[PATCH v2] winebus.sys: Do not report HID report read errors unconditionally

[Kai Krakow]

Device reports may come in faster than our consumers could possibly read
them, this is especially true for multi-axis events: When you move a
stick across its range, it will always generate at least two events, one
for the x axis, and one for the y axis. This is not really an error
situation, so let's just discard the error message.

We cannot completely get rid of `last_report_read`, tho, as it is used
to track delivery of the report in different places and if true, try to
deliver the report to queued readers.

v2: Do not add complexity and just discard the message as suggested
by Alexandre Julliard and approved by Aric Stewart.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=43125
CC: Aric Stewart <aric(a)codeweavers.com>
CC: Alexandre Julliard <julliard(a)winehq.org>
Signed-off-by: Kai Krakow <kai(a)kaishome.de>
---
 dlls/winebus.sys/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/dlls/winebus.sys/main.c b/dlls/winebus.sys/main.c
index deb2c86d96..ce68e4af86 100644
--- a/dlls/winebus.sys/main.c
+++ b/dlls/winebus.sys/main.c
@@ -641,9 +641,6 @@ void process_hid_report(DEVICE_OBJECT *device, BYTE *report, DWORD length)
             ext->buffer_size = length;
     }
 
-    if (!ext->last_report_read)
-        ERR_(hid_report)("Device reports coming in too fast, last report not read yet!\n");
-
     memcpy(ext->last_report, report, length);
     ext->last_report_size = length;
     ext->last_report_read = FALSE;
-- 
2.16.4

[Aric Stewart]

Signed-off-by: Aric Stewart <aric(a)codeweavers.com>

On 8/15/18 1:29 PM, Kai Krakow wrote:
> Device reports may come in faster than our consumers could possibly read
> them, this is especially true for multi-axis events: When you move a
> stick across its range, it will always generate at least two events, one
> for the x axis, and one for the y axis. This is not really an error
> situation, so let's just discard the error message.
>
> We cannot completely get rid of `last_report_read`, tho, as it is used
> to track delivery of the report in different places and if true, try to
> deliver the report to queued readers.
>
> v2: Do not add complexity and just discard the message as suggested
> by Alexandre Julliard and approved by Aric Stewart.
>
> Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=43125
> CC: Aric Stewart <aric(a)codeweavers.com>
> CC: Alexandre Julliard <julliard(a)winehq.org>
> Signed-off-by: Kai Krakow <kai(a)kaishome.de>
> ---
>   dlls/winebus.sys/main.c | 3 ---
>   1 file changed, 3 deletions(-)
>
> diff --git a/dlls/winebus.sys/main.c b/dlls/winebus.sys/main.c
> index deb2c86d96..ce68e4af86 100644
> --- a/dlls/winebus.sys/main.c
> +++ b/dlls/winebus.sys/main.c
> @@ -641,9 +641,6 @@ void process_hid_report(DEVICE_OBJECT *device, BYTE *report, DWORD length)
>               ext->buffer_size = length;
>       }
>
> -    if (!ext->last_report_read)
> -        ERR_(hid_report)("Device reports coming in too fast, last report not read yet!\n");
> -
>       memcpy(ext->last_report, report, length);
>       ext->last_report_size = length;
>       ext->last_report_read = FALSE;