Re: [PATCH 3.4 214/214] HID: logitech: dont use stack based dj_report structures

From: Marc Dionne
Date: Thu Jun 05 2014 - 07:30:34 EST


On Thu, Jun 5, 2014 at 12:19 AM, Greg Kroah-Hartman
<gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> 3.4-stable review patch. If anyone has any objections, please let me know.
>
> ------------------
>
> From: Marc Dionne <marc.c.dionne@xxxxxxxxx>
>
> commit d8dc3494f77a5cc3b274bae36f7e74e85cf8a407 upstream.
>
> On a system with a logitech wireless keyboard/mouse and DMA-API debugging
> enabled, this warning appears at boot:
>
> kernel: WARNING: at lib/dma-debug.c:929 check_for_stack.part.12+0x70/0xa7()
> kernel: Hardware name: MS-7593
> kernel: uhci_hcd 0000:00:1d.1: DMA-API: device driver maps memory fromstack [addr=ffff8801b0079c29]
>
> Make logi_dj_recv_query_paired_devices and logi_dj_recv_switch_to_dj_mode
> use a structure allocated with kzalloc rather than a stack based one.
>
> Signed-off-by: Marc Dionne <marc.c.dionne@xxxxxxxxx>
> Signed-off-by: Jiri Kosina <jkosina@xxxxxxx>
> Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
> Cc: Yijing Wang <wangyijing@xxxxxxxxxx>
> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
>
> ---
> drivers/hid/hid-logitech-dj.c | 38 ++++++++++++++++++++++++--------------
> 1 file changed, 24 insertions(+), 14 deletions(-)
>
> --- a/drivers/hid/hid-logitech-dj.c
> +++ b/drivers/hid/hid-logitech-dj.c
> @@ -474,28 +474,38 @@ static int logi_dj_recv_send_report(stru
>
> static int logi_dj_recv_query_paired_devices(struct dj_receiver_dev *djrcv_dev)
> {
> - struct dj_report dj_report;
> + struct dj_report *dj_report;
> + int retval;
>
> - memset(&dj_report, 0, sizeof(dj_report));
> - dj_report.report_id = REPORT_ID_DJ_SHORT;
> - dj_report.device_index = 0xFF;
> - dj_report.report_type = REPORT_TYPE_CMD_GET_PAIRED_DEVICES;
> - return logi_dj_recv_send_report(djrcv_dev, &dj_report);
> + dj_report = kzalloc(sizeof(dj_report), GFP_KERNEL);

This patch was unfortunately incorrect as is - it needs to allocate
sizeof(struct dj_report) here and a little further down. This was
later fixed in mainline by 8a55ade7655, which should also be included
if you take this - I don't see it in the current set.

> + if (!dj_report)
> + return -ENOMEM;
> + dj_report->report_id = REPORT_ID_DJ_SHORT;
> + dj_report->device_index = 0xFF;
> + dj_report->report_type = REPORT_TYPE_CMD_GET_PAIRED_DEVICES;
> + retval = logi_dj_recv_send_report(djrcv_dev, dj_report);
> + kfree(dj_report);
> + return retval;
> }
>
>
> static int logi_dj_recv_switch_to_dj_mode(struct dj_receiver_dev *djrcv_dev,
> unsigned timeout)
> {
> - struct dj_report dj_report;
> + struct dj_report *dj_report;
> + int retval;
>
> - memset(&dj_report, 0, sizeof(dj_report));
> - dj_report.report_id = REPORT_ID_DJ_SHORT;
> - dj_report.device_index = 0xFF;
> - dj_report.report_type = REPORT_TYPE_CMD_SWITCH;
> - dj_report.report_params[CMD_SWITCH_PARAM_DEVBITFIELD] = 0x3F;
> - dj_report.report_params[CMD_SWITCH_PARAM_TIMEOUT_SECONDS] = (u8)timeout;
> - return logi_dj_recv_send_report(djrcv_dev, &dj_report);
> + dj_report = kzalloc(sizeof(dj_report), GFP_KERNEL);

Same here.

> + if (!dj_report)
> + return -ENOMEM;
> + dj_report->report_id = REPORT_ID_DJ_SHORT;
> + dj_report->device_index = 0xFF;
> + dj_report->report_type = REPORT_TYPE_CMD_SWITCH;
> + dj_report->report_params[CMD_SWITCH_PARAM_DEVBITFIELD] = 0x3F;
> + dj_report->report_params[CMD_SWITCH_PARAM_TIMEOUT_SECONDS] = (u8)timeout;
> + retval = logi_dj_recv_send_report(djrcv_dev, dj_report);
> + kfree(dj_report);
> + return retval;
> }

Marc
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/