Re: [PATCH v1 2/2] of: Let of_for_each_phandle fallback to non-negative cell_count

[Marek Szyprowski <m.szyprowski@samsung.com>]

Hi Uwe,

On 17.09.2019 14:25, Uwe Kleine-König wrote:
> On Tue, Sep 17, 2019 at 11:25:46AM +0000, Peter Rosin wrote:
>> On 2019-09-17 12:13, Uwe Kleine-König wrote:
>>> Hello Geert,
>>>
>>> On Tue, Sep 17, 2019 at 11:40:25AM +0200, Geert Uytterhoeven wrote:
>>>> Hi Rob, Uwe,
>>>>
>>>> On Fri, Sep 13, 2019 at 11:58 PM Rob Herring <robh@kernel.org> wrote:
>>>>> On Sat, 24 Aug 2019 15:28:46 +0200, =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?=          wrote:
>>>>>> Referencing device tree nodes from a property allows to pass arguments.
>>>>>> This is for example used for referencing gpios. This looks as follows:
>>>>>>
>>>>>>        gpio_ctrl: gpio-controller {
>>>>>>                #gpio-cells = <2>
>>>>>>                ...
>>>>>>        }
>>>>>>
>>>>>>        someothernode {
>>>>>>                gpios = <&gpio_ctrl 5 0 &gpio_ctrl 3 0>;
>>>>>>                ...
>>>>>>        }
>>>>>>
>>>>>> To know the number of arguments this must be either fixed, or the
>>>>>> referenced node is checked for a $cells_name (here: "#gpio-cells")
>>>>>> property and with this information the start of the second reference can
>>>>>> be determined.
>>>>>>
>>>>>> Currently regulators are referenced with no additional arguments. To
>>>>>> allow some optional arguments without having to change all referenced
>>>>>> nodes this change introduces a way to specify a default cell_count. So
>>>>>> when a phandle is parsed we check for the $cells_name property and use
>>>>>> it as before if present. If it is not present we fall back to
>>>>>> cells_count if non-negative and only fail if cells_count is smaller than
>>>>>> zero.
>>>>>>
>>>>>> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>>>> This is now commit e42ee61017f58cd9 ("of: Let of_for_each_phandle fallback
>>>> to non-negative cell_count") in robh/for-next, which causes a lock-up when
>>>> booting a shmobile_defconfig kernel on r8a7791/koelsch:
>>>>
>>>> rcu: INFO: rcu_sched self-detected stall on CPU
>>>> rcu:     0-....: (2099 ticks this GP) idle=6fe/1/0x40000002
>>>> softirq=29/29 fqs=1050
>>>>   (t=2100 jiffies g=-1131 q=0)
>>>> NMI backtrace for cpu 0
>>>> CPU: 0 PID: 1 Comm: swapper/0 Not tainted
>>>> 5.3.0-rc2-shmobile-00050-ge42ee61017f58cd9 #376
>>>> Hardware name: Generic R-Car Gen2 (Flattened Device Tree)
>>>> [<c010f8ac>] (unwind_backtrace) from [<c010b620>] (show_stack+0x10/0x14)
>>>> [<c010b620>] (show_stack) from [<c073d038>] (dump_stack+0x7c/0x9c)
>>>> [<c073d038>] (dump_stack) from [<c0742e80>] (nmi_cpu_backtrace+0xa0/0xb8)
>>>> [<c0742e80>] (nmi_cpu_backtrace) from [<c0742f1c>] (nmi_trigger_cpumask_backtrace+0x84/0x114)
>>>> [<c0742f1c>] (nmi_trigger_cpumask_backtrace) from [<c017d684>] (rcu_dump_cpu_stacks+0xac/0xc8)
>>>> [<c017d684>] (rcu_dump_cpu_stacks) from [<c017a598>] (rcu_sched_clock_irq+0x2ac/0x6b4)
>>>> [<c017a598>] (rcu_sched_clock_irq) from [<c0183980>] (update_process_times+0x30/0x5c)
>>>> [<c0183980>] (update_process_times) from [<c01941a8>] (tick_nohz_handler+0xcc/0x120)
>>>> [<c01941a8>] (tick_nohz_handler) from [<c05b1d40>] (arch_timer_handler_virt+0x28/0x30)
>>>> [<c05b1d40>] (arch_timer_handler_virt) from [<c016c9e0>] (handle_percpu_devid_irq+0xe8/0x21c)
>>>> [<c016c9e0>] (handle_percpu_devid_irq) from [<c0167a8c>] (generic_handle_irq+0x18/0x28)
>>>> [<c0167a8c>] (generic_handle_irq) from [<c0167b3c>] (__handle_domain_irq+0xa0/0xb4)
>>>> [<c0167b3c>] (__handle_domain_irq) from [<c03673ec>] (gic_handle_irq+0x58/0x90)
>>>> [<c03673ec>] (gic_handle_irq) from [<c0101a8c>] (__irq_svc+0x6c/0x90)
>>>> Exception stack(0xeb08dd30 to 0xeb08dd78)
>>>> dd20:                                     c0cc7514 20000013 00000005 00003b27
>>>> dd40: eb7c4020 c0cc750c 00000051 00000051 20000013 c0c66b08 eb1cdc00 00000018
>>>> dd60: 00000000 eb08dd80 c05c1a38 c0756c00 20000013 ffffffff
>>>> [<c0101a8c>] (__irq_svc) from [<c0756c00>] (_raw_spin_unlock_irqrestore+0x1c/0x20)
>>>> [<c0756c00>] (_raw_spin_unlock_irqrestore) from [<c05c1a38>] (of_find_node_by_phandle+0xcc/0xf0)
>>>> [<c05c1a38>] (of_find_node_by_phandle) from [<c05c1bb8>] (of_phandle_iterator_next+0x68/0x178)
>>>> [<c05c1bb8>] (of_phandle_iterator_next) from [<c05c22bc>] (of_count_phandle_with_args+0x5c/0x7c)
>>>> [<c05c22bc>] (of_count_phandle_with_args) from [<c053fc38>] (i2c_demux_pinctrl_probe+0x24/0x1fc)
>>>> [<c053fc38>] (i2c_demux_pinctrl_probe) from [<c04463c4>] (platform_drv_probe+0x48/0x94)
>>>> [<c04463c4>] (platform_drv_probe) from [<c0444a20>] (really_probe+0x1f0/0x2b8)
>>>> [<c0444a20>] (really_probe) from [<c0444e68>] (driver_probe_device+0x140/0x158)
>>>> [<c0444e68>] (driver_probe_device) from [<c0444ff0>] (device_driver_attach+0x44/0x5c)
>>>> [<c0444ff0>] (device_driver_attach) from [<c04450b4>] (__driver_attach+0xac/0xb4)
>>>> [<c04450b4>] (__driver_attach) from [<c0443178>] (bus_for_each_dev+0x64/0xa0)
>>>> [<c0443178>] (bus_for_each_dev) from [<c04438a8>] (bus_add_driver+0x148/0x1a8)
>>>> [<c04438a8>] (bus_add_driver) from [<c0445ad0>] (driver_register+0xac/0xf0)
>>>> [<c0445ad0>] (driver_register) from [<c0b010b0>] (do_one_initcall+0xa8/0x1d4)
>>>> [<c0b010b0>] (do_one_initcall) from [<c0b01448>] (kernel_init_freeable+0x26c/0x2c8)
>>>> [<c0b01448>] (kernel_init_freeable) from [<c0751c70>] (kernel_init+0x8/0x10c)
>>>> [<c0751c70>] (kernel_init) from [<c01010e8>] (ret_from_fork+0x14/0x2c)
>>>> Exception stack(0xeb08dfb0 to 0xeb08dff8)
>>>> dfa0:                                     00000000 00000000 00000000 00000000
>>>> dfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
>>>> dfe0: 00000000 00000000 00000000 00000000 00000013 00000000
>>>>
>>>> Presumably it loops forever, due to a conversion of -1 to unsigned
>>>> somewhere?
>>> Hmm, I fail to see the culprit. i2c_demux_pinctrl_probe calls
>>> of_count_phandle_with_args with cells_name=NULL. With that I don't see
>>> how my patch changes anything as the only change is in an if
>>> (it->cells_name) block that shouldn't be relevant in your case.
>>>
>>> Can you please verify that the loop in of_count_phandle_with_args is
>>> indeed not terminating, e.g. with
>> The below indicated else-branch was not touched by e42ee61017f58cd9,
>> which ends up setting the count to -1 (aka 0xff...ff in this case).
>> No?
>>
>> int of_phandle_iterator_next(struct of_phandle_iterator *it)
>> {
>>
>> 	...
>>
>> 		if (it->cells_name) {
>>
>> 			...
>>
>> 		} else {
>> 			count = it->cell_count;    /* <---- SUSPECT!!! */
>> 		}
> Oh yeah, you're right. I'm a bit disappointed that I didn't spot this
> myself :-|
>
> Untested patch to fix this problem:

Yesterday I've noticed that sound driver fails to initialize on TM2(e) 
board (arm64) and today I've bisected to this commit. Nice to see that 
the issue has been already investigated.

> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index 2f25d2dfecfa..26f7a21d7187 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -1284,6 +1284,13 @@ int of_phandle_iterator_init(struct of_phandle_iterator *it,
>   	const __be32 *list;
>   	int size;
>   
> +	/*
> +	 * one of cell_count or cells_name must be provided to determine the
> +	 * argument length.
> +	 */
> +	if (cell_count < 0 && !cells_name)
> +		return -EINVAL;
> +
>   	memset(it, 0, sizeof(*it));
>   
>   	list = of_get_property(np, list_name, &size);
> @@ -1765,6 +1772,18 @@ int of_count_phandle_with_args(const struct device_node *np, const char *list_na
>   	struct of_phandle_iterator it;
>   	int rc, cur_index = 0;
>   
> +	/* If cells_name is NULL we assume an cell_count of 0 */
> +	if (cells_name == NULL) {
> +		const __be32 *list;
> +		int size;
> +
> +		list = of_get_property(np, list_name, &size);
> +		if (!list)
> +			return -ENOENT;
> +
> +		return size / sizeof(*list);
> +	}
> +
>   	rc = of_phandle_iterator_init(&it, np, list_name, cells_name, -1);
>   	if (rc)
>   		return rc;
>
I confirm that the above code works. The patch fixes my TM2(e) sound 
issue, feel free to add:

Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu