| Message ID | 20200715113842.30680-1-jonathanh@nvidia.com |
|---|---|
| State | Accepted |
| Headers | show |
| Series | [V2] usb: tegra: Fix allocation for the FPCI context | expand |
On Wed, Jul 15, 2020 at 12:38:42PM +0100, Jon Hunter wrote: > Commit 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB > context save/restore") is using the IPFS 'num_offsets' value when > allocating memory for FPCI context instead of the FPCI 'num_offsets'. > > After commit cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") > was added system suspend started failing on Tegra186. The kernel log > showed that the Tegra XHCI driver was crashing on entry to suspend when > attempting the save the USB context. On Tegra186, the IPFS context has a > zero length but the FPCI content has a non-zero length, and because of > the bug in the Tegra XHCI driver we are incorrectly allocating a zero > length array for the FPCI context. The crash seen on entering suspend > when we attempt to save the FPCI context and following commit > cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") this now > causes a NULL pointer deference when we access the memory. Fix this by > correcting the amount of memory we are allocating for FPCI contexts. > > Cc: stable@vger.kernel.org > > Fixes: 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB context save/restore") > > Signed-off-by: Jon Hunter <jonathanh@nvidia.com> > Acked-by: Thierry Reding <treding@nvidia.com> > --- > > Changes since V1: > - Corrected commit message > - Added Thierry's ACK > > drivers/usb/host/xhci-tegra.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) No cc: to linux-usb@vger? :( I'll go queue this up, but I would have caught it sooner if you had done so... thanks, greg k-h
On 23/07/2020 12:19, Greg Kroah-Hartman wrote: > On Wed, Jul 15, 2020 at 12:38:42PM +0100, Jon Hunter wrote: >> Commit 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB >> context save/restore") is using the IPFS 'num_offsets' value when >> allocating memory for FPCI context instead of the FPCI 'num_offsets'. >> >> After commit cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") >> was added system suspend started failing on Tegra186. The kernel log >> showed that the Tegra XHCI driver was crashing on entry to suspend when >> attempting the save the USB context. On Tegra186, the IPFS context has a >> zero length but the FPCI content has a non-zero length, and because of >> the bug in the Tegra XHCI driver we are incorrectly allocating a zero >> length array for the FPCI context. The crash seen on entering suspend >> when we attempt to save the FPCI context and following commit >> cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") this now >> causes a NULL pointer deference when we access the memory. Fix this by >> correcting the amount of memory we are allocating for FPCI contexts. >> >> Cc: stable@vger.kernel.org >> >> Fixes: 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB context save/restore") >> >> Signed-off-by: Jon Hunter <jonathanh@nvidia.com> >> Acked-by: Thierry Reding <treding@nvidia.com> >> --- >> >> Changes since V1: >> - Corrected commit message >> - Added Thierry's ACK >> >> drivers/usb/host/xhci-tegra.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) > > No cc: to linux-usb@vger? :( > > I'll go queue this up, but I would have caught it sooner if you had done > so... Sorry about that. Thanks for queuing up! Jon
diff --git a/drivers/usb/host/xhci-tegra.c b/drivers/usb/host/xhci-tegra.c index 9ce28ab47f4b..014d79334f50 100644 --- a/drivers/usb/host/xhci-tegra.c +++ b/drivers/usb/host/xhci-tegra.c @@ -856,7 +856,7 @@ static int tegra_xusb_init_context(struct tegra_xusb *tegra) if (!tegra->context.ipfs) return -ENOMEM; - tegra->context.fpci = devm_kcalloc(tegra->dev, soc->ipfs.num_offsets, + tegra->context.fpci = devm_kcalloc(tegra->dev, soc->fpci.num_offsets, sizeof(u32), GFP_KERNEL); if (!tegra->context.fpci) return -ENOMEM;
Commit 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB context save/restore") is using the IPFS 'num_offsets' value when allocating memory for FPCI context instead of the FPCI 'num_offsets'. After commit cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") was added system suspend started failing on Tegra186. The kernel log showed that the Tegra XHCI driver was crashing on entry to suspend when attempting the save the USB context. On Tegra186, the IPFS context has a zero length but the FPCI content has a non-zero length, and because of the bug in the Tegra XHCI driver we are incorrectly allocating a zero length array for the FPCI context. The crash seen on entering suspend when we attempt to save the FPCI context and following commit cad064f1bd52 ("devres: handle zero size in devm_kmalloc()") this now causes a NULL pointer deference when we access the memory. Fix this by correcting the amount of memory we are allocating for FPCI contexts. Cc: stable@vger.kernel.org Fixes: 5c4e8d3781bc ("usb: host: xhci-tegra: Add support for XUSB context save/restore") Signed-off-by: Jon Hunter <jonathanh@nvidia.com> Acked-by: Thierry Reding <treding@nvidia.com> --- Changes since V1: - Corrected commit message - Added Thierry's ACK drivers/usb/host/xhci-tegra.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)