parport: Fix race between port and client registration
Subsystems (2)
PATCH 1/1 parport: Fix race between port and client registration
The parport subsystem registers port devices before they are fully initialised, resulting in a race condition where client drivers such as lp can attach to ports that are not completely initialised or even being torn down.
When the port and client drivers are built as modules and loaded around the same time during boot, this occasionally results in a crash. I was able to make this happen reliably in a VM with a PC-style parallel port by patching parport_pc to fail probing:
drivers/parport/parport_pc.c static struct parport *__parport_pc_probe_port(unsigned long int base, · line 2075
Quoted 11 lines
drivers/parport/parport_pc.c +1
-1@@ -2069,7 +2069,7 @@ static struct parport *__parport_pc_probe_port(unsigned long int base,if (!p) goto out3; - base_res =request_region (base,3 , p->name); + base_res =NULL ;if (!base_res) goto out4;
and then running:
while true; do
modprobe lp & modprobe parport_pc
wait
rmmod lp parport_pc
done
for a few seconds.
In the long term I think port registration should be changed to put the call to device_add() inside parport_announce_port(), but since the latter currently cannot fail this will require changing all port drivers.
For now, add a flag to indicate whether a port has been "announced" and only try to attach client drivers to ports when the flag is set.
--- drivers/parport/share.c | 11 +++++++++-- include/linux/parport.h | 1 + 2 files changed, 10 insertions(+), 2 deletions(-)
drivers/parport/share.c +9 -2
@@ -214,10 +214,14 @@ static void get_lowlevel_driver(void)static int port_check (struct device *dev,void *dev_drv) {struct parport_driver *drv = dev_drv; +struct parport *port;/* only send ports, do not send other devices connected to bus */ -if (is_parport (dev)) - drv->match_port (to_parport_dev (dev)); +if (is_parport (dev)) { + port =to_parport_dev (dev); +if (test_bit (PARPORT_ANNOUNCED , &port->devflags)) + drv->match_port (port); + }return 0 ; }
@@ -532,6 +536,7 @@ void parport_announce_port(struct parport *port)if (slave)attach_driver_chain (slave); } +set_bit (PARPORT_ANNOUNCED , &port->devflags);mutex_unlock (®istration_lock); }EXPORT_SYMBOL (parport_announce_port);
@@ -561,6 +566,8 @@ void parport_remove_port(struct parport *port)mutex_lock (®istration_lock); +clear_bit (PARPORT_ANNOUNCED , &port->devflags); +/* Spread the word. */ detach_driver_chain (port);
include/linux/parport.h +1 -0
@@ -240,6 +240,7 @@ struct parport {unsigned long devflags;#define PARPORT_DEVPROC_REGISTERED 0 +#define PARPORT_ANNOUNCED 1struct pardevice *proc_device;/* Currently register proc device */ struct list_head full_list;
Quoted 5 lines · Ben Hutchings
The parport subsystem registers port devices before they are fully initialised, resulting in a race condition where client drivers such as lp can attach to ports that are not completely initialised or even being torn down.
Thanks Ben.
Quoted 5 lines
When the port and client drivers are built as modules and loaded around the same time during boot, this occasionally results in a crash. I was able to make this happen reliably in a VM with a PC-style parallel port by patching
parport_pcto fail probing:
<snip>
Quoted 6 lines
Fixes 6fa45a226897 ("parport: add device-model to parport subsystem")Signed-off-by Ben Hutchings <[email protected]>
Signature · 2 lines
Regards Sudip