diff options
author | Andrew Donnellan <andrew.donnellan@au1.ibm.com> | 2017-02-06 12:07:17 +1100 |
---|---|---|
committer | Michael Ellerman <mpe@ellerman.id.au> | 2017-02-21 21:32:52 +1100 |
commit | 171ed0fcd8966d82c45376f1434678e7b9d4d9b1 (patch) | |
tree | 866ae9bbb1d2a386362b1d0fb3bb9b054437981b /drivers/misc/cxl/pci.c | |
parent | 5e48dc0aa4d9daf93e9bff2a274473623a134ec8 (diff) | |
download | linux-171ed0fcd8966d82c45376f1434678e7b9d4d9b1.tar.bz2 |
cxl: fix nested locking hang during EEH hotplug
Commit 14a3ae34bfd0 ("cxl: Prevent read/write to AFU config space while AFU
not configured") introduced a rwsem to fix an invalid memory access that
occurred when someone attempts to access the config space of an AFU on a
vPHB whilst the AFU is deconfigured, such as during EEH recovery.
It turns out that it's possible to run into a nested locking issue when EEH
recovery fails and a full device hotplug is required.
cxl_pci_error_detected() deconfigures the AFU, taking a writer lock on
configured_rwsem. When EEH recovery fails, the EEH code calls
pci_hp_remove_devices() to remove the device, which in turn calls
cxl_remove() -> cxl_pci_remove_afu() -> pci_deconfigure_afu(), which tries
to grab the writer lock that's already held.
Standard rwsem semantics don't express what we really want to do here and
don't allow for nested locking. Fix this by replacing the rwsem with an
atomic_t which we can control more finely. Allow the AFU to be locked
multiple times so long as there are no readers.
Fixes: 14a3ae34bfd0 ("cxl: Prevent read/write to AFU config space while AFU not configured")
Cc: stable@vger.kernel.org # v4.9+
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Diffstat (limited to 'drivers/misc/cxl/pci.c')
-rw-r--r-- | drivers/misc/cxl/pci.c | 11 |
1 files changed, 9 insertions, 2 deletions
diff --git a/drivers/misc/cxl/pci.c b/drivers/misc/cxl/pci.c index cca938845ffd..91f645992c94 100644 --- a/drivers/misc/cxl/pci.c +++ b/drivers/misc/cxl/pci.c @@ -1129,7 +1129,7 @@ static int pci_configure_afu(struct cxl_afu *afu, struct cxl *adapter, struct pc if ((rc = cxl_native_register_psl_irq(afu))) goto err2; - up_write(&afu->configured_rwsem); + atomic_set(&afu->configured_state, 0); return 0; err2: @@ -1142,7 +1142,14 @@ err1: static void pci_deconfigure_afu(struct cxl_afu *afu) { - down_write(&afu->configured_rwsem); + /* + * It's okay to deconfigure when AFU is already locked, otherwise wait + * until there are no readers + */ + if (atomic_read(&afu->configured_state) != -1) { + while (atomic_cmpxchg(&afu->configured_state, 0, -1) != -1) + schedule(); + } cxl_native_release_psl_irq(afu); if (afu->adapter->native->sl_ops->release_serr_irq) afu->adapter->native->sl_ops->release_serr_irq(afu); |