In commit 221abd4d478a ("staging: r8188eu: Remove no more necessary definitions
and code"), two entries were removed from RTW_ChannelPlanMap[], but not replaced
with zeros. The position within this table is important, thus the patch broke
systems operating in regulatory domains osted later than entry 0x13 in the table.
Unfortunately, the FCC entry comes before that point and most testers did not see
this problem.
Reported-and-tested-by: Zameer Manji <[email protected]>
Reported-by: kernel test robot <[email protected]>
Fixes: 221abd4d478a ("staging: r8188eu: Remove no more necessary definitions and code")
Cc: Stable <[email protected]> # v5.5+
Signed-off-by: Larry Finger <[email protected]>
---
v2 - fixed use of () rsther than {} - found by kernel test robot
---
drivers/staging/r8188eu/core/rtw_mlme_ext.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/staging/r8188eu/core/rtw_mlme_ext.c b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
index 55c3d4a6faeb..5b60e6df5f87 100644
--- a/drivers/staging/r8188eu/core/rtw_mlme_ext.c
+++ b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
@@ -107,6 +107,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
{0x01}, /* 0x10, RT_CHANNEL_DOMAIN_JAPAN */
{0x02}, /* 0x11, RT_CHANNEL_DOMAIN_FCC_NO_DFS */
{0x01}, /* 0x12, RT_CHANNEL_DOMAIN_JAPAN_NO_DFS */
+ {0x00}, /* 0x13 */
{0x02}, /* 0x14, RT_CHANNEL_DOMAIN_TAIWAN_NO_DFS */
{0x00}, /* 0x15, RT_CHANNEL_DOMAIN_ETSI_NO_DFS */
{0x00}, /* 0x16, RT_CHANNEL_DOMAIN_KOREA_NO_DFS */
@@ -118,6 +119,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
{0x00}, /* 0x1C, */
{0x00}, /* 0x1D, */
{0x00}, /* 0x1E, */
+ {0x00}, /* 0x1F, */
/* 0x20 ~ 0x7F , New Define ===== */
{0x00}, /* 0x20, RT_CHANNEL_DOMAIN_WORLD_NULL */
{0x01}, /* 0x21, RT_CHANNEL_DOMAIN_ETSI1_NULL */
--
2.33.1
From: Larry Finger
> Sent: 07 November 2021 17:36
>
> In commit 221abd4d478a ("staging: r8188eu: Remove no more necessary definitions
> and code"), two entries were removed from RTW_ChannelPlanMap[], but not replaced
> with zeros. The position within this table is important, thus the patch broke
> systems operating in regulatory domains osted later than entry 0x13 in the table.
> Unfortunately, the FCC entry comes before that point and most testers did not see
> this problem.
...
> drivers/staging/r8188eu/core/rtw_mlme_ext.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/staging/r8188eu/core/rtw_mlme_ext.c b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
> index 55c3d4a6faeb..5b60e6df5f87 100644
> --- a/drivers/staging/r8188eu/core/rtw_mlme_ext.c
> +++ b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
> @@ -107,6 +107,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
> {0x01}, /* 0x10, RT_CHANNEL_DOMAIN_JAPAN */
> {0x02}, /* 0x11, RT_CHANNEL_DOMAIN_FCC_NO_DFS */
> {0x01}, /* 0x12, RT_CHANNEL_DOMAIN_JAPAN_NO_DFS */
> + {0x00}, /* 0x13 */
> {0x02}, /* 0x14, RT_CHANNEL_DOMAIN_TAIWAN_NO_DFS */
> {0x00}, /* 0x15, RT_CHANNEL_DOMAIN_ETSI_NO_DFS */
> {0x00}, /* 0x16, RT_CHANNEL_DOMAIN_KOREA_NO_DFS */
> @@ -118,6 +119,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
> {0x00}, /* 0x1C, */
> {0x00}, /* 0x1D, */
> {0x00}, /* 0x1E, */
> + {0x00}, /* 0x1F, */
> /* 0x20 ~ 0x7F , New Define ===== */
> {0x00}, /* 0x20, RT_CHANNEL_DOMAIN_WORLD_NULL */
> {0x01}, /* 0x21, RT_CHANNEL_DOMAIN_ETSI1_NULL */
Is it worth changing that to use designated array initialisers?
so:
[0x21] = {0x01} /* RT_CHANNEL_DOMAIN_ETS11_NULL */
Then the {0x00} entries can be missed out (or just commented as not used).
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
On 11/8/21 04:55, David Laight wrote:
> From: Larry Finger
>> Sent: 07 November 2021 17:36
>>
>> In commit 221abd4d478a ("staging: r8188eu: Remove no more necessary definitions
>> and code"), two entries were removed from RTW_ChannelPlanMap[], but not replaced
>> with zeros. The position within this table is important, thus the patch broke
>> systems operating in regulatory domains osted later than entry 0x13 in the table.
>> Unfortunately, the FCC entry comes before that point and most testers did not see
>> this problem.
> ...
>> drivers/staging/r8188eu/core/rtw_mlme_ext.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/staging/r8188eu/core/rtw_mlme_ext.c b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
>> index 55c3d4a6faeb..5b60e6df5f87 100644
>> --- a/drivers/staging/r8188eu/core/rtw_mlme_ext.c
>> +++ b/drivers/staging/r8188eu/core/rtw_mlme_ext.c
>> @@ -107,6 +107,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
>> {0x01}, /* 0x10, RT_CHANNEL_DOMAIN_JAPAN */
>> {0x02}, /* 0x11, RT_CHANNEL_DOMAIN_FCC_NO_DFS */
>> {0x01}, /* 0x12, RT_CHANNEL_DOMAIN_JAPAN_NO_DFS */
>> + {0x00}, /* 0x13 */
>> {0x02}, /* 0x14, RT_CHANNEL_DOMAIN_TAIWAN_NO_DFS */
>> {0x00}, /* 0x15, RT_CHANNEL_DOMAIN_ETSI_NO_DFS */
>> {0x00}, /* 0x16, RT_CHANNEL_DOMAIN_KOREA_NO_DFS */
>> @@ -118,6 +119,7 @@ static struct rt_channel_plan_map RTW_ChannelPlanMap[RT_CHANNEL_DOMAIN_MAX] = {
>> {0x00}, /* 0x1C, */
>> {0x00}, /* 0x1D, */
>> {0x00}, /* 0x1E, */
>> + {0x00}, /* 0x1F, */
>> /* 0x20 ~ 0x7F , New Define ===== */
>> {0x00}, /* 0x20, RT_CHANNEL_DOMAIN_WORLD_NULL */
>> {0x01}, /* 0x21, RT_CHANNEL_DOMAIN_ETSI1_NULL */
>
> Is it worth changing that to use designated array initialisers?
> so:
> [0x21] = {0x01} /* RT_CHANNEL_DOMAIN_ETS11_NULL */
>
> Then the {0x00} entries can be missed out (or just commented as not used).
Yes, we could save a few bytes but risk getting into trouble aligning with some
other array that I have not located. I think it should be left the way it is.
Larry