Re: [PATCH v2] i2c: virtio: add a virtio i2c frontend driver
From: Andy Shevchenko
Date: Mon Sep 14 2020 - 12:33:18 EST
On Mon, Sep 14, 2020 at 06:47:27PM +0300, Dan Carpenter wrote:
> On Mon, Sep 14, 2020 at 06:24:55PM +0300, Andy Shevchenko wrote:
> > On Mon, Sep 14, 2020 at 05:48:07PM +0300, Dan Carpenter wrote:
> > > Hi Jie,
> > >
> > > url: https://github.com/0day-ci/linux/commits/Jie-Deng/i2c-virtio-add-a-virtio-i2c-frontend-driver/20200911-115013
> > > base: https://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git i2c/for-next
> > > config: parisc-randconfig-m031-20200913 (attached as .config)
> > > compiler: hppa-linux-gcc (GCC) 9.3.0
> > >
> > > If you fix the issue, kindly add following tag as appropriate
> > > Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > > Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> > >
> > > smatch warnings:
> > > drivers/i2c/busses/i2c-virtio.c:160 virtio_i2c_xfer() error: we previously assumed 'vmsg' could be null (see line 137)
> > >
> >
> > It's quite possible a false positive. Look at 122. But I agree that for-loop is
> > not the best for such things to understand. Perhaps switching to do {} while ()
> > will make it better.
> >
>
> Smatch is assuming that virtqueue_get_buf() can return NULL on the last
> iteration through the loop.
I see now. Thanks.
> > > # https://github.com/0day-ci/linux/commit/0a54ec771966748fcbc86256b830b5f786168b7d
> > > git remote add linux-review https://github.com/0day-ci/linux
> > > git fetch --no-tags linux-review Jie-Deng/i2c-virtio-add-a-virtio-i2c-frontend-driver/20200911-115013
> > > git checkout 0a54ec771966748fcbc86256b830b5f786168b7d
> > > vim +/vmsg +160 drivers/i2c/busses/i2c-virtio.c
> > >
> > > 0a54ec77196674 Jie Deng 2020-09-11 109 static int virtio_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
> > > 0a54ec77196674 Jie Deng 2020-09-11 110 {
> > > 0a54ec77196674 Jie Deng 2020-09-11 111 struct virtio_i2c *vi = i2c_get_adapdata(adap);
> > > 0a54ec77196674 Jie Deng 2020-09-11 112 struct virtqueue *vq = vi->vq;
> > > 0a54ec77196674 Jie Deng 2020-09-11 113 struct virtio_i2c_msg *vmsg;
> > > 0a54ec77196674 Jie Deng 2020-09-11 114 unsigned long time_left;
> > > 0a54ec77196674 Jie Deng 2020-09-11 115 int len, i, ret = 0;
> > > 0a54ec77196674 Jie Deng 2020-09-11 116
> > > 0a54ec77196674 Jie Deng 2020-09-11 117 mutex_lock(&vi->i2c_lock);
> > > 0a54ec77196674 Jie Deng 2020-09-11 118 vmsg = &vi->vmsg;
> > > 0a54ec77196674 Jie Deng 2020-09-11 119 vmsg->buf = NULL;
> > > 0a54ec77196674 Jie Deng 2020-09-11 120
> > > 0a54ec77196674 Jie Deng 2020-09-11 121 for (i = 0; i < num; i++) {
> > > 0a54ec77196674 Jie Deng 2020-09-11 122 ret = virtio_i2c_add_msg(vq, vmsg, &msgs[i]);
> > > 0a54ec77196674 Jie Deng 2020-09-11 123 if (ret) {
> > > 0a54ec77196674 Jie Deng 2020-09-11 124 dev_err(&adap->dev, "failed to add msg[%d] to virtqueue.\n", i);
> > > 0a54ec77196674 Jie Deng 2020-09-11 125 break;
> > > 0a54ec77196674 Jie Deng 2020-09-11 126 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 127
> > > 0a54ec77196674 Jie Deng 2020-09-11 128 virtqueue_kick(vq);
> > > 0a54ec77196674 Jie Deng 2020-09-11 129
> > > 0a54ec77196674 Jie Deng 2020-09-11 130 time_left = wait_for_completion_timeout(&vi->completion, adap->timeout);
> > > 0a54ec77196674 Jie Deng 2020-09-11 131 if (!time_left) {
> > > 0a54ec77196674 Jie Deng 2020-09-11 132 dev_err(&adap->dev, "msg[%d]: addr=0x%x timeout.\n", i, msgs[i].addr);
> > > 0a54ec77196674 Jie Deng 2020-09-11 133 break;
> > > 0a54ec77196674 Jie Deng 2020-09-11 134 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 135
> > > 0a54ec77196674 Jie Deng 2020-09-11 136 vmsg = (struct virtio_i2c_msg *)virtqueue_get_buf(vq, &len);
> > > 0a54ec77196674 Jie Deng 2020-09-11 @137 if (vmsg) {
> > > ^^^^
> > > Check for NULL.
> > >
> > > 0a54ec77196674 Jie Deng 2020-09-11 138 /* vmsg should point to the same address with &vi->vmsg */
> > > 0a54ec77196674 Jie Deng 2020-09-11 139 if (vmsg != &vi->vmsg) {
> > > 0a54ec77196674 Jie Deng 2020-09-11 140 dev_err(&adap->dev, "msg[%d]: addr=0x%x virtqueue error.\n",
> > > 0a54ec77196674 Jie Deng 2020-09-11 141 i, le16_to_cpu(vmsg->hdr.addr));
> > > 0a54ec77196674 Jie Deng 2020-09-11 142 break;
> > > 0a54ec77196674 Jie Deng 2020-09-11 143 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 144 if (vmsg->status != VIRTIO_I2C_MSG_OK) {
> > > 0a54ec77196674 Jie Deng 2020-09-11 145 dev_err(&adap->dev, "msg[%d]: addr=0x%x error=%d.\n",
> > > 0a54ec77196674 Jie Deng 2020-09-11 146 i, le16_to_cpu(vmsg->hdr.addr), vmsg->status);
> > > 0a54ec77196674 Jie Deng 2020-09-11 147 break;
> > > 0a54ec77196674 Jie Deng 2020-09-11 148 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 149 if ((msgs[i].flags & I2C_M_RD) && msgs[i].len)
> > > 0a54ec77196674 Jie Deng 2020-09-11 150 memcpy(msgs[i].buf, vmsg->buf, msgs[i].len);
> > > 0a54ec77196674 Jie Deng 2020-09-11 151
> > > 0a54ec77196674 Jie Deng 2020-09-11 152 kfree(vmsg->buf);
> > > 0a54ec77196674 Jie Deng 2020-09-11 153 vmsg->buf = NULL;
> > > 0a54ec77196674 Jie Deng 2020-09-11 154 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 155
> > > 0a54ec77196674 Jie Deng 2020-09-11 156 reinit_completion(&vi->completion);
> > > 0a54ec77196674 Jie Deng 2020-09-11 157 }
> > > 0a54ec77196674 Jie Deng 2020-09-11 158
> > > 0a54ec77196674 Jie Deng 2020-09-11 159 mutex_unlock(&vi->i2c_lock);
> > > 0a54ec77196674 Jie Deng 2020-09-11 @160 kfree(vmsg->buf);
> > > ^^^^^^^^^
> > > Unchecked dereference.
> > >
> > > 0a54ec77196674 Jie Deng 2020-09-11 161 return ((ret < 0) ? ret : i);
> > > 0a54ec77196674 Jie Deng 2020-09-11 162 }
> > >
> > > ---
> > > 0-DAY CI Kernel Test Service, Intel Corporation
> > > https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx
> >
> >
> >
> > --
> > With Best Regards,
> > Andy Shevchenko
> >
--
With Best Regards,
Andy Shevchenko