linux-kernel - Re: [PATCH v3] staging: vc04_services: vchiq

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] [day] [month] [year] [list]
Date: Mon, 24 Jun 2024 12:44:11 +0530
From: Umang Jain <umang.jain@...asonboard.com>
To: Kieran Bingham <kieran.bingham@...asonboard.com>,
 Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
 Florian Fainelli <florian.fainelli@...adcom.com>,
 linux-rpi-kernel@...ts.infradead.org, Stefan Wahren <wahrenst@....net>
Cc: Broadcom internal kernel review list
 <bcm-kernel-feedback-list@...adcom.com>,
 Laurent Pinchart <laurent.pinchart@...asonboard.com>,
 Dan Carpenter <dan.carpenter@...aro.org>,
 Yang Li <yang.lee@...ux.alibaba.com>,
 "moderated list:BROADCOM BCM2711/BCM2835 ARM ARCHITECTURE"
 <linux-arm-kernel@...ts.infradead.org>,
 "open list:STAGING SUBSYSTEM" <linux-staging@...ts.linux.dev>,
 open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3] staging: vc04_services: vchiq_arm: Fix initialisation
 check

Hi Kieran,

On 21/06/24 3:40 am, Kieran Bingham wrote:
> The vchiq_state used to be obtained through an accessor which would
> validate that the VCHIQ had been initialised correctly with the remote,
> or return a null state.
>
> In commit 42a2f6664e18 ("staging: vc04_services: Move global g_state to
> vchiq_state") the global state was moved to the vchiq_mgnt structures
> stored as a vchiq instance specific context. This conversion removed the
> helpers and instead replaced users of this helper with the assumption
> that the state is always available and the remote connected.
>
> The conversion does ensure that the state is always available, so some
> remaining state null pointer checks that remain are unnecessary, but the
> assumption that the remote is present and initialised is incorrect.
>
> Fix this broken assumption by re-introducing the logic that was lost
> during the conversion.

Yes, the logic was broken. thanks for noticing this.
>
> Fixes: 42a2f6664e18 ("staging: vc04_services: Move global g_state to vchiq_state")
> Signed-off-by: Kieran Bingham <kieran.bingham@...asonboard.com>

Reviewed-by: Umang Jain <umang.jain@...asonboard.com>
>
> ---
> v2:
>   - Just a resend
>
> v3:
>   - Downgrade vchiq_open() error path print to dbg
>   - Clarify commit message about unnecessary state checks.
>
>   .../staging/vc04_services/interface/vchiq_arm/vchiq_arm.c  | 4 ++--
>   .../staging/vc04_services/interface/vchiq_arm/vchiq_core.h | 5 +++++
>   .../staging/vc04_services/interface/vchiq_arm/vchiq_dev.c  | 7 ++++++-
>   3 files changed, 13 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> index 69daeba974f2..5f518e5a9273 100644
> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> @@ -707,7 +707,7 @@ int vchiq_initialise(struct vchiq_state *state, struct vchiq_instance **instance
>   	 * block forever.
>   	 */
>   	for (i = 0; i < VCHIQ_INIT_RETRIES; i++) {
> -		if (state)
> +		if (vchiq_remote_initialised(state))
>   			break;
>   		usleep_range(500, 600);
>   	}
> @@ -1202,7 +1202,7 @@ void vchiq_dump_platform_instances(struct vchiq_state *state, struct seq_file *f
>   {
>   	int i;
>   
> -	if (!state)
> +	if (!vchiq_remote_initialised(state))
>   		return;
>   
>   	/*
> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.h b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.h
> index 8af209e34fb2..382ec08f6a14 100644
> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.h
> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.h
> @@ -413,6 +413,11 @@ struct vchiq_state {
>   	struct opaque_platform_state *platform_state;
>   };
>   
> +static inline bool vchiq_remote_initialised(const struct vchiq_state *state)
> +{
> +	return state->remote && state->remote->initialised;
> +}
> +
>   struct bulk_waiter {
>   	struct vchiq_bulk *bulk;
>   	struct completion event;
> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_dev.c b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_dev.c
> index 67ba9ceaad3e..9cd2a64dce5e 100644
> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_dev.c
> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_dev.c
> @@ -1170,6 +1170,11 @@ static int vchiq_open(struct inode *inode, struct file *file)
>   
>   	dev_dbg(state->dev, "arm: vchiq open\n");
>   
> +	if (!vchiq_remote_initialised(state)) {
> +		dev_dbg(state->dev, "arm: vchiq has no connection to VideoCore\n");
> +		return -ENOTCONN;
> +	}
> +
>   	instance = kzalloc(sizeof(*instance), GFP_KERNEL);
>   	if (!instance)
>   		return -ENOMEM;
> @@ -1200,7 +1205,7 @@ static int vchiq_release(struct inode *inode, struct file *file)
>   
>   	dev_dbg(state->dev, "arm: instance=%p\n", instance);
>   
> -	if (!state) {
> +	if (!vchiq_remote_initialised(state)) {
>   		ret = -EPERM;
>   		goto out;
>   	}