On Wed, Apr 02, 2014 at 11:46:51AM +0200, Daniel Mack wrote: > When the platform initialization fails due to missing resources, it will > return -EPROBE_DEFER after dsps_musb_init() has been called. > > dsps_musb_init() calls dsps_musb_dbg_init() to allocate the debugfs > nodes. At a later point in time, the probe will be retried, and > dsps_musb_dbg_init() will be called again. debugfs_create_dir() will > fail this time, as the node already exists, and so the entire device > probe will fail with -ENOMEM. > > Fix this by moving debugfs_remove_recursive() from dsps_remove() to the > plaform's exit function, so it will be cleanly torn down when the probe > fails. It also feels more natural this way, as .exit is the counterpart > to .init. > > Signed-off-by: Daniel Mack <zonque@xxxxxxxxx> > --- > Only the current 'next' branch which will be merged into 3.15 is affected. > > drivers/usb/musb/musb_dsps.c | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/drivers/usb/musb/musb_dsps.c b/drivers/usb/musb/musb_dsps.c > index 3372ded..e2fd263 100644 > --- a/drivers/usb/musb/musb_dsps.c > +++ b/drivers/usb/musb/musb_dsps.c > @@ -470,8 +470,9 @@ static int dsps_musb_exit(struct musb *musb) > struct dsps_glue *glue = dev_get_drvdata(dev->parent); > > del_timer_sync(&glue->timer); > - > usb_phy_shutdown(musb->xceiv); > + debugfs_remove_recursive(glue->dbgfs_root); > + > return 0; > } > > @@ -708,8 +709,6 @@ static int dsps_remove(struct platform_device *pdev) > pm_runtime_put(&pdev->dev); > pm_runtime_disable(&pdev->dev); > > - debugfs_remove_recursive(glue->dbgfs_root); I don't think this is the right fix. debugfs_remove_recursive is supposed to remove the directory as well. Why isn't dsps_musb_exit() called ? I can see it being called from musb_remove() -> musb_shutdown() -> musb_platform_exit(). Looks like the problem is elsewhere. -- balbi
Attachment:
signature.asc
Description: Digital signature
