Re: [PATCH] ipmi: Don't allow device module unload when in use

[tony camuso]

On 10/16/19 8:18 AM, tony camuso wrote:
> On 10/14/19 11:46 AM, minyard@xxxxxxx wrote:
> > From: Corey Minyard <cminyard@xxxxxxxxxx>
> >
> > If something has the IPMI driver open, don't allow the device
> > module to be unloaded. Before it would unload and the user would
> > get errors on use.
> >
> > This change is made on user request, and it makes it consistent
> > with the I2C driver, which has the same behavior.
> >
> > It does change things a little bit with respect to kernel users.
> > If the ACPI or IPMI watchdog (or any other kernel user) has
> > created a user, then the device module cannot be unloaded. Before
> > it could be unloaded,
> >
> > This does not affect hot-plug. If the device goes away (it's on
> > something removable that is removed or is hot-removed via sysfs)
> > then it still behaves as it did before.
> >
> > Reported-by: tony camuso <tcamuso@xxxxxxxxxx>
> > Signed-off-by: Corey Minyard <cminyard@xxxxxxxxxx>
> > ---
> > Tony, here is a suggested change for this. Can you look it over and
> > see if it looks ok?
> >
> > Thanks,
> >
> > -corey
> >
> > Â drivers/char/ipmi/ipmi_msghandler.c | 23 ++++++++++++++++-------
> > Â include/linux/ipmi_smi.hÂÂÂÂÂÂÂÂÂÂÂ | 12 ++++++++----
> > Â 2 files changed, 24 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c
> > index 2aab80e19ae0..15680de18625 100644
> > --- a/drivers/char/ipmi/ipmi_msghandler.c
> > +++ b/drivers/char/ipmi/ipmi_msghandler.c
> > @@ -448,6 +448,8 @@ enum ipmi_stat_indexes {
> > Â #define IPMI_IPMB_NUM_SEQÂÂÂ 64
> > Â struct ipmi_smi {
> > +ÂÂÂ struct module *owner;
> > +
> > ÂÂÂÂÂ /* What interface number are we? */
> > ÂÂÂÂÂ int intf_num;
> > @@ -1220,6 +1222,11 @@ int ipmi_create_user(unsigned intÂÂÂÂÂÂÂÂÂ if_num,
> > ÂÂÂÂÂ if (rv)
> > ÂÂÂÂÂÂÂÂÂ goto out_kfree;
> > +ÂÂÂ if (!try_module_get(intf->owner)) {
> > +ÂÂÂÂÂÂÂ rv = -ENODEV;
> > +ÂÂÂÂÂÂÂ goto out_kfree;
> > +ÂÂÂ }
> > +
> > ÂÂÂÂÂ /* Note that each existing user holds a refcount to the interface. */
> > ÂÂÂÂÂ kref_get(&intf->refcount);
> > @@ -1349,6 +1356,7 @@ static void _ipmi_destroy_user(struct ipmi_user *user)
> > ÂÂÂÂÂ }
> > ÂÂÂÂÂ kref_put(&intf->refcount, intf_free);
> > +ÂÂÂ module_put(intf->owner);
> > Â }
> > Â int ipmi_destroy_user(struct ipmi_user *user)
> > @@ -2459,7 +2467,7 @@ static int __get_device_id(struct ipmi_smi *intf, struct bmc_device *bmc)
> >  * been recently fetched, this will just use the cached data. Otherwise
> > ÂÂ * it will run a new fetch.
> > ÂÂ *
> > - * Except for the first time this is called (in ipmi_register_smi()),
> > + * Except for the first time this is called (in ipmi_add_smi()),
> > ÂÂ * this will always return good data;
> > ÂÂ */
> > Â static int __bmc_get_device_id(struct ipmi_smi *intf, struct bmc_device *bmc,
> > @@ -3377,10 +3385,11 @@ static void redo_bmc_reg(struct work_struct *work)
> > ÂÂÂÂÂ kref_put(&intf->refcount, intf_free);
> > Â }
> > -int ipmi_register_smi(const struct ipmi_smi_handlers *handlers,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ voidÂÂÂÂÂÂÂÂÂÂÂÂÂÂ *send_info,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ struct deviceÂÂÂÂÂÂÂÂÂÂÂ *si_dev,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ unsigned charÂÂÂÂÂÂÂÂÂÂÂ slave_addr)
> > +int ipmi_add_smi(struct moduleÂÂÂÂÂÂÂÂ *owner,
> > +ÂÂÂÂÂÂÂÂ const struct ipmi_smi_handlers *handlers,
> > +ÂÂÂÂÂÂÂÂ voidÂÂÂÂÂÂÂÂÂÂÂÂÂÂ *send_info,
> > +ÂÂÂÂÂÂÂÂ struct deviceÂÂÂÂÂÂÂÂ *si_dev,
> > +ÂÂÂÂÂÂÂÂ unsigned charÂÂÂÂÂÂÂÂ slave_addr)
> > Â {
> > ÂÂÂÂÂ intÂÂÂÂÂÂÂÂÂÂÂÂÂ i, j;
> > ÂÂÂÂÂ intÂÂÂÂÂÂÂÂÂÂÂÂÂ rv;
> > @@ -3406,7 +3415,7 @@ int ipmi_register_smi(const struct ipmi_smi_handlers *handlers,
> > ÂÂÂÂÂÂÂÂÂ return rv;
> > ÂÂÂÂÂ }
> > -
> > +ÂÂÂ intf->owner = owner;
> > ÂÂÂÂÂ intf->bmc = &intf->tmp_bmc;
> > ÂÂÂÂÂ INIT_LIST_HEAD(&intf->bmc->intfs);
> > ÂÂÂÂÂ mutex_init(&intf->bmc->dyn_mutex);
> > @@ -3514,7 +3523,7 @@ int ipmi_register_smi(const struct ipmi_smi_handlers *handlers,
> > ÂÂÂÂÂ return rv;
> > Â }
> > -EXPORT_SYMBOL(ipmi_register_smi);
> > +EXPORT_SYMBOL(ipmi_add_smi);
> > Â static void deliver_smi_err_response(struct ipmi_smi *intf,
> > ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ struct ipmi_smi_msg *msg,
> > diff --git a/include/linux/ipmi_smi.h b/include/linux/ipmi_smi.h
> > index 4dc66157d872..deec18b8944a 100644
> > --- a/include/linux/ipmi_smi.h
> > +++ b/include/linux/ipmi_smi.h
> > @@ -224,10 +224,14 @@ static inline int ipmi_demangle_device_id(uint8_t netfn, uint8_t cmd,
> > ÂÂ * is called, and the lower layer must get the interface from that
> > ÂÂ * call.
> > ÂÂ */
> > -int ipmi_register_smi(const struct ipmi_smi_handlers *handlers,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ voidÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ *send_info,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ struct deviceÂÂÂÂÂÂÂÂÂÂÂ *dev,
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂ unsigned charÂÂÂÂÂÂÂÂÂÂÂ slave_addr);
> > +int ipmi_add_smi(struct moduleÂÂÂÂÂÂÂÂÂÂÂ *owner,
> > +ÂÂÂÂÂÂÂÂ const struct ipmi_smi_handlers *handlers,
> > +ÂÂÂÂÂÂÂÂ voidÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ *send_info,
> > +ÂÂÂÂÂÂÂÂ struct deviceÂÂÂÂÂÂÂÂÂÂÂ *dev,
> > +ÂÂÂÂÂÂÂÂ unsigned charÂÂÂÂÂÂÂÂÂÂÂ slave_addr);
> > +
> > +#define ipmi_register_smi(handlers, send_info, dev, slave_addr) \
> > +ÂÂÂ ipmi_add_smi(THIS_MODULE, handlers, send_info, dev, slave_addr)
> > Â /*
> >  * Remove a low-level interface from the IPMI driver. This will
>
> Thanks, Corey.
>
> Regards,
> Tony

And I meant to add that I will be testing this over the next couple days.