netdev - Re: [PATCH net-next v4 1/5] netdevsim: maintain a list of probed netdevsims

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Thu, 21 Dec 2023 16:45:58 -0800
From: David Wei <dw@...idwei.uk>
To: Jiri Pirko <jiri@...nulli.us>
Cc: Jakub Kicinski <kuba@...nel.org>, Sabrina Dubroca <sd@...asysnail.net>,
 netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
 Eric Dumazet <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next v4 1/5] netdevsim: maintain a list of probed
 netdevsims

On 2023-12-20 00:57, Jiri Pirko wrote:
> Wed, Dec 20, 2023 at 02:47:43AM CET, dw@...idwei.uk wrote:
>> This patch adds a linked list nsim_dev_list of probed netdevsims, added
>> during nsim_drv_probe() and removed during nsim_drv_remove(). A mutex
>> nsim_dev_list_lock protects the list.
> 
> In the commit message, you should use imperative mood, command
> the codebase what to do:
> https://www.kernel.org/doc/html/v6.6/process/submitting-patches.html#describe-your-changes

Thanks, I didn't know about this. Will edit the commit messages.

> 
> 
>>
>> Signed-off-by: David Wei <dw@...idwei.uk>
>> ---
>> drivers/net/netdevsim/dev.c       | 17 +++++++++++++++++
>> drivers/net/netdevsim/netdevsim.h |  1 +
>> 2 files changed, 18 insertions(+)
>>
>> diff --git a/drivers/net/netdevsim/dev.c b/drivers/net/netdevsim/dev.c
>> index b4d3b9cde8bd..e30a12130e07 100644
>> --- a/drivers/net/netdevsim/dev.c
>> +++ b/drivers/net/netdevsim/dev.c
>> @@ -35,6 +35,9 @@
>>
>> #include "netdevsim.h"
>>
>> +static LIST_HEAD(nsim_dev_list);
>> +static DEFINE_MUTEX(nsim_dev_list_lock);
>> +
>> static unsigned int
>> nsim_dev_port_index(enum nsim_dev_port_type type, unsigned int port_index)
>> {
>> @@ -1531,6 +1534,7 @@ int nsim_drv_probe(struct nsim_bus_dev *nsim_bus_dev)
>> 				 nsim_bus_dev->initial_net, &nsim_bus_dev->dev);
>> 	if (!devlink)
>> 		return -ENOMEM;
>> +	mutex_lock(&nsim_dev_list_lock);
> 
> I don't follow. You claim you use this mutex to protect the list.
> a) why don't you use spin-lock?

I'm using a mutex unless I know (or someone else who knows better point
out) that a spinlock is better. It is simple, there are fewer gotchas,
and I anticipate actual contention here to be near 0. The
nsim_bus_dev_list is also protected by a mutex.

Is a spinlock better here and if so why?

> b) why don't don't you take the lock just for list manipulation?

Many code paths interact here, touching drivers and netdevs. There is an
ordering of locks being taken:

1. nsim_bus_dev->dev.mutex
2. devlink->lock
3. rtnl_lock

I was careful to avoid deadlocking by acquiring locks in the same order.
But looking at it again, I can reduce the critical section by acquiring
nsim_dev_list_lock after devlink->lock, thanks.

> 
> 
>> 	devl_lock(devlink);
>> 	nsim_dev = devlink_priv(devlink);
>> 	nsim_dev->nsim_bus_dev = nsim_bus_dev;
>> @@ -1544,6 +1548,7 @@ int nsim_drv_probe(struct nsim_bus_dev *nsim_bus_dev)
>> 	spin_lock_init(&nsim_dev->fa_cookie_lock);
>>
>> 	dev_set_drvdata(&nsim_bus_dev->dev, nsim_dev);
>> +	list_add(&nsim_dev->list, &nsim_dev_list);
>>
>> 	nsim_dev->vfconfigs = kcalloc(nsim_bus_dev->max_vfs,
>> 				      sizeof(struct nsim_vf_config),
>> @@ -1607,6 +1612,7 @@ int nsim_drv_probe(struct nsim_bus_dev *nsim_bus_dev)
>>
>> 	nsim_dev->esw_mode = DEVLINK_ESWITCH_MODE_LEGACY;
>> 	devl_unlock(devlink);
>> +	mutex_unlock(&nsim_dev_list_lock);
>> 	return 0;
>>
>> err_hwstats_exit:
>> @@ -1668,8 +1674,18 @@ void nsim_drv_remove(struct nsim_bus_dev *nsim_bus_dev)
>> {
>> 	struct nsim_dev *nsim_dev = dev_get_drvdata(&nsim_bus_dev->dev);
>> 	struct devlink *devlink = priv_to_devlink(nsim_dev);
>> +	struct nsim_dev *pos, *tmp;
>>
>> +	mutex_lock(&nsim_dev_list_lock);
>> 	devl_lock(devlink);
>> +
>> +	list_for_each_entry_safe(pos, tmp, &nsim_dev_list, list) {
>> +		if (pos == nsim_dev) {
>> +			list_del(&nsim_dev->list);
>> +			break;
>> +		}
>> +	}
>> +
>> 	nsim_dev_reload_destroy(nsim_dev);
>>
>> 	nsim_bpf_dev_exit(nsim_dev);
>> @@ -1681,6 +1697,7 @@ void nsim_drv_remove(struct nsim_bus_dev *nsim_bus_dev)
>> 	kfree(nsim_dev->vfconfigs);
>> 	kfree(nsim_dev->fa_cookie);
>> 	devl_unlock(devlink);
>> +	mutex_unlock(&nsim_dev_list_lock);
>> 	devlink_free(devlink);
>> 	dev_set_drvdata(&nsim_bus_dev->dev, NULL);
>> }
>> diff --git a/drivers/net/netdevsim/netdevsim.h b/drivers/net/netdevsim/netdevsim.h
>> index 028c825b86db..babb61d7790b 100644
>> --- a/drivers/net/netdevsim/netdevsim.h
>> +++ b/drivers/net/netdevsim/netdevsim.h
>> @@ -277,6 +277,7 @@ struct nsim_vf_config {
>>
>> struct nsim_dev {
>> 	struct nsim_bus_dev *nsim_bus_dev;
>> +	struct list_head list;
>> 	struct nsim_fib_data *fib_data;
>> 	struct nsim_trap_data *trap_data;
>> 	struct dentry *ddir;
>> -- 
>> 2.39.3
>>