ivampiresp/Spdk
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

env_dpdk/pci: fix segfault on simultaneous VFIO hotremove and user detach

Browse Source 
There was a chance we scheduled a device removal to the DPDK thread
while that thread was already removing the device from a VFIO hotremove
notification (on the DPDK interrupt thread). The second hotremove
attempt touches some freed memory and segfaults.

The VFIO hotremove notification already checks pending_removal flag
under a mutex and sets it to true, so do the same in spdk_detach_rte()
(called from the SPDK init thread).

Change-Id: Ib3f0eb7c0c5c6e1ab8cf253b7711fd149925a143
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1730
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
This commit is contained in:
Darek Stojaczyk 2020-04-07 13:57:43 +02:00 committed by Jim Harris
parent d3bcd1ca5b
commit 7c6f0ef001
1 changed files with 10 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
lib/env_dpdk/pci.c
View File

@ -130,17 +130,19 @@ detach_rte(struct spdk_pci_device *dev)
int i;
bool removed;
/* The device was already marked as available and could be attached
* again while we go asynchronous, so we explicitly forbid that.
*/
dev->internal.pending_removal = true;
if (!spdk_process_is_primary()) {
remove_rte_dev(rte_dev);
return;
}
pthread_mutex_lock(&g_pci_mutex);
/* prevent the hotremove notification from removing this device */
dev->internal.pending_removal = true;
pthread_mutex_unlock(&g_pci_mutex);
rte_eal_alarm_set(1, detach_rte_cb, rte_dev);
/* wait up to 2s for the cb to finish executing */
/* wait up to 2s for the cb to execute */
for (i = 2000; i > 0; i--) {
spdk_delay_us(1000);
@ -209,7 +211,9 @@ pci_device_rte_hotremove(const char *device_name,
pthread_mutex_unlock(&g_pci_mutex);
if (dev != NULL && can_detach) {
/* if device is not attached we can remove it right away. */
/* if device is not attached we can remove it right away.
* Otherwise it will be removed at detach.
*/
remove_rte_dev(dev->dev_handle);
}
}