* [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
@ 2020-02-05 12:38 Dan Carpenter
2020-02-05 17:47 ` Dan Williams
0 siblings, 1 reply; 9+ messages in thread
From: Dan Carpenter @ 2020-02-05 12:38 UTC (permalink / raw)
To: dan.j.williams; +Cc: linux-nvdimm
Hello Dan Williams,
The patch 4d88a97aa9e8: "libnvdimm, nvdimm: dimm driver and base
libnvdimm device-driver infrastructure" from May 31, 2015, leads to
the following static checker warning:
drivers/nvdimm/bus.c:511 nd_async_device_register()
error: dereferencing freed memory 'dev'
drivers/nvdimm/bus.c
502 static void nd_async_device_register(void *d, async_cookie_t cookie)
503 {
504 struct device *dev = d;
505
506 if (device_add(dev) != 0) {
507 dev_err(dev, "%s: failed\n", __func__);
508 put_device(dev);
^^^^^^^^^^^^^^^
509 }
510 put_device(dev);
^^^^^^^^^^^^^^
511 if (dev->parent)
512 put_device(dev->parent);
513 }
We call get_device() from __nd_device_register(), I guess. It seems
buggy to call put device twice on error.
regards,
dan carpenter
regards,
dan carpenter
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 12:38 [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure Dan Carpenter
@ 2020-02-05 17:47 ` Dan Williams
2020-02-05 18:10 ` Dan Carpenter
0 siblings, 1 reply; 9+ messages in thread
From: Dan Williams @ 2020-02-05 17:47 UTC (permalink / raw)
To: Dan Carpenter; +Cc: linux-nvdimm
On Wed, Feb 5, 2020 at 4:38 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> Hello Dan Williams,
>
> The patch 4d88a97aa9e8: "libnvdimm, nvdimm: dimm driver and base
> libnvdimm device-driver infrastructure" from May 31, 2015, leads to
> the following static checker warning:
>
> drivers/nvdimm/bus.c:511 nd_async_device_register()
> error: dereferencing freed memory 'dev'
>
> drivers/nvdimm/bus.c
> 502 static void nd_async_device_register(void *d, async_cookie_t cookie)
> 503 {
> 504 struct device *dev = d;
> 505
> 506 if (device_add(dev) != 0) {
> 507 dev_err(dev, "%s: failed\n", __func__);
> 508 put_device(dev);
> ^^^^^^^^^^^^^^^
> 509 }
> 510 put_device(dev);
> ^^^^^^^^^^^^^^
> 511 if (dev->parent)
> 512 put_device(dev->parent);
> 513 }
>
> We call get_device() from __nd_device_register(), I guess. It seems
> buggy to call put device twice on error.
The registration path does:
get_device(dev);
async_schedule_dev_domain(nd_async_device_register, dev,
&nd_async_domain);
...and device_add() does its own get_device(). I could add a comment
to clarify which put_device() is correlated to which put_device(), but
this seems a false positive to me.
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 17:47 ` Dan Williams
@ 2020-02-05 18:10 ` Dan Carpenter
2020-02-05 18:23 ` Dan Williams
0 siblings, 1 reply; 9+ messages in thread
From: Dan Carpenter @ 2020-02-05 18:10 UTC (permalink / raw)
To: Dan Williams; +Cc: linux-nvdimm
On Wed, Feb 05, 2020 at 09:47:01AM -0800, Dan Williams wrote:
> On Wed, Feb 5, 2020 at 4:38 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
> >
> > Hello Dan Williams,
> >
> > The patch 4d88a97aa9e8: "libnvdimm, nvdimm: dimm driver and base
> > libnvdimm device-driver infrastructure" from May 31, 2015, leads to
> > the following static checker warning:
> >
> > drivers/nvdimm/bus.c:511 nd_async_device_register()
> > error: dereferencing freed memory 'dev'
> >
> > drivers/nvdimm/bus.c
> > 502 static void nd_async_device_register(void *d, async_cookie_t cookie)
> > 503 {
> > 504 struct device *dev = d;
> > 505
> > 506 if (device_add(dev) != 0) {
> > 507 dev_err(dev, "%s: failed\n", __func__);
> > 508 put_device(dev);
> > ^^^^^^^^^^^^^^^
> > 509 }
> > 510 put_device(dev);
> > ^^^^^^^^^^^^^^
> > 511 if (dev->parent)
> > 512 put_device(dev->parent);
> > 513 }
> >
> > We call get_device() from __nd_device_register(), I guess. It seems
> > buggy to call put device twice on error.
>
> The registration path does:
>
> get_device(dev);
>
> async_schedule_dev_domain(nd_async_device_register, dev,
> &nd_async_domain);
>
> ...and device_add() does its own get_device().
device_add() does its own put_device() at the end so it's a net zero.
regards,
dan carpenter
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 18:10 ` Dan Carpenter
@ 2020-02-05 18:23 ` Dan Williams
2020-02-05 19:08 ` Dan Carpenter
0 siblings, 1 reply; 9+ messages in thread
From: Dan Williams @ 2020-02-05 18:23 UTC (permalink / raw)
To: Dan Carpenter; +Cc: linux-nvdimm
On Wed, Feb 5, 2020 at 10:11 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> On Wed, Feb 05, 2020 at 09:47:01AM -0800, Dan Williams wrote:
> > On Wed, Feb 5, 2020 at 4:38 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
> > >
> > > Hello Dan Williams,
> > >
> > > The patch 4d88a97aa9e8: "libnvdimm, nvdimm: dimm driver and base
> > > libnvdimm device-driver infrastructure" from May 31, 2015, leads to
> > > the following static checker warning:
> > >
> > > drivers/nvdimm/bus.c:511 nd_async_device_register()
> > > error: dereferencing freed memory 'dev'
> > >
> > > drivers/nvdimm/bus.c
> > > 502 static void nd_async_device_register(void *d, async_cookie_t cookie)
> > > 503 {
> > > 504 struct device *dev = d;
> > > 505
> > > 506 if (device_add(dev) != 0) {
> > > 507 dev_err(dev, "%s: failed\n", __func__);
> > > 508 put_device(dev);
> > > ^^^^^^^^^^^^^^^
> > > 509 }
> > > 510 put_device(dev);
> > > ^^^^^^^^^^^^^^
> > > 511 if (dev->parent)
> > > 512 put_device(dev->parent);
> > > 513 }
> > >
> > > We call get_device() from __nd_device_register(), I guess. It seems
> > > buggy to call put device twice on error.
> >
> > The registration path does:
> >
> > get_device(dev);
> >
> > async_schedule_dev_domain(nd_async_device_register, dev,
> > &nd_async_domain);
> >
> > ...and device_add() does its own get_device().
>
> device_add() does its own put_device() at the end so it's a net zero.
>
It does it's own, yes, but the put_device() after device_add() failure
is there to drop the reference taken by device_initialize().
Otherwise, device_add() has always documented:
* NOTE: _Never_ directly free @dev after calling this function, even
* if it returned an error! Always use put_device() to give up your
* reference instead.
...so what am I missing?
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 18:23 ` Dan Williams
@ 2020-02-05 19:08 ` Dan Carpenter
2020-02-05 19:16 ` Dan Williams
0 siblings, 1 reply; 9+ messages in thread
From: Dan Carpenter @ 2020-02-05 19:08 UTC (permalink / raw)
To: Dan Williams; +Cc: linux-nvdimm
On Wed, Feb 05, 2020 at 10:23:00AM -0800, Dan Williams wrote:
> > > > 506 if (device_add(dev) != 0) {
> > > > 507 dev_err(dev, "%s: failed\n", __func__);
> > > > 508 put_device(dev);
> > > > ^^^^^^^^^^^^^^^
> > > > 509 }
> > > > 510 put_device(dev);
> > > > ^^^^^^^^^^^^^^
> > > > 511 if (dev->parent)
> > > > 512 put_device(dev->parent);
> > > > 513 }
> > > >
> > > > We call get_device() from __nd_device_register(), I guess. It seems
> > > > buggy to call put device twice on error.
> > >
> > > The registration path does:
> > >
> > > get_device(dev);
> > >
> > > async_schedule_dev_domain(nd_async_device_register, dev,
> > > &nd_async_domain);
> > >
> > > ...and device_add() does its own get_device().
> >
> > device_add() does its own put_device() at the end so it's a net zero.
> >
>
> It does it's own, yes, but the put_device() after device_add() failure
> is there to drop the reference taken by device_initialize().
> Otherwise, device_add() has always documented:
>
> * NOTE: _Never_ directly free @dev after calling this function, even
> * if it returned an error! Always use put_device() to give up your
> * reference instead.
>
> ...so what am I missing?
The "never call kfree" is hopefully straight forward because the kobject
needs to do its own cleanup.
__nvdimm_create() allocates the dev.
nd_device_register() calls device_initialize() which call kobject_init()
so the refcount is 1.
__nd_device_register() call get_device() so the refcount is now two.
nd_async_device_register() decrements the refcount once on success.
But if device_add() fails then it decrements it twice. Now the refcount
is zero so we call nvdimm_release(). This leads to a use after free on
the next line:
put_device(dev);
if (dev->parent)
There is a trick here because depending on the debug options it
might free immediately or it might call nvdimm_release() after 4
seconds. See kobject_release() for details.
Either way if device_add() fails we return back to __nvdimm_create()
and return the zero reference count "nvdimm" pointer, which is going
to be a problem.
regards,
dan carpenter
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 19:08 ` Dan Carpenter
@ 2020-02-05 19:16 ` Dan Williams
2020-02-05 19:28 ` Dan Carpenter
0 siblings, 1 reply; 9+ messages in thread
From: Dan Williams @ 2020-02-05 19:16 UTC (permalink / raw)
To: Dan Carpenter; +Cc: linux-nvdimm
On Wed, Feb 5, 2020 at 11:08 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> On Wed, Feb 05, 2020 at 10:23:00AM -0800, Dan Williams wrote:
> > > > > 506 if (device_add(dev) != 0) {
> > > > > 507 dev_err(dev, "%s: failed\n", __func__);
> > > > > 508 put_device(dev);
> > > > > ^^^^^^^^^^^^^^^
> > > > > 509 }
> > > > > 510 put_device(dev);
> > > > > ^^^^^^^^^^^^^^
> > > > > 511 if (dev->parent)
> > > > > 512 put_device(dev->parent);
> > > > > 513 }
> > > > >
> > > > > We call get_device() from __nd_device_register(), I guess. It seems
> > > > > buggy to call put device twice on error.
> > > >
> > > > The registration path does:
> > > >
> > > > get_device(dev);
> > > >
> > > > async_schedule_dev_domain(nd_async_device_register, dev,
> > > > &nd_async_domain);
> > > >
> > > > ...and device_add() does its own get_device().
> > >
> > > device_add() does its own put_device() at the end so it's a net zero.
> > >
> >
> > It does it's own, yes, but the put_device() after device_add() failure
> > is there to drop the reference taken by device_initialize().
> > Otherwise, device_add() has always documented:
> >
> > * NOTE: _Never_ directly free @dev after calling this function, even
> > * if it returned an error! Always use put_device() to give up your
> > * reference instead.
> >
> > ...so what am I missing?
>
> The "never call kfree" is hopefully straight forward because the kobject
> needs to do its own cleanup.
>
> __nvdimm_create() allocates the dev.
> nd_device_register() calls device_initialize() which call kobject_init()
> so the refcount is 1.
> __nd_device_register() call get_device() so the refcount is now two.
> nd_async_device_register() decrements the refcount once on success.
>
> But if device_add() fails then it decrements it twice. Now the refcount
> is zero so we call nvdimm_release(). This leads to a use after free on
> the next line:
>
> put_device(dev);
> if (dev->parent)
>
> There is a trick here because depending on the debug options it
> might free immediately or it might call nvdimm_release() after 4
> seconds. See kobject_release() for details.
>
> Either way if device_add() fails we return back to __nvdimm_create()
> and return the zero reference count "nvdimm" pointer, which is going
> to be a problem.
Ugh, sorry I thought you were pointing out that there's too many
put_device() not the use after free. Yes, the use after free is a bug
that needs fixing.
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 19:16 ` Dan Williams
@ 2020-02-05 19:28 ` Dan Carpenter
2020-02-05 20:04 ` Dan Williams
0 siblings, 1 reply; 9+ messages in thread
From: Dan Carpenter @ 2020-02-05 19:28 UTC (permalink / raw)
To: Dan Williams; +Cc: linux-nvdimm
On Wed, Feb 05, 2020 at 11:16:22AM -0800, Dan Williams wrote:
> Ugh, sorry I thought you were pointing out that there's too many
> put_device() not the use after free. Yes, the use after free is a bug
> that needs fixing.
I am complaining about the device_puts... If we call device_put()
twice then it cause a problem in __nvdimm_create()
drivers/nvdimm/dimm_devs.c
506 nvdimm->sec.flags = nvdimm_security_flags(nvdimm, NVDIMM_USER);
507 nvdimm->sec.ext_flags = nvdimm_security_flags(nvdimm, NVDIMM_MASTER);
508 nd_device_register(dev);
509
510 return nvdimm;
^^^^^^
If we call device_put() twice then we this pointer within 4 seconds.
511 }
The fix is probably to make nd_device_register() return an error code so
we can do:
ret = nd_device_register(dev);
if (ret) {
device_put(&nvdimm->dev);
return NULL;
}
return nvdimm;
regards,
dan carpenter
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 19:28 ` Dan Carpenter
@ 2020-02-05 20:04 ` Dan Williams
2020-02-05 20:18 ` Dan Carpenter
0 siblings, 1 reply; 9+ messages in thread
From: Dan Williams @ 2020-02-05 20:04 UTC (permalink / raw)
To: Dan Carpenter; +Cc: linux-nvdimm
On Wed, Feb 5, 2020 at 11:28 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> On Wed, Feb 05, 2020 at 11:16:22AM -0800, Dan Williams wrote:
> > Ugh, sorry I thought you were pointing out that there's too many
> > put_device() not the use after free. Yes, the use after free is a bug
> > that needs fixing.
>
> I am complaining about the device_puts... If we call device_put()
> twice then it cause a problem in __nvdimm_create()
>
> drivers/nvdimm/dimm_devs.c
> 506 nvdimm->sec.flags = nvdimm_security_flags(nvdimm, NVDIMM_USER);
> 507 nvdimm->sec.ext_flags = nvdimm_security_flags(nvdimm, NVDIMM_MASTER);
> 508 nd_device_register(dev);
> 509
> 510 return nvdimm;
> ^^^^^^
> If we call device_put() twice then we this pointer within 4 seconds.
"we this pointer"? We "what" this pointer. 4 seconds is relative to a
runtime test case?
...but yes, point taken this looks like an obvious leak in isolation.
>
> 511 }
>
> The fix is probably to make nd_device_register() return an error code so
> we can do:
>
> ret = nd_device_register(dev);
> if (ret) {
> device_put(&nvdimm->dev);
> return NULL;
> }
>
> return nvdimm;
Ok, so this is meant to be mitigated by the fact that all consumers of
nvdimm_create() perform a nvdimm_bus_check_dimm_count() to validate
that device_add() did not fail. The reason for this organization is to
allow nvdimm initialization to happen in parallel.
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure
2020-02-05 20:04 ` Dan Williams
@ 2020-02-05 20:18 ` Dan Carpenter
0 siblings, 0 replies; 9+ messages in thread
From: Dan Carpenter @ 2020-02-05 20:18 UTC (permalink / raw)
To: Dan Williams; +Cc: linux-nvdimm
On Wed, Feb 05, 2020 at 12:04:15PM -0800, Dan Williams wrote:
> On Wed, Feb 5, 2020 at 11:28 AM Dan Carpenter <dan.carpenter@oracle.com> wrote:
> >
> > On Wed, Feb 05, 2020 at 11:16:22AM -0800, Dan Williams wrote:
> > > Ugh, sorry I thought you were pointing out that there's too many
> > > put_device() not the use after free. Yes, the use after free is a bug
> > > that needs fixing.
> >
> > I am complaining about the device_puts... If we call device_put()
> > twice then it cause a problem in __nvdimm_create()
> >
> > drivers/nvdimm/dimm_devs.c
> > 506 nvdimm->sec.flags = nvdimm_security_flags(nvdimm, NVDIMM_USER);
> > 507 nvdimm->sec.ext_flags = nvdimm_security_flags(nvdimm, NVDIMM_MASTER);
> > 508 nd_device_register(dev);
> > 509
> > 510 return nvdimm;
> > ^^^^^^
> > If we call device_put() twice then we this pointer within 4 seconds.
>
> "we this pointer"? We "what" this pointer. 4 seconds is relative to a
> runtime test case?
>
Sorry. I meant we *free* it. The second device_put() leads to a
nvdimm_release(dev) where dev is "&nvdimm->dev" within 0-4 seconds.
Most times it will free it immediately but if you have
CONFIG_DEBUG_KOBJECT_RELEASE enabled then it will wait between 1-4
seconds and then free nvdimm. It's a config option, not a runtime
thing.
regards,
dan carpenter
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2020-02-05 20:19 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-05 12:38 [bug report] libnvdimm, nvdimm: dimm driver and base libnvdimm device-driver infrastructure Dan Carpenter
2020-02-05 17:47 ` Dan Williams
2020-02-05 18:10 ` Dan Carpenter
2020-02-05 18:23 ` Dan Williams
2020-02-05 19:08 ` Dan Carpenter
2020-02-05 19:16 ` Dan Williams
2020-02-05 19:28 ` Dan Carpenter
2020-02-05 20:04 ` Dan Williams
2020-02-05 20:18 ` Dan Carpenter
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).