summaryrefslogtreecommitdiffstats
path: root/drivers/firewire/fw-device.h
diff options
context:
space:
mode:
authorStefan Richter <stefanr@s5r6.in-berlin.de>2009-05-13 21:42:14 +0200
committerStefan Richter <stefanr@s5r6.in-berlin.de>2009-05-17 14:13:47 +0200
commit92368890d551794ee8d7e90477d8498bb7f82a9b (patch)
tree0c9fe72c0d770ce4ac358235fafd18803415b1ac /drivers/firewire/fw-device.h
parent32a0f488ce5e8a9a148491f15edc508ab5e8265b (diff)
downloadop-kernel-dev-92368890d551794ee8d7e90477d8498bb7f82a9b.zip
op-kernel-dev-92368890d551794ee8d7e90477d8498bb7f82a9b.tar.gz
firewire: core: improve check for local node
My recently added test for a device being local in fw-cdev.c got it slightly wrong: Comparisons of node IDs are only valid if the generation is current, which I forgot to check. Normally, serialization by card->lock takes care of this, but a device in FW_DEVICE_GONE state will necessarily have a wrong generation and invalid node_id. The "is it local?" check is made 100% correct and simpler now by means of a struct fw_device flag which is set at fw_device creation. Besides the fw-cdev site which was to be fixed, there is another site which can make use of the new flag, and an RFC-2734 driver will benefit from it too. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Diffstat (limited to 'drivers/firewire/fw-device.h')
-rw-r--r--drivers/firewire/fw-device.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/drivers/firewire/fw-device.h b/drivers/firewire/fw-device.h
index 9758893..623cfee 100644
--- a/drivers/firewire/fw-device.h
+++ b/drivers/firewire/fw-device.h
@@ -80,6 +80,7 @@ struct fw_device {
u32 *config_rom;
size_t config_rom_length;
int config_rom_retries;
+ unsigned is_local:1;
unsigned cmc:1;
unsigned bc_implemented:2;
OpenPOWER on IntegriCloud