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: <20211109123140.GA5208@Peter>
Date:   Tue, 9 Nov 2021 20:31:40 +0800
From:   Peter Chen <peter.chen@...nel.org>
To:     Qihang Hu <huqihang@...o.com>
Cc:     balbi@...nel.org, gregkh@...uxfoundation.org,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] usb: gadget: composite: Show warning if function
 driver's descriptors are incomplete.

On 21-11-09 18:19:36, Qihang Hu wrote:
> In the config_ep_by_speed_and_alt function, select the corresponding
> descriptor through g->speed. But some unsound function drivers may

But some 'legacy or not well designed' function drivers

> not support the corresponding speed. So, we can directly display
> warnings instead of panicking the kernel.

instead of 'causing kernel panic'

> At the same time, it
> indicates a problem with the function in the warning message.

It indicates the reasons in warning message.
> 
> Signed-off-by: Qihang Hu <huqihang@...o.com>
> ---
> Changes in v2:
> -Add warning message
> 
> Changes in v3:
> -Change commit log
> -Delete extra blank lines
> -Modify 'incomplete_desc' to bool type

The latest changelog should be showed at the first.

> ---
>  drivers/usb/gadget/composite.c | 39 ++++++++++++++++++++++------------
>  1 file changed, 26 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
> index 72a9797dbbae..cb9c7edf9bbf 100644
> --- a/drivers/usb/gadget/composite.c
> +++ b/drivers/usb/gadget/composite.c
> @@ -159,6 +159,8 @@ int config_ep_by_speed_and_alt(struct usb_gadget *g,
>  	int want_comp_desc = 0;
>  
>  	struct usb_descriptor_header **d_spd; /* cursor for speed desc */
> +	struct usb_composite_dev *cdev;
> +	bool incomplete_desc = false;
>  
>  	if (!g || !f || !_ep)
>  		return -EIO;
> @@ -167,28 +169,43 @@ int config_ep_by_speed_and_alt(struct usb_gadget *g,
>  	switch (g->speed) {
>  	case USB_SPEED_SUPER_PLUS:
>  		if (gadget_is_superspeed_plus(g)) {
> -			speed_desc = f->ssp_descriptors;
> -			want_comp_desc = 1;
> -			break;
> +			if (f->ssp_descriptors) {
> +				speed_desc = f->ssp_descriptors;
> +				want_comp_desc = 1;
> +				break;
> +			}
> +			incomplete_desc = true;
>  		}
>  		fallthrough;
>  	case USB_SPEED_SUPER:
>  		if (gadget_is_superspeed(g)) {
> -			speed_desc = f->ss_descriptors;
> -			want_comp_desc = 1;
> -			break;
> +			if (f->ss_descriptors) {
> +				speed_desc = f->ss_descriptors;
> +				want_comp_desc = 1;
> +				break;
> +			}
> +			incomplete_desc = true;
>  		}
>  		fallthrough;
>  	case USB_SPEED_HIGH:
>  		if (gadget_is_dualspeed(g)) {
> -			speed_desc = f->hs_descriptors;
> -			break;
> +			if (f->hs_descriptors) {
> +				speed_desc = f->hs_descriptors;
> +				break;
> +			}
> +			incomplete_desc = true;
>  		}
>  		fallthrough;
>  	default:
>  		speed_desc = f->fs_descriptors;
>  	}
>  
> +	cdev = get_gadget_data(g);
> +	if (incomplete_desc)
> +		WARNING(cdev,
> +			"%s doesn't hold the descriptors for current speed\n",
> +			f->name);
> +
>  	/* find correct alternate setting descriptor */
>  	for_each_desc(speed_desc, d_spd, USB_DT_INTERFACE) {
>  		int_desc = (struct usb_interface_descriptor *)*d_spd;
> @@ -244,12 +261,8 @@ int config_ep_by_speed_and_alt(struct usb_gadget *g,
>  			_ep->maxburst = comp_desc->bMaxBurst + 1;
>  			break;
>  		default:
> -			if (comp_desc->bMaxBurst != 0) {
> -				struct usb_composite_dev *cdev;
> -
> -				cdev = get_gadget_data(g);
> +			if (comp_desc->bMaxBurst != 0)
>  				ERROR(cdev, "ep0 bMaxBurst must be 0\n");
> -			}
>  			_ep->maxburst = 1;
>  			break;
>  		}
> -- 
> 2.25.1
> 

Except the typo issues, other things are ok for me. It shows an warning
message for not well designed function driver, instead of panic
the kernel. It depends on Greg that whether it should be panic directly
or show warning message to indicate the issue.

-- 

Thanks,
Peter Chen

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ