Re: [PATCH 2/2] coresight: core: fix memory leak in dict->fwnode_list

[Suzuki K Poulose]

On 17/08/2023 15:46, James Clark wrote:
> On 17/08/2023 15:39, Suzuki K Poulose wrote:
> > On 17/08/2023 09:59, Junhao He wrote:
> > > There are memory leaks reported by kmemleak:
> > > ...
> > > unreferenced object 0xffff2020103c3200 (size 256):
> > >     comm "insmod", pid 4476, jiffies 4294978252 (age 50072.536s)
> > >     hex dump (first 32 bytes):
> > >       10 60 40 06 28 20 ff ff 10 c0 59 06 20 20 ff ff  .`@.( ....Y.  ..
> > >       10 e0 47 06 28 20 ff ff 10 00 49 06 28 20 ff ff  ..G.( ....I.( ..
> > >     backtrace:
> > >       [<0000000034ec4724>] __kmem_cache_alloc_node+0x2f8/0x348
> > >       [<0000000057fbc15d>] __kmalloc_node_track_caller+0x5c/0x110
> > >       [<00000055d5e34b>] krealloc+0x8c/0x178
> > >       [<00000000a4635beb>] coresight_alloc_device_name+0x128/0x188
> > > [coresight]
> > >       [<00000000a92ddfee>] funnel_cs_ops+0x10/0xfffffffffffedaa0
> > > [coresight_funnel]
> > >       [<00000000449e20f8>] dynamic_funnel_ids+0x80/0xfffffffffffed840
> > > [coresight_funnel]
> > > ...
> > >
> > > when remove driver, the golab variables defined by the macro
> > > DEFINE_CORESIGHT_DEVLIST will be released, dict->nr_idx and
> > > dict->fwnode_list are cleared to 0. The lifetime of the golab
> > > variable has ended. So the buffer pointer is lost.
> > >
> > > Use the callback of devm_add_action_or_reset() to free memory.
> >
> > Thanks for the report. But please see below:
> >
> > > Fixes: 0f5f9b6ba9e1 ("coresight: Use platform agnostic names")
> > > Signed-off-by: Junhao He <hejunhao3@xxxxxxxxxx>
> > > ---
> > >    drivers/hwtracing/coresight/coresight-core.c | 20 +++++++++++++++++++-
> > >    1 file changed, 19 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/hwtracing/coresight/coresight-core.c
> > > b/drivers/hwtracing/coresight/coresight-core.c
> > > index 9fabe00a40d6..6849faad697d 100644
> > > --- a/drivers/hwtracing/coresight/coresight-core.c
> > > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > > @@ -1756,6 +1756,20 @@ bool coresight_loses_context_with_cpu(struct
> > > device *dev)
> > >    }
> > >    EXPORT_SYMBOL_GPL(coresight_loses_context_with_cpu);
> > >    +void coresight_release_dev_list(void *data)
> > > +{
> > > +    struct coresight_dev_list *dict = data;
> > > +
> > > +    mutex_lock(&coresight_mutex);
> > > +
> > > +    if (dict->nr_idx) {
> > > +        kfree(dict->fwnode_list);
> > > +        dict->nr_idx = 0;
> > > +    }
> > > +
> > > +    mutex_unlock(&coresight_mutex);
> > > +}
> > > +
> > >    /*
> > >     * coresight_alloc_device_name - Get an index for a given device in the
> > >     * device index list specific to a driver. An index is allocated for a
> > > @@ -1766,12 +1780,16 @@
> > > EXPORT_SYMBOL_GPL(coresight_loses_context_with_cpu);
> > >    char *coresight_alloc_device_name(struct coresight_dev_list *dict,
> > >                      struct device *dev)
> > >    {
> > > -    int idx;
> > > +    int idx, ret;
> > >        char *name = NULL;
> > >        struct fwnode_handle **list;
> > >          mutex_lock(&coresight_mutex);
> > >    +    ret = devm_add_action_or_reset(dev, coresight_release_dev_list,
> > > dict);
> > > +    if (ret)
> > > +        goto done;
> >
> > This looks wrong. The devlist should be only released on the "driver"
> > unload, not on every device release. The list retains the fwnode to
> > assign the same name for a device, if it is re-probed (e.g., due to
> > -EPROBE_DEFER error).
> >
> > Suzuki
>
> I think in that case my suggestion to change it to devm_krealloc_array
> might be wrong then. Probably worth putting an explicit comment there in
> case someone tries to tidy up all the non devm allocators to devm ones.
>
> But how do you release a single device without releasing the whole driver?

You could unbind a device from the driver and that will make another 
driver probe it. I know it works for PCI/platform with VFIO.

Suzuki

> James