linux-fpga.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Frank Rowand <frowand.list@gmail.com>
To: Alan Tull <atull@kernel.org>, Rob Herring <robh@kernel.org>
Cc: Pantelis Antoniou <pantelis.antoniou@konsulko.com>,
	Moritz Fischer <mdf@kernel.org>,
	"devicetree@vger.kernel.org" <devicetree@vger.kernel.org>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	linux-fpga@vger.kernel.org
Subject: Re: [RFC 1/2] of: overlay: add whitelist
Date: Wed, 29 Nov 2017 04:25:59 -0500	[thread overview]
Message-ID: <791d0e87-c3c4-faa6-6a55-9141be2ca0ac@gmail.com> (raw)
In-Reply-To: <CANk1AXQ756EeKDDWXT5W6T8CJqqoFR7pGNkL5BGPN8gYBneRfw@mail.gmail.com>

On 11/28/17 14:26, Alan Tull wrote:
> On Tue, Nov 28, 2017 at 9:15 AM, Rob Herring <robh@kernel.org> wrote:
>> On Mon, Nov 27, 2017 at 02:58:03PM -0600, Alan Tull wrote:
>>> Add simple whitelist.  When an overlay is submitted, if any target in
>>> the overlay is not in the whitelist, the overlay is rejected.  Drivers
>>> that support dynamic configuration can register their device node with:
>>>
>>>   int of_add_whitelist_node(struct device_node *np)
>>>
>>> and remove themselves with:
>>>
>>>   void of_remove_whitelist_node(struct device_node *np)
>>
>> I think these should be named for what they do, not how it is
>> implemented.
> 
> Sure, such as of_node_overlay_enable and of_node_overlay_disable?
of_allow_overlay_on_node(), of_disallow_overlay_on_node()?


> 
>>
>>>
>>> Signed-off-by: Alan Tull <atull@kernel.org>
>>> ---
>>>  drivers/of/overlay.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>  include/linux/of.h   | 12 +++++++++
>>>  2 files changed, 85 insertions(+)
>>>
>>> diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
>>> index c150abb..5f952a1 100644
>>> --- a/drivers/of/overlay.c
>>> +++ b/drivers/of/overlay.c
>>> @@ -21,6 +21,7 @@
>>>  #include <linux/slab.h>
>>>  #include <linux/err.h>
>>>  #include <linux/idr.h>
>>> +#include <linux/spinlock.h>
>>>
>>>  #include "of_private.h"
>>>
>>> @@ -646,6 +647,74 @@ static void free_overlay_changeset(struct overlay_changeset *ovcs)
>>>       kfree(ovcs);
>>>  }
>>>
>>> +/* lock for adding/removing device nodes to the whitelist */
>>> +static spinlock_t whitelist_lock;
>>> +
>>> +static struct list_head whitelist_list = LIST_HEAD_INIT(whitelist_list);
>>> +
>>> +struct dt_overlay_whitelist {
>>> +     struct device_node *np;
>>> +     struct list_head node;
>>> +};
>>
>> Can't we just add a flags bit in device_node.flags? That would be much
>> simpler.
> 
> Yes, much simpler.  Such as:
> 
> #define OF_OVERLAY_ENABLED     5 /* allow DT overlay targeting this node */
> 
>>
>>> +
>>> +int of_add_whitelist_node(struct device_node *np)
>>> +{
>>> +     unsigned long flags;
>>> +     struct dt_overlay_whitelist *wln;
>>> +
>>> +     wln = kzalloc(sizeof(*wln), GFP_KERNEL);
>>> +     if (!wln)
>>> +             return -ENOMEM;
>>> +
>>> +     wln->np = np;
>>> +
>>> +     spin_lock_irqsave(&whitelist_lock, flags);
>>> +     list_add(&wln->node, &whitelist_list);
>>> +     spin_unlock_irqrestore(&whitelist_lock, flags);
>>> +
>>> +     return 0;
>>> +}
>>> +EXPORT_SYMBOL_GPL(of_add_whitelist_node);
>>> +
>>> +void of_remove_whitelist_node(struct device_node *np)
>>> +{
>>> +     struct dt_overlay_whitelist *wln;
>>> +     unsigned long flags;
>>> +
>>> +     list_for_each_entry(wln, &whitelist_list, node) {
>>> +             if (np == wln->np) {
>>> +                     spin_lock_irqsave(&whitelist_lock, flags);
>>> +                     list_del(&wln->node);
>>> +                     spin_unlock_irqrestore(&whitelist_lock, flags);
>>> +                     kfree(wln);
>>> +                     return;
>>> +             }
>>> +     }
>>> +}
>>> +EXPORT_SYMBOL_GPL(of_remove_whitelist_node);
>>> +
>>> +static int of_check_whitelist(struct overlay_changeset *ovcs)
>>> +{
>>> +     struct dt_overlay_whitelist *wln;
>>> +     struct device_node *target;
>>> +     int i;
>>> +
>>> +     for (i = 0; i < ovcs->count; i++) {
>>> +             target = ovcs->fragments[i].target;
>>> +             if (!of_node_cmp(target->name, "__symbols__"))
>>> +                     continue;
>>> +
>>> +             list_for_each_entry(wln, &whitelist_list, node)
>>> +                     if (target == wln->np)
>>> +                             break;
>>> +
>>> +             if (target != wln->np)
>>> +                     return -ENODEV;
>>> +     }
>>> +
>>> +     return 0;
>>> +}
>>> +
>>>  /**
>>>   * of_overlay_apply() - Create and apply an overlay changeset
>>>   * @tree:    Expanded overlay device tree
>>> @@ -717,6 +786,10 @@ int of_overlay_apply(struct device_node *tree, int *ovcs_id)
>>>       if (ret)
>>>               goto err_free_overlay_changeset;
>>>
>>> +     ret = of_check_whitelist(ovcs);
>>> +     if (ret)
>>> +             goto err_free_overlay_changeset;
>>
>> This will break you until the next patch and breaks any other users. I
>> think this is now just the unittest as tilcdc overlay is getting
>> removed.
>>
>> You have to make this chunk the last patch in the series.
> 
> I'd rather squash the two patches.  In either case, the contents of
> second patch are dependent on stuff in char-misc-testing today, so it
> won't be able to apply yet on linux-next or anywhere else.
> 
> Thanks
> Alan
> 
>>
>> Rob
> 

  reply	other threads:[~2017-11-29  9:25 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-27 20:58 [RFC 0/2] of: Add whitelist Alan Tull
2017-11-27 20:58 ` [RFC 1/2] of: overlay: add whitelist Alan Tull
2017-11-28 15:15   ` Rob Herring
2017-11-28 19:26     ` Alan Tull
2017-11-29  9:25       ` Frank Rowand [this message]
2017-11-27 20:58 ` [RFC 2/2] fpga: of region: add of-fpga-region to whitelist Alan Tull
2017-11-29  9:20 ` [RFC 0/2] of: Add whitelist Frank Rowand
2017-11-29 13:31   ` Rob Herring
2017-11-29 16:11     ` Alan Tull
2017-11-30 12:46       ` Frank Rowand
2017-12-05 16:33         ` Alan Tull
2017-12-06 11:56           ` Frank Rowand
2017-12-07 19:22             ` Alan Tull
2017-11-30 12:18     ` Frank Rowand
2017-12-05 16:55       ` Alan Tull
2017-12-06 11:47         ` Frank Rowand
2017-11-29 22:47   ` Frank Rowand
2017-11-30 14:39     ` Rob Herring
2017-12-06 11:44       ` Frank Rowand
2017-11-30  0:58   ` Frank Rowand

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=791d0e87-c3c4-faa6-6a55-9141be2ca0ac@gmail.com \
    --to=frowand.list@gmail.com \
    --cc=atull@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-fpga@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mdf@kernel.org \
    --cc=pantelis.antoniou@konsulko.com \
    --cc=robh@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).