component: remove device from master match list on failed add
Subsystems (2)
PATCH 1/1 component: remove device from master match list on failed add
Calling component_add() may result in the completion of a set of devices, which will try to bring up a master. In bringing the master up, we populate its match array with the current set of children.
If binding any of the devices fails, component_add() itself will fail, free the struct component entry, and return to the caller. The now-freed entry is never removed from the master's match array, and will later be used in a futile attempt to bind to freed memory.
Bring component_add's behaviour on failure to bring up a master into line with component_del by removing the (to-be-freed) component from the master's match array.
The specific case which broke was:
- rockchip_drm_drv adds a component master
- dwhdmi_rockchip adds a child component in probe (master incomplete)
- rockchip_drm_vop adds two children in probe, which completes the set
- inside component_add, we try to bring up the master, having populated the master's match array, and fail with EPROBE_DEFER from dwhdmi_rockchip; we delete the putative component
- rockchip_drm_vop's probe fails and returns EPROBE_DEFER
- we later re-probe rockchip_drm_vop and add the component; the master is complete, so we attempt to bring it up again
- walking the match array, we find the previous child, whose master pointer doesn't match (as it has been freed in the meantime)
- rockchip_drm_vop probe fails, and will never be attempted again
--- drivers/base/component.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/base/component.c +2 -0
@@ -491,6 +491,8 @@ int component_add(struct device *dev, const struct component_ops *ops) ret =try_to_bring_up_masters (component);if (ret <0 ) { +if (component->master) +remove_component (component->master, component);list_del (&component->node);kfree (component);
Signature · 1 line
2.5.0
Quoted 3 lines · Daniel Stone <daniels <at> collabora.com> writes:
Fixes ffc30b74fd6d01588bd3fdebc3b1acc0857e6fc8Signed-off-by Daniel Stone <daniels <at> collabora.com>
Tested on imx6 processor based board where re-probe was broken after a probe deferral.
Russell,
Quoted 8 lines · Akshay Bhat
Quoted 2 lines · Daniel Stone <daniels <at> collabora.com> writes:
Fixes ffc30b74fd6d01588bd3fdebc3b1acc0857e6fc8Signed-off-by Daniel Stone <daniels <at> collabora.com>Tested-by Akshay Bhat <[email protected]>Tested on imx6 processor based board where re-probe was broken after a probe deferral.
One-week ping; this breaks quite a few drivers, even despite 57480484f9 already being present. Another option could just be to revert the original match-array commit (and the subsequent two fixups) until you can work out a fix.
Cheers,
Daniel
Quoted 9 lines · Daniel Stone
Russell,
Quoted 6 lines · Akshay Bhat
Quoted 2 lines · Daniel Stone <daniels <at> collabora.com> writes:
Fixes ffc30b74fd6d01588bd3fdebc3b1acc0857e6fc8Signed-off-by Daniel Stone <daniels <at> collabora.com>Tested-by Akshay Bhat <[email protected]>
Quoted 7 lines
Quoted 2 lines
Tested on imx6 processor based board where re-probe was broken after a probe deferral.
One-week ping; this breaks quite a few drivers, even despite
57480484f9already being present. Another option could just be to revert the original match-array commit (and the subsequent two fixups) until you can work out a fix.
I assume you haven't checked before sending out reminder emails... it's in -rc4, and it's been in linux-next since last Thursday.
Signature · 3 lines
RMK's Patch system: http://www.arm.linux.org.uk/developer/patches/ FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up according to speedtest.net.