linux-kernel - Re: [PATCH] drm: Fix possible memleak in drm_client_modeset

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]

Message-ID: <e082f0fa-d524-ae4f-053d-48d817d2c7d4@huawei.com>
Date:   Tue, 9 Nov 2021 10:06:40 +0800
From:   He Ying <heying24@...wei.com>
To:     <maarten.lankhorst@...ux.intel.com>, <mripard@...nel.org>,
        <tzimmermann@...e.de>, <airlied@...ux.ie>, <daniel@...ll.ch>
CC:     <dri-devel@...ts.freedesktop.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm: Fix possible memleak in drm_client_modeset_probe

kindly ping...

在 2021/10/22 9:26, He Ying 写道:
> I got memory leak as follows when doing fault injection test:
>
> WARNING: CPU: 0 PID: 1214 at drm_mode_config_cleanup+0x689/0x890 [drm]
> RIP: 0010:drm_mode_config_cleanup+0x689/0x890 [drm]
> Call Trace:
>   ? tracer_hardirqs_on+0x33/0x520
>   ? drm_mode_config_reset+0x3f0/0x3f0 [drm]
>   ? trace_event_raw_event_rcu_torture_read+0x2c0/0x2c0
>   ? __sanitizer_cov_trace_pc+0x1d/0x50
>   ? call_rcu+0x489/0x15e0
>   ? trace_hardirqs_on+0x63/0x2d0
>   ? write_comp_data+0x2a/0x90
>   ? drm_mode_config_cleanup+0x890/0x890 [drm]
>   drm_managed_release+0x1fa/0x4f0 [drm]
>   drm_dev_release+0x72/0xb0 [drm]
>   devm_drm_dev_init_release+0x81/0xa0 [drm]
>   release_nodes+0xba/0x3b0
>   ...
>   entry_SYSCALL_64_after_hwframe+0x44/0xae
>
> and then an error message:
>
> [drm:drm_mode_config_cleanup [drm]] *ERROR* connector SPI-1 leaked!
>
> When krealloc() in drm_client_modeset_probe() fails, it
> goes to the label 'free_connectors'. However, krealloc()
> is between drm_connector_list_iter_begin() and *_end().
> Going to the label directly is not a good idea. Because
> iter->conn is not cleaned up if so. Fix the problem by
> going to the label outside drm_connector_list_iter_end().
>
> Signed-off-by: He Ying <heying24@...wei.com>
> ---
>   drivers/gpu/drm/drm_client_modeset.c | 5 ++++-
>   1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/drm_client_modeset.c b/drivers/gpu/drm/drm_client_modeset.c
> index ced09c7c06f9..7ac88ae93f38 100644
> --- a/drivers/gpu/drm/drm_client_modeset.c
> +++ b/drivers/gpu/drm/drm_client_modeset.c
> @@ -789,7 +789,7 @@ int drm_client_modeset_probe(struct drm_client_dev *client, unsigned int width,
>   		tmp = krealloc(connectors, (connector_count + 1) * sizeof(*connectors), GFP_KERNEL);
>   		if (!tmp) {
>   			ret = -ENOMEM;
> -			goto free_connectors;
> +			break;
>   		}
>   
>   		connectors = tmp;
> @@ -798,6 +798,9 @@ int drm_client_modeset_probe(struct drm_client_dev *client, unsigned int width,
>   	}
>   	drm_connector_list_iter_end(&conn_iter);
>   
> +	if (ret)
> +		goto free_connectors;
> +
>   	if (!connector_count)
>   		return 0;
>