Re: Module init call vs symbols exporting race?

[Jan Glauber <jang@linux.vnet.ibm.com>]

Hi Rusty,

I've seen a symbol-resolving race on s390. The qeth module uses symbols
from qdio and although the loading order seems correct and the qdio
symbols should be available the following error appears:

qdio: loading QDIO base support version 2
qeth: Unknown symbol qdio_synchronize
qeth: Unknown symbol do_QDIO
qeth: Unknown symbol qdio_initialize
qeth: Unknown symbol qdio_cleanup
qeth: Unknown symbol qdio_activate
qeth: Unknown symbol qdio_synchronize
qeth: Unknown symbol do_QDIO
qeth: Unknown symbol qdio_initialize
qeth: Unknown symbol qdio_cleanup
qeth: Unknown symbol qdio_activate
qeth: loading qeth S/390 OSA-Express driver
qeth: Device 0.0.f5f0/0.0.f5f1/0.0.f5f2 is a OSD Express card (level: 087a) with link type OSD_1000 (portname: OSAPORT)
qeth: Hardware IP fragmentation not supported on eth0
qeth: VLAN enabled 
qeth: Multicast enabled
qeth: IPV6 enabled 
qeth: Broadcast enabled 
qeth: Using SW checksumming on eth0.
qeth: Outbound TSO enabled

After that both drivers work fine but I'm curious why this happens.
Cheers,
Jan


On Tue, 2007-11-06 at 23:41 +1100, Rusty Russell wrote:
> On Tuesday 06 November 2007 20:20:58 Pavel Emelyanov wrote:
> > Hi.
> >
> > I looked at the sys_init_module() and found that the ->init callback
> > for the module is called without the module_mutex held and *after*
> > the module's symbols are exported. Doesn't this create the race when
> > loading two modules in parallel? Like this.
> 
> Hi Pavel,
> 
> 	In a word, no.  See "strong_try_module_get()".
> 
> Cheers,
> Rusty.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/