_AG_RESV_CRITICAL,
> > > +};
> > >
> > > int
> > > -xfs_error_test(int error_tag, int *fsidp, char *expression,
> > > - int line, char *file, unsigned long randfactor)
> > > +xfs_errortag_init(
> > > + struct xfs_mount *mp)
> > > {
> > > - int i;
> > > - int64_t fsid;
> > > + mp->m_errortag = kmem_zalloc(sizeof(unsigned int) * XFS_ERRTAG_MAX,
> > > + KM_SLEEP | KM_MAYFAIL);
> > > + if (!mp->m_errortag)
> > > + return -ENOMEM;
> > > + return 0;
> > > +}
> > >
> > > - if (prandom_u32() % randfactor)
> > > - return 0;
> > > +void
> > > +xfs_errortag_del(
> > > + struct xfs_mount *mp)
> > > +{
> > > + kmem_free(mp->m_errortag);
> > > +}
> > >
> > > - memcpy(&fsid, fsidp, sizeof(xfs_fsid_t));
> > > +bool
> > > +xfs_errortag_test(
> > > + struct xfs_mount *mp,
> > > + const char *expression,
> > > + const char *file,
> > > + int line,
> > > + unsigned int error_tag)
> > > +{
> > > + unsigned int randfactor;
> > >
> > > - for (i = 0; i < XFS_NUM_INJECT_ERROR; i++) {
> > > - if (xfs_etest[i] == error_tag && xfs_etest_fsid[i] == fsid) {
> > > - xfs_warn(NULL,
> > > - "Injecting error (%s) at file %s, line %d, on filesystem \"%s\"",
> > > - expression, file, line, xfs_etest_fsname[i]);
> > > - return 1;
> > > - }
> > > - }
> > > + ASSERT(error_tag < XFS_ERRTAG_MAX);
> > > + randfactor = mp->m_errortag[error_tag];
> > > + if (!randfactor || prandom_u32() % randfactor)
> > > + return false;
> > >
> > > - return 0;
> > > + xfs_warn(mp,
> > > +"Injecting error (%s) at file %s, line %d, on filesystem \"%s\"",
> > > + expression, file, line, mp->m_fsname);
> > > + return true;
> >
> > I wonder if we're going to want/need to ratelimit this now that we have
> > dynamic frequency support. I'm thinking cases like 100% error frequency
> > for the purpose of AIL item pinning could generate quite a bit of
> > output.
>
> Yeah, we should xfs_warn_ratelimited, the one I made to force repair
> to repair undamaged metadata is particularly noisy.
>
> >
> > > }
> > >
> > > int
> > > -xfs_errortag_add(unsigned int error_tag, xfs_mount_t *mp)
> > > +xfs_errortag_set(
> > > + struct xfs_mount *mp,
> > > + unsigned int error_tag,
> > > + unsigned int tag_value)
> > > {
> > > - int i;
> > > - int len;
> > > - int64_t fsid;
> > > -
> > > if (error_tag >= XFS_ERRTAG_MAX)
> > > return -EINVAL;
> > >
> > > - memcpy(&fsid, mp->m_fixedfsid, sizeof(xfs_fsid_t));
> > > + mp->m_errortag[error_tag] = tag_value;
> > > + xfs_warn(mp,
> > > +"Turned %s XFS error tag #%d",
> > > + tag_value ? "on" : "off", error_tag);
> > > + return 0;
> >
> > This is unnecessarily noisy IMO as well. If we wanted to preserve it for
> > the ioctl() case, we could pull it up into the caller. Unless it's
> > actually used somewhere that I'm not aware of, I'd be fine with killing
> > it altogether.
>
> Sure. We cough out an error message every time we actually inject
> an error anyway.
>
> >
> > > +}
> > >
> > > - for (i = 0; i < XFS_NUM_INJECT_ERROR; i++) {
> > > - if (xfs_etest_fsid[i] == fsid && xfs_etest[i] == error_tag) {
> > > - xfs_warn(mp, "error tag #%d on", error_tag);
> > > - return 0;
> > > - }
> > > - }
> > > +int
> > > +xfs_errortag_add(
> > > + struct xfs_mount *mp,
> > > + unsigned int error_tag)
> > > +{
> > > + if (error_tag >= XFS_ERRTAG_MAX)
> > > + return -EINVAL;
> > >
> > > - for (i = 0; i < XFS_NUM_INJECT_ERROR; i++) {
> > > - if (xfs_etest[i] == 0) {
> > > - xfs_warn(mp, "Turned on XFS error tag #%d",
> > > - error_tag);
> > > - xfs_etest[i] = error_tag;
> > > - xfs_etest_fsid[i] = fsid;
> > > - len = strlen(mp->m_fsname);
> > > - xfs_etest_fsname[i] = kmem_alloc(len + 1, KM_SLEEP);
> > > - strcpy(xfs_etest_fsname[i], mp->m_fsname);
> > > - xfs_error_test_active++;
> > > - return 0;
> > > - }
> > > + if (mp->m_errortag[error_tag]) {
> > > + xfs_warn(mp, "error tag #%d on", error_tag);
> > > + return 0;
> >
> > It probably doesn't matter that much, but I wonder if we should still
> > set the _random_default value even if the tag is already enabled. The
>
> The ioctl errored out if the tag was already enabled, so this preserves
> that behavior. Note that xfs_errortag_add is only called by the ioctl,
> whereas xfs_errortag_set is called by sysfs and it doesn't care what
> the previous value was.
>
Yeah, that's what I figured. I'm more weighing whether it's better for
the ioctl() to force set the value rather than error out. Previously,
returning an error implied what the value was (whatever the hardcoded
value is). With the new interface being available, it just tells us that
it's on and the value could be anything.
> Granted, maybe we don't care about checking if it's already set and
> should just go ahead and change the value, but that would change the
> behavior of an existing (debug) interface.
>
It would be a slight change, indeed. Granted this is all debug code so
I'm less hesitant about preserving things like error semantics and
whatnot. I think it's probably best to just go with whatever seems more
useful/sane. I don't care too much either way for that same reason, I
just wanted to call it out for discussion.
Brian
> > idea being to (sort of) preserve backwards compatibility for the ioctl()
> > in the event that the tag was already enabled via sysfs.
> >
> > Now that I think of it, does the ioctl() interface even support turning
> > off an individual error tag? :/
>
> Correct. You only can clear all of them, at least via ioctls.
>
> --D
>
> >
> > Brian
> >
> > > }
> > >
> > > - xfs_warn(mp, "error tag overflow, too many turned on");
> > > -
> > > - return 1;
> > > + return xfs_errortag_set(mp, error_tag,
> > > + xfs_errortag_random_default[error_tag]);
> > > }
> > >
> > > int
> > > -xfs_errortag_clearall(xfs_mount_t *mp, int loud)
> > > +xfs_errortag_clearall(
> > > + struct xfs_mount *mp,
> > > + bool loud)
> > > {
> > > - int64_t fsid;
> > > - int cleared = 0;
> > > - int i;
> > > -
> > > - memcpy(&fsid, mp->m_fixedfsid, sizeof(xfs_fsid_t));
> > > -
> > > -
> > > - for (i = 0; i < XFS_NUM_INJECT_ERROR; i++) {
> > > - if ((fsid == 0LL || xfs_etest_fsid[i] == fsid) &&
> > > - xfs_etest[i] != 0) {
> > > - cleared = 1;
> > > - xfs_warn(mp, "Clearing XFS error tag #%d",
> > > - xfs_etest[i]);
> > > - xfs_etest[i] = 0;
> > > - xfs_etest_fsid[i] = 0LL;
> > > - kmem_free(xfs_etest_fsname[i]);
> > > - xfs_etest_fsname[i] = NULL;
> > > - xfs_error_test_active--;
> > > + bool cleared = false;
> > > + unsigned int i;
> > > +
> > > + for (i = 0; i < XFS_ERRTAG_MAX; i++) {
> > > + if (mp->m_errortag[i]) {
> > > + cleared = true;
> > > + break;
> > > }
> > > }
> > >
> > > + memset(mp->m_errortag, 0, sizeof(unsigned int) * XFS_ERRTAG_MAX);
> > > +
> > > if (loud || cleared)
> > > xfs_warn(mp, "Cleared all XFS error tags for filesystem");
> > >
> > > diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h
> > > index 05f8666..341e8a0 100644
> > > --- a/fs/xfs/xfs_error.h
> > > +++ b/fs/xfs/xfs_error.h
> > > @@ -131,20 +131,23 @@ extern void xfs_verifier_error(struct xfs_buf *bp);
> > > #define XFS_RANDOM_AG_RESV_CRITICAL 4
> > >
> > > #ifdef DEBUG
> > > -extern int xfs_error_test_active;
> > > -extern int xfs_error_test(int, int *, char *, int, char *, unsigned long);
> > > -
> > > -#define XFS_NUM_INJECT_ERROR 10
> > > +extern int xfs_errortag_init(struct xfs_mount *mp);
> > > +extern void xfs_errortag_del(struct xfs_mount *mp);
> > > +extern bool xfs_errortag_test(struct xfs_mount *mp, const char *expression,
> > > + const char *file, int line, unsigned int error_tag);
> > > #define XFS_TEST_ERROR(expr, mp, tag, rf) \
> > > - ((expr) || (xfs_error_test_active && \
> > > - xfs_error_test((tag), (mp)->m_fixedfsid, "expr", __LINE__, __FILE__, \
> > > - (rf))))
> > > + ((expr) || xfs_errortag_test((mp), #expr, __FILE__, __LINE__, (tag)))
> > >
> > > -extern int xfs_errortag_add(unsigned int error_tag, struct xfs_mount *mp);
> > > -extern int xfs_errortag_clearall(struct xfs_mount *mp, int loud);
> > > +extern int xfs_errortag_set(struct xfs_mount *mp, unsigned int error_tag,
> > > + unsigned int tag_value);
> > > +extern int xfs_errortag_add(struct xfs_mount *mp, unsigned int error_tag);
> > > +extern int xfs_errortag_clearall(struct xfs_mount *mp, bool loud);
> > > #else
> > > +#define xfs_errortag_init(mp) (0)
> > > +#define xfs_errortag_del(mp)
> > > #define XFS_TEST_ERROR(expr, mp, tag, rf) (expr)
> > > -#define xfs_errortag_add(tag, mp) (ENOSYS)
> > > +#define xfs_errortag_set(mp, tag, val) (ENOSYS)
> > > +#define xfs_errortag_add(mp, tag) (ENOSYS)
> > > #define xfs_errortag_clearall(mp, loud) (ENOSYS)
> > > #endif /* DEBUG */
> > >
> > > diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
> > > index 8ffe4ea..ab0894a 100644
> > > --- a/fs/xfs/xfs_ioctl.c
> > > +++ b/fs/xfs/xfs_ioctl.c
> > > @@ -2037,14 +2037,14 @@ xfs_file_ioctl(
> > > if (copy_from_user(&in, arg, sizeof(in)))
> > > return -EFAULT;
> > >
> > > - return xfs_errortag_add(in.errtag, mp);
> > > + return xfs_errortag_add(mp, in.errtag);
> > > }
> > >
> > > case XFS_IOC_ERROR_CLEARALL:
> > > if (!capable(CAP_SYS_ADMIN))
> > > return -EPERM;
> > >
> > > - return xfs_errortag_clearall(mp, 1);
> > > + return xfs_errortag_clearall(mp, true);
> > >
> > > case XFS_IOC_FREE_EOFBLOCKS: {
> > > struct xfs_fs_eofblocks eofb;
> > > diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
> > > index cc6789d..5cdcf43 100644
> > > --- a/fs/xfs/xfs_mount.c
> > > +++ b/fs/xfs/xfs_mount.c
> > > @@ -720,10 +720,13 @@ xfs_mountfs(
> > > if (error)
> > > goto out_del_stats;
> > >
> > > + error = xfs_errortag_init(mp);
> > > + if (error)
> > > + goto out_remove_error_sysfs;
> > >
> > > error = xfs_uuid_mount(mp);
> > > if (error)
> > > - goto out_remove_error_sysfs;
> > > + goto out_remove_errortag;
> > >
> > > /*
> > > * Set the minimum read and write sizes
> > > @@ -1042,6 +1045,8 @@ xfs_mountfs(
> > > xfs_da_unmount(mp);
> > > out_remove_uuid:
> > > xfs_uuid_unmount(mp);
> > > + out_remove_errortag:
> > > + xfs_errortag_del(mp);
> > > out_remove_error_sysfs:
> > > xfs_error_sysfs_del(mp);
> > > out_del_stats:
> > > @@ -1145,10 +1150,11 @@ xfs_unmountfs(
> > > xfs_uuid_unmount(mp);
> > >
> > > #if defined(DEBUG)
> > > - xfs_errortag_clearall(mp, 0);
> > > + xfs_errortag_clearall(mp, false);
> > > #endif
> > > xfs_free_perag(mp);
> > >
> > > + xfs_errortag_del(mp);
> > > xfs_error_sysfs_del(mp);
> > > xfs_sysfs_del(&mp->m_stats.xs_kobj);
> > > xfs_sysfs_del(&mp->m_kobj);
> > > diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
> > > index 305d953..e002ac5 100644
> > > --- a/fs/xfs/xfs_mount.h
> > > +++ b/fs/xfs/xfs_mount.h
> > > @@ -199,6 +199,13 @@ typedef struct xfs_mount {
> > > bool m_fail_unmount;
> > > #ifdef DEBUG
> > > /*
> > > + * Frequency with which errors are injected. Replaces xfs_etest; the
> > > + * value stored in here is the inverse of the frequency with which the
> > > + * error triggers. 1 = always, 2 = half the time, etc.
> > > + */
> > > + unsigned int *m_errortag;
> > > +
> > > + /*
> > > * DEBUG mode instrumentation to test and/or trigger delayed allocation
> > > * block killing in the event of failed writes. When enabled, all
> > > * buffered writes are silenty dropped and handled as if they failed.
> > >
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> > > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
