| 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
| ||
|
Message-ID: <20200810133450.GD5493@kadam>
Date: Mon, 10 Aug 2020 16:34:51 +0300
From: Dan Carpenter <dan.carpenter@...cle.com>
To: Christoph Hellwig <hch@....de>
Cc: kernel test robot <lkp@...el.com>, kbuild-all@...ts.01.org,
linux-kernel@...r.kernel.org,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: drivers/usb/core/devio.c:1155 do_proc_control() error:
copy_from_user() 'tbuf' too small (4096 vs 8192)
On Mon, Aug 10, 2020 at 02:14:33PM +0200, Christoph Hellwig wrote:
> On Mon, Aug 10, 2020 at 02:12:23PM +0200, Christoph Hellwig wrote:
> > As far as I can tell the warning is valid as we copy a user controlled
> > amount into a fixed sized buffer. But this an old condition not actually
> > created by this commit..
>
> Actually, is is not. do_proc_control checks the legnth, but does it
> using a somewhat pointless local variable, which might have confused
> smatch.
I think what's happening is that this is using PAGE_SIZE = 8096 and
smatch assumes that it's always 4096. It seems obvious in retrospect
that you can't assume page size if 4096 but I'm not sure how to
calculate the actual page size...
Normally I filter all Smatch kbuild bot warnings manually before
forwarding them but this one was sent directly.
regards,
dan carpenter
>
> >
> > On Sat, Aug 08, 2020 at 01:26:39PM +0800, kernel test robot wrote:
> > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > > head: 049eb096da48db0421dd5e358b9b082a1a8a2025
> > > commit: c17536d0abde2fd24afca542e3bb73b45a299633 usb: usbfs: stop using compat_alloc_user_space
> > > date: 2 weeks ago
> > > config: nds32-randconfig-m031-20200808 (attached as .config)
> > > compiler: nds32le-linux-gcc (GCC) 9.3.0
> > >
> > > If you fix the issue, kindly add following tag as appropriate
> > > Reported-by: kernel test robot <lkp@...el.com>
> > >
> > > smatch warnings:
> > > drivers/usb/core/devio.c:1155 do_proc_control() error: copy_from_user() 'tbuf' too small (4096 vs 8192)
> > >
> > > vim +/tbuf +1155 drivers/usb/core/devio.c
> > >
> > > 1104
> > > 1105 static int do_proc_control(struct usb_dev_state *ps,
> > > 1106 struct usbdevfs_ctrltransfer *ctrl)
> > > 1107 {
> > > 1108 struct usb_device *dev = ps->dev;
> > > 1109 unsigned int tmo;
> > > 1110 unsigned char *tbuf;
> > > 1111 unsigned wLength;
> > > 1112 int i, pipe, ret;
> > > 1113
> > > 1114 ret = check_ctrlrecip(ps, ctrl->bRequestType, ctrl->bRequest,
> > > 1115 ctrl->wIndex);
> > > 1116 if (ret)
> > > 1117 return ret;
> > > 1118 wLength = ctrl->wLength; /* To suppress 64k PAGE_SIZE warning */
> > > 1119 if (wLength > PAGE_SIZE)
> > > 1120 return -EINVAL;
> > > 1121 ret = usbfs_increase_memory_usage(PAGE_SIZE + sizeof(struct urb) +
> > > 1122 sizeof(struct usb_ctrlrequest));
> > > 1123 if (ret)
> > > 1124 return ret;
> > > 1125 tbuf = (unsigned char *)__get_free_page(GFP_KERNEL);
> > > 1126 if (!tbuf) {
> > > 1127 ret = -ENOMEM;
> > > 1128 goto done;
> > > 1129 }
> > > 1130 tmo = ctrl->timeout;
> > > 1131 snoop(&dev->dev, "control urb: bRequestType=%02x "
> > > 1132 "bRequest=%02x wValue=%04x "
> > > 1133 "wIndex=%04x wLength=%04x\n",
> > > 1134 ctrl->bRequestType, ctrl->bRequest, ctrl->wValue,
> > > 1135 ctrl->wIndex, ctrl->wLength);
> > > 1136 if (ctrl->bRequestType & 0x80) {
> > > 1137 pipe = usb_rcvctrlpipe(dev, 0);
> > > 1138 snoop_urb(dev, NULL, pipe, ctrl->wLength, tmo, SUBMIT, NULL, 0);
> > > 1139
> > > 1140 usb_unlock_device(dev);
> > > 1141 i = usb_control_msg(dev, pipe, ctrl->bRequest,
> > > 1142 ctrl->bRequestType, ctrl->wValue, ctrl->wIndex,
> > > 1143 tbuf, ctrl->wLength, tmo);
> > > 1144 usb_lock_device(dev);
> > > 1145 snoop_urb(dev, NULL, pipe, max(i, 0), min(i, 0), COMPLETE,
> > > 1146 tbuf, max(i, 0));
> > > 1147 if ((i > 0) && ctrl->wLength) {
> > > 1148 if (copy_to_user(ctrl->data, tbuf, i)) {
> > > 1149 ret = -EFAULT;
> > > 1150 goto done;
> > > 1151 }
> > > 1152 }
> > > 1153 } else {
> > > 1154 if (ctrl->wLength) {
> > > > 1155 if (copy_from_user(tbuf, ctrl->data, ctrl->wLength)) {
> > > 1156 ret = -EFAULT;
> > > 1157 goto done;
> > > 1158 }
> > > 1159 }
> > > 1160 pipe = usb_sndctrlpipe(dev, 0);
> > > 1161 snoop_urb(dev, NULL, pipe, ctrl->wLength, tmo, SUBMIT,
> > > 1162 tbuf, ctrl->wLength);
> > > 1163
> > > 1164 usb_unlock_device(dev);
> > > 1165 i = usb_control_msg(dev, usb_sndctrlpipe(dev, 0), ctrl->bRequest,
> > > 1166 ctrl->bRequestType, ctrl->wValue, ctrl->wIndex,
> > > 1167 tbuf, ctrl->wLength, tmo);
> > > 1168 usb_lock_device(dev);
> > > 1169 snoop_urb(dev, NULL, pipe, max(i, 0), min(i, 0), COMPLETE, NULL, 0);
> > > 1170 }
> > > 1171 if (i < 0 && i != -EPIPE) {
> > > 1172 dev_printk(KERN_DEBUG, &dev->dev, "usbfs: USBDEVFS_CONTROL "
> > > 1173 "failed cmd %s rqt %u rq %u len %u ret %d\n",
> > > 1174 current->comm, ctrl->bRequestType, ctrl->bRequest,
> > > 1175 ctrl->wLength, i);
> > > 1176 }
> > > 1177 ret = i;
> > > 1178 done:
> > > 1179 free_page((unsigned long) tbuf);
> > > 1180 usbfs_decrease_memory_usage(PAGE_SIZE + sizeof(struct urb) +
> > > 1181 sizeof(struct usb_ctrlrequest));
> > > 1182 return ret;
> > > 1183 }
> > > 1184
> > >
> > > ---
> > > 0-DAY CI Kernel Test Service, Intel Corporation
> > > https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
> >
> >
> > ---end quoted text---
> ---end quoted text---
Powered by blists - more mailing lists