Martin Matuska cddafdcbc5 Illumos #1313: Integer overflow in txg_delay()
The function txg_delay() is used to delay txg (transaction group)
threads in ZFS.  The timeout value for this function is calculated
using:

    int timeout = ddi_get_lbolt() + ticks;

Later, the actual wait is performed:

    while (ddi_get_lbolt() < timeout &&
        tx->tx_syncing_txg < txg-1 && !txg_stalled(dp))
            (void) cv_timedwait(&tx->tx_quiesce_more_cv, &tx->tx_sync_lock,
                timeout - ddi_get_lbolt());

The ddi_get_lbolt() function returns current uptime in clock ticks
and is typed as clock_t.  The clock_t type on 64-bit architectures
is int64_t.

The "timeout" variable will overflow depending on the tick frequency
(e.g. for 1000 it will overflow in 28.855 days). This will make the
expression "ddi_get_lbolt() < timeout" always false - txg threads will
not be delayed anymore at all. This leads to a slowdown in ZFS writes.

The attached patch initializes timeout as clock_t to match the return
value of ddi_get_lbolt().

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #352
2011-08-01 12:09:43 -07:00
2011-07-31 11:28:44 -07:00
2011-08-01 12:09:11 -07:00
2010-08-31 13:41:27 -07:00
2011-06-17 16:35:49 -07:00
2011-07-26 10:15:35 -07:00
2010-08-31 13:41:27 -07:00
2010-05-18 10:32:23 -07:00
2011-03-17 16:52:04 -07:00
2011-07-01 15:24:34 -07:00
2008-12-01 14:49:34 -08:00
2010-09-15 09:09:37 -07:00
2010-08-26 14:24:34 -07:00

Native ZFS for Linux! ZFS is an advanced file system and volume manager which was originally developed for Solaris. It has been successfully ported to FreeBSD and now there is a functional Linux ZFS kernel port too. The port currently includes a fully functional and stable SPA, DMU, and ZVOL with a ZFS Posix Layer (ZPL) on the way!

$ ./configure
$ make pkg

Full documentation for building, configuring, and using ZFS can be found at: http://zfsonlinux.org

S
Description
No description provided
Readme 122 MiB
Languages
C 70.2%
Shell 19.9%
Assembly 5.1%
M4 1.9%
Python 1.6%
Other 1.3%