On Wed, 22 Jun 2016, Wilfried Klaebe wrote:
> Am Wed, Jun 22, 2016 at 07:56:49AM -0700 schrieb Greg KH:
> > On Wed, Jun 22, 2016 at 12:11:08PM +0200, Wilfried Klaebe wrote:
> > > Hi.
> > >
> > > In linux-4.7-rc3, I activated CONFIG_UBSAN and started reporting results
> > > by opening bugs in bugzilla.kernel.org. This was first about a UBSAN
> > > splat in drivers/usb/host/ehci-hub.c:877:47 about which I wrote in
> > > <https://bugzilla.kernel.org/show_bug.cgi?id=120361>. GregKH asked if
> > > this was fixed in linux-next, but in linux-next-20160616 there was a
> > > different splat (index 2 out of range instead of index -1 out of range)
> > > in drivers/usb/host/ehci-hub.c:889:34. I got asked to report here.
> > > That last splat persists until at least linux-next-20160622.
> > >
> > > Is there anything else you need?
> >
> > Is this still in 4.7-rc4? All of the reported ubsan fixes for USB
> > should now be in there.
>
> It is.
>
> > If not, can you provide the full "splat" here?
>
> dmesg attached, including UBSAN message and following further ehci messages.
>
> Regards,
> Wilfried
Please test this patch.
Index: usb-4.x/include/linux/usb/ehci_def.h
===================================================================
--- usb-4.x.orig/include/linux/usb/ehci_def.h
+++ usb-4.x/include/linux/usb/ehci_def.h
@@ -180,11 +180,11 @@ struct ehci_regs {
* PORTSCx
*/
/* HOSTPC: offset 0x84 */
- u32 hostpc[1]; /* HOSTPC extension */
+ u32 hostpc[0]; /* HOSTPC extension */
#define HOSTPC_PHCD (1<<22) /* Phy clock disable */
#define HOSTPC_PSPD (3<<25) /* Port speed detection */
- u32 reserved5[16];
+ u32 reserved5[17];
/* USBMODE_EX: offset 0xc8 */
u32 usbmode_ex; /* USB Device mode extension */
By coincidence, it turns out that I'm the person responsible for the
other UBSAN complaint in your log. Please see if the following patch
fixes it.
Alan Stern
Index: usb-4.x/drivers/scsi/scsi_devinfo.c
===================================================================
--- usb-4.x.orig/drivers/scsi/scsi_devinfo.c
+++ usb-4.x/drivers/scsi/scsi_devinfo.c
@@ -429,7 +429,7 @@ static struct scsi_dev_info_list *scsi_d
* here, and we don't know what device it is
* trying to work with, leave it as-is.
*/
- vmax = 8; /* max length of vendor */
+ vmax = sizeof(devinfo->vendor);
vskip = vendor;
while (vmax > 0 && *vskip == ' ') {
vmax--;
@@ -439,7 +439,7 @@ static struct scsi_dev_info_list *scsi_d
while (vmax > 0 && vskip[vmax - 1] == ' ')
--vmax;
- mmax = 16; /* max length of model */
+ mmax = sizeof(devinfo->model);
mskip = model;
while (mmax > 0 && *mskip == ' ') {
mmax--;
@@ -455,10 +455,12 @@ static struct scsi_dev_info_list *scsi_d
* Behave like the older version of get_device_flags.
*/
if (memcmp(devinfo->vendor, vskip, vmax) ||
- devinfo->vendor[vmax])
+ (vmax < sizeof(devinfo->vendor) &&
+ devinfo->vendor[vmax]))
continue;
if (memcmp(devinfo->model, mskip, mmax) ||
- devinfo->model[mmax])
+ (mmax < sizeof(devinfo->model) &&
+ devinfo->model[mmax]))
continue;
return devinfo;
} else {
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
