On Tue, Oct 29, 2019 at 05:07:58AM -0600, shuah wrote: > On 10/25/19 9:40 PM, Suwan Kim wrote: > > On Thu, Oct 24, 2019 at 04:52:52PM -0600, shuah wrote: > > > On 10/24/19 1:45 PM, Dan Carpenter wrote: > > > > On Wed, Oct 23, 2019 at 04:11:20PM +0900, Suwan Kim wrote: > > > > > On Tue, Oct 22, 2019 at 12:28:39PM +0300, Dan Carpenter wrote: > > > > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master > > > > > > head: 7d194c2100ad2a6dded545887d02754948ca5241 > > > > > > commit: ea44d190764b4422af4d1c29eaeb9e69e353b406 usbip: Implement SG support to vhci-hcd and stub driver > > > > > > date: 7 weeks ago > > > > > > > > > > > > If you fix the issue, kindly add following tag > > > > > > Reported-by: kbuild test robot <lkp@xxxxxxxxx> > > > > > > Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> > > > > > > > > > > > > New smatch warnings: > > > > > > drivers/usb/usbip/stub_rx.c:505 stub_recv_cmd_submit() error: uninitialized symbol 'nents'. > > > > > > > > > > > > Old smatch warnings: > > > > > > drivers/usb/usbip/stub_rx.c:450 stub_recv_xbuff() error: uninitialized symbol 'ret'. > > > > > > > > > > > > # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ea44d190764b4422af4d1c29eaeb9e69e353b406 > > > > > > git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git > > > > > > git remote update linus > > > > > > git checkout ea44d190764b4422af4d1c29eaeb9e69e353b406 > > > > > > vim +/nents +505 drivers/usb/usbip/stub_rx.c > > > > > > > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 453 static void stub_recv_cmd_submit(struct stub_device *sdev, > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 454 struct usbip_header *pdu) > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 455 { > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 456 struct stub_priv *priv; > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 457 struct usbip_device *ud = &sdev->ud; > > > > > > 2d8f4595d1f275 drivers/staging/usbip/stub_rx.c Max Vozeler 2011-01-12 458 struct usb_device *udev = sdev->udev; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 459 struct scatterlist *sgl = NULL, *sg; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 460 void *buffer = NULL; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 461 unsigned long long buf_len; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 462 int nents; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 463 int num_urbs = 1; > > > > > > c6688ef9f29762 drivers/usb/usbip/stub_rx.c Shuah Khan 2017-12-07 464 int pipe = get_pipe(sdev, pdu); > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 465 int use_sg = pdu->u.cmd_submit.transfer_flags & URB_DMA_MAP_SG; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 466 int support_sg = 1; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 467 int np = 0; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 468 int ret, i; > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 469 > > > > > > 635f545a7e8be7 drivers/usb/usbip/stub_rx.c Shuah Khan 2017-12-07 470 if (pipe == -1) > > > > > > 635f545a7e8be7 drivers/usb/usbip/stub_rx.c Shuah Khan 2017-12-07 471 return; > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 472 > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 473 priv = stub_priv_alloc(sdev, pdu); > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 474 if (!priv) > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 475 return; > > > > > > 4d7b5c7f8ad49b drivers/staging/usbip/stub_rx.c Takahiro Hirofuchi 2008-07-09 476 > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 477 buf_len = (unsigned long long)pdu->u.cmd_submit.transfer_buffer_length; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 478 > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 479 /* allocate urb transfer buffer, if needed */ > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 480 if (buf_len) { > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 481 if (use_sg) { > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 482 sgl = sgl_alloc(buf_len, GFP_KERNEL, &nents); > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 483 if (!sgl) > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 484 goto err_malloc; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 485 } else { > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 486 buffer = kzalloc(buf_len, GFP_KERNEL); > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 487 if (!buffer) > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 488 goto err_malloc; > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 489 } > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 490 } > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 491 > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 492 /* Check if the server's HCD supports SG */ > > > > > > ea44d190764b44 drivers/usb/usbip/stub_rx.c Suwan Kim 2019-08-28 493 if (use_sg && !udev->bus->sg_tablesize) { > > > > > > > > > > > > Smatch thinks "use_sg" can be true when "buf_len" is zero. It's hard > > > > > > to tell if Smatch is right or wrong without more context... > > > > > > > > > > This is a bit strange. The meaning of "use_sg" is that client will > > > > > use scatter-gather and client's urb->num_sgs is not zero. And buffer > > > > > length should not be zero. > > > > > > > > > > usb_sg and buf_len are both client-dependent variables, so I think > > > > > if they have wrong value in the server side, the client must have > > > > > sent use_sg and buf_len with incorrect values. > > > > > > > > > > Did this error occur when compiling? > > > > > > > > Smatch is doing static analysis, yes. > > > > > > > > > If then, Did Smatch also consider vhci tx side? > > > > > > > > I'm not really sure... I can't reproduce the warning because on my > > > > system Smatch doesn't parse usbip_recv() correctly so it ends up > > > > silencing that warning. :/ > > > > > > > > > > Hi Suwan, > > > > > > This is a problem that needs fixing. nents > > > > > > /* allocate urb transfer buffer, if needed */ > > > if (buf_len) { > > > if (use_sg) { > > > sgl = sgl_alloc(buf_len, GFP_KERNEL, &nents); > > > > > > nents gets initialized here by sgl_alloc() > > > > > > if (!sgl) > > > goto err_malloc; > > > } else { > > > buffer = kzalloc(buf_len, GFP_KERNEL); > > > if (!buffer) > > > goto err_malloc; > > > } > > > } > > > > > > /* Check if the server's HCD supports SG */ > > > if (use_sg && !udev->bus->sg_tablesize) { > > > /* > > > * If the server's HCD doesn't support SG, break a single SG > > > * request into several URBs and map each SG list entry to > > > * corresponding URB buffer. The previously allocated SG > > > * list is stored in priv->sgl (If the server's HCD support > > > SG, > > > * SG list is stored only in urb->sg) and it is used as an > > > * indicator that the server split single SG request into > > > * several URBs. Later, priv->sgl is used by stub_complete() > > > and > > > * stub_send_ret_submit() to reassemble the divied URBs. > > > */ > > > support_sg = 0; > > > num_urbs = nents; > > > > > > I think nents will be valid here. Is there need for this additional > > > check here? You can fold this into the previous use_sg check, right > > > after the sg_alloc() success, I would think. > > > > Or Is it your intention to check as follows? > > I think this looks good. > > > > ... > > int nents = 0 > > > > ... > > > > /* allocate urb transfer buffer, if needed */ > > if (buf_len) { > > if (use_sg) { > > sgl = sgl_alloc(buf_len, GFP_KERNEL, &nents); > > if (!sgl) > > goto err_malloc; > > Why can't we move the > > /* Check if the server's HCD supports SG */ > > if (use_sg && !udev->bus->sg_tablesize) { > > > > under this check? I understood. Moving this check after sgl_alloc() does not seem to require any additional checks on nents. But I think we need to check for the case that Smatch reported that use_sg is true and buf_len is zero. If there is no error check and an error condition occurs, the URB will be passed to the next step without a buffer. I attached the code. If you are okay, I will send a patch. Regards, Suwan Kim --- diff --git a/drivers/usb/usbip/stub_rx.c b/drivers/usb/usbip/stub_rx.c index 66edfeea68fe..0b6c4736ffd6 100644 --- a/drivers/usb/usbip/stub_rx.c +++ b/drivers/usb/usbip/stub_rx.c @@ -476,12 +476,39 @@ static void stub_recv_cmd_submit(struct stub_device *sdev, buf_len = (unsigned long long)pdu->u.cmd_submit.transfer_buffer_length; + if (use_sg && !buf_len) { + dev_err(&udev->dev, "sg buffer with zero length\n"); + goto err_malloc; + } + /* allocate urb transfer buffer, if needed */ if (buf_len) { if (use_sg) { sgl = sgl_alloc(buf_len, GFP_KERNEL, &nents); if (!sgl) goto err_malloc; + + /* Check if the server's HCD supports SG */ + if (!udev->bus->sg_tablesize) { + /* + * If the server's HCD doesn't support SG, break + * a single SG request into several URBs and map + * each SG list entry to corresponding URB + * buffer. The previously allocated SG list is + * stored in priv->sgl (If the server's HCD + * support SG, SG list is stored only in + * urb->sg) and it is used as an indicator that + * the server split single SG request into + * several URBs. Later, priv->sgl is used by + * stub_complete() and stub_send_ret_submit() to + * reassemble the divied URBs. + */ + support_sg = 0; + num_urbs = nents; + priv->completed_urbs = 0; + pdu->u.cmd_submit.transfer_flags &= + ~URB_DMA_MAP_SG; + } } else { buffer = kzalloc(buf_len, GFP_KERNEL); if (!buffer) @@ -489,24 +516,6 @@ static void stub_recv_cmd_submit(struct stub_device *sdev, } } - /* Check if the server's HCD supports SG */ - if (use_sg && !udev->bus->sg_tablesize) { - /* - * If the server's HCD doesn't support SG, break a single SG - * request into several URBs and map each SG list entry to - * corresponding URB buffer. The previously allocated SG - * list is stored in priv->sgl (If the server's HCD support SG, - * SG list is stored only in urb->sg) and it is used as an - * indicator that the server split single SG request into - * several URBs. Later, priv->sgl is used by stub_complete() and - * stub_send_ret_submit() to reassemble the divied URBs. - */ - support_sg = 0; - num_urbs = nents; - priv->completed_urbs = 0; - pdu->u.cmd_submit.transfer_flags &= ~URB_DMA_MAP_SG; - } - /* allocate urb array */ priv->num_urbs = num_urbs; priv->urbs = kmalloc_array(num_urbs, sizeof(*priv->urbs), GFP_KERNEL);
