x86_64 allmodconfig build with W=1 gives these warnings:
drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
u32 reg, dummy;
^~~~~
drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
u32 reg, dummy;
^~~~~
cc1: all warnings being treated as errors
Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
stubbed, which is the same as what we already do in pata_cs5536.c .
Signed-off-by: John Garry <[email protected]>
diff --git a/drivers/ata/pata_cs5535.c b/drivers/ata/pata_cs5535.c
index 6725931f3c35..c2c3238ff84b 100644
--- a/drivers/ata/pata_cs5535.c
+++ b/drivers/ata/pata_cs5535.c
@@ -90,7 +90,7 @@ static void cs5535_set_piomode(struct ata_port *ap, struct ata_device *adev)
static const u16 pio_cmd_timings[5] = {
0xF7F4, 0x53F3, 0x13F1, 0x5131, 0x1131
};
- u32 reg, dummy;
+ u32 reg, __maybe_unused dummy;
struct ata_device *pair = ata_dev_pair(adev);
int mode = adev->pio_mode - XFER_PIO_0;
@@ -129,7 +129,7 @@ static void cs5535_set_dmamode(struct ata_port *ap, struct ata_device *adev)
static const u32 mwdma_timings[3] = {
0x7F0FFFF3, 0x7F035352, 0x7F024241
};
- u32 reg, dummy;
+ u32 reg, __maybe_unused dummy;
int mode = adev->dma_mode;
rdmsr(ATAC_CH0D0_DMA + 2 * adev->devno, reg, dummy);
--
2.35.3
On 6/27/22 22:12, John Garry wrote:
> x86_64 allmodconfig build with W=1 gives these warnings:
>
> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
> u32 reg, dummy;
> ^~~~~
> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
> u32 reg, dummy;
> ^~~~~
> cc1: all warnings being treated as errors
>
> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
> stubbed, which is the same as what we already do in pata_cs5536.c .
>
> Signed-off-by: John Garry <[email protected]>
Looks good, but I wonder why I am not getting this warning. I always do
W=1 and C=1 builds. I tried allmodconfig now and I am not getting the
warning...
>
> diff --git a/drivers/ata/pata_cs5535.c b/drivers/ata/pata_cs5535.c
> index 6725931f3c35..c2c3238ff84b 100644
> --- a/drivers/ata/pata_cs5535.c
> +++ b/drivers/ata/pata_cs5535.c
> @@ -90,7 +90,7 @@ static void cs5535_set_piomode(struct ata_port *ap, struct ata_device *adev)
> static const u16 pio_cmd_timings[5] = {
> 0xF7F4, 0x53F3, 0x13F1, 0x5131, 0x1131
> };
> - u32 reg, dummy;
> + u32 reg, __maybe_unused dummy;
> struct ata_device *pair = ata_dev_pair(adev);
>
> int mode = adev->pio_mode - XFER_PIO_0;
> @@ -129,7 +129,7 @@ static void cs5535_set_dmamode(struct ata_port *ap, struct ata_device *adev)
> static const u32 mwdma_timings[3] = {
> 0x7F0FFFF3, 0x7F035352, 0x7F024241
> };
> - u32 reg, dummy;
> + u32 reg, __maybe_unused dummy;
> int mode = adev->dma_mode;
>
> rdmsr(ATAC_CH0D0_DMA + 2 * adev->devno, reg, dummy);
--
Damien Le Moal
Western Digital Research
On 28/06/2022 09:59, Damien Le Moal wrote:
> On 6/27/22 22:12, John Garry wrote:
>> x86_64 allmodconfig build with W=1 gives these warnings:
>>
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> cc1: all warnings being treated as errors
>>
>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>
>> Signed-off-by: John Garry <[email protected]>
>
> Looks good, but I wonder why I am not getting this warning. I always do
> W=1 and C=1 builds. I tried allmodconfig now and I am not getting the
> warning...
eh, I'm not sure why. It's the only driver in drivers/ata/ which gives
such a problem.
I'm using gcc 9:
gcc version 9.3.1 20200406 [revision
6db837a5288ee3ca5ec504fbd5a765817e556ac2] (SUSE Linux)
Then I cross-compiled from my arm64 machine (to x86) and still got it:
john@ubuntu:~/kernel-dev$ /usr/bin/x86_64-linux-gnu-gcc --version
x86_64-linux-gnu-gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
>
>>
>> diff --git a/drivers/ata/pata_cs5535.c b/drivers/ata/pata_cs5535.c
>> index 6725931f3c35..c2c3238ff84b 100644
>> --- a/drivers/ata/pata_cs5535.c
>> +++ b/drivers/ata/pata_cs5535.c
>> @@ -90,7 +90,7 @@ static void cs5535_set_piomode(struct ata_port *ap, struct ata_device *adev)
>> static const u16 pio_cmd_timings[5] = {
>> 0xF7F4, 0x53F3, 0x13F1, 0x5131, 0x1131
>> };
>> - u32 reg, dummy;
>> + u32 reg, __maybe_unused dummy;
>> struct ata_device *pair = ata_dev_pair(adev);
>>
>> int mode = adev->pio_mode - XFER_PIO_0;
>> @@ -129,7 +129,7 @@ static void cs5535_set_dmamode(struct ata_port *ap, struct ata_device *adev)
>> static const u32 mwdma_timings[3] = {
>> 0x7F0FFFF3, 0x7F035352, 0x7F024241
>> };
>> - u32 reg, dummy;
>> + u32 reg, __maybe_unused dummy;
>> int mode = adev->dma_mode;
>>
>> rdmsr(ATAC_CH0D0_DMA + 2 * adev->devno, reg, dummy);
>
>
Hello!
On 6/28/22 11:59 AM, Damien Le Moal wrote:
[...]
>> x86_64 allmodconfig build with W=1 gives these warnings:
>>
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> cc1: all warnings being treated as errors
>>
>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>
>> Signed-off-by: John Garry <[email protected]>
Reviewed-by: Sergey Shtylyov <[email protected]>
> Looks good, but I wonder why I am not getting this warning. I always do
> W=1 and C=1 builds. I tried allmodconfig now and I am not getting the
> warning...
I can confirm the (fatal) warnings with RedHat gcc 10.3.1... but somehow they only
occur on x86_64 with allmodconfig indeed (which is strange)...
[...]
MBR, Sergey
On 6/28/22 1:05 PM, Sergey Shtylyov wrote:
[...]
>>> x86_64 allmodconfig build with W=1 gives these warnings:
>>>
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> cc1: all warnings being treated as errors
>>>
>>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>>> stubbed, which is the same as what we already do in pata_cs5536.c .
Wait, what do you mean by "stubbed", the version in <asm/paravirt.h>?
>>>
>>> Signed-off-by: John Garry <[email protected]>
>
> Reviewed-by: Sergey Shtylyov <[email protected]>
[...]
MBR, Sergey
On 6/28/22 19:05, Sergey Shtylyov wrote:
> Hello!
>
> On 6/28/22 11:59 AM, Damien Le Moal wrote:
> [...]
>>> x86_64 allmodconfig build with W=1 gives these warnings:
>>>
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> cc1: all warnings being treated as errors
>>>
>>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>>
>>> Signed-off-by: John Garry <[email protected]>
>
>
> Reviewed-by: Sergey Shtylyov <[email protected]>
>
>> Looks good, but I wonder why I am not getting this warning. I always do
>> W=1 and C=1 builds. I tried allmodconfig now and I am not getting the
>> warning...
>
> I can confirm the (fatal) warnings with RedHat gcc 10.3.1... but somehow they only
> occur on x86_64 with allmodconfig indeed (which is strange)...
Compiling on x86_64 with allmodconfig with gcc 12 and sparse-0.6.4-2
(Fedora 36) does not give the warning... Weird.
>
> [...]
>
> MBR, Sergey
>
--
Damien Le Moal
Western Digital Research
On 6/27/22 22:12, John Garry wrote:
> x86_64 allmodconfig build with W=1 gives these warnings:
>
> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
> u32 reg, dummy;
> ^~~~~
> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
> u32 reg, dummy;
> ^~~~~
> cc1: all warnings being treated as errors
>
> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
> stubbed, which is the same as what we already do in pata_cs5536.c .
>
> Signed-off-by: John Garry <[email protected]>
Applied to for-5.20. Thanks !
>
> diff --git a/drivers/ata/pata_cs5535.c b/drivers/ata/pata_cs5535.c
> index 6725931f3c35..c2c3238ff84b 100644
> --- a/drivers/ata/pata_cs5535.c
> +++ b/drivers/ata/pata_cs5535.c
> @@ -90,7 +90,7 @@ static void cs5535_set_piomode(struct ata_port *ap, struct ata_device *adev)
> static const u16 pio_cmd_timings[5] = {
> 0xF7F4, 0x53F3, 0x13F1, 0x5131, 0x1131
> };
> - u32 reg, dummy;
> + u32 reg, __maybe_unused dummy;
> struct ata_device *pair = ata_dev_pair(adev);
>
> int mode = adev->pio_mode - XFER_PIO_0;
> @@ -129,7 +129,7 @@ static void cs5535_set_dmamode(struct ata_port *ap, struct ata_device *adev)
> static const u32 mwdma_timings[3] = {
> 0x7F0FFFF3, 0x7F035352, 0x7F024241
> };
> - u32 reg, dummy;
> + u32 reg, __maybe_unused dummy;
> int mode = adev->dma_mode;
>
> rdmsr(ATAC_CH0D0_DMA + 2 * adev->devno, reg, dummy);
--
Damien Le Moal
Western Digital Research
On 28/06/2022 19:39, Sergey Shtylyov wrote:
> On 6/28/22 1:05 PM, Sergey Shtylyov wrote:
> [...]
>>>> x86_64 allmodconfig build with W=1 gives these warnings:
>>>>
>>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>>>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>>> u32 reg, dummy;
>>>> ^~~~~
>>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>>>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>>> u32 reg, dummy;
>>>> ^~~~~
>>>> cc1: all warnings being treated as errors
>>>>
>>>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>>>> stubbed, which is the same as what we already do in pata_cs5536.c .
>
> Wait, what do you mean by "stubbed", the version in <asm/paravirt.h>?
Ah, this mentioning of stubbed is wrong on my part. I was distracted by
the stub in pata_cs5536.c - that's what my text editor showed me for the
defintion of rdmsr()
Hi Damien,
I see that you applied this patch. If it has not been pushed can we hand
edit the commit message or I can send a v2? Or maybe we can live with it...
Thanks,
John
>
>>>>
>>>> Signed-off-by: John Garry <[email protected]>
>>
>> Reviewed-by: Sergey Shtylyov <[email protected]>
>
> [...]
>
> MBR, Sergey
> .
On 6/29/22 16:41, John Garry wrote:
> On 28/06/2022 19:39, Sergey Shtylyov wrote:
>> On 6/28/22 1:05 PM, Sergey Shtylyov wrote:
>> [...]
>>>>> x86_64 allmodconfig build with W=1 gives these warnings:
>>>>>
>>>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>>>>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>>>> u32 reg, dummy;
>>>>> ^~~~~
>>>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>>>>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>>>> u32 reg, dummy;
>>>>> ^~~~~
>>>>> cc1: all warnings being treated as errors
>>>>>
>>>>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>>>>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>
>> Wait, what do you mean by "stubbed", the version in <asm/paravirt.h>?
>
> Ah, this mentioning of stubbed is wrong on my part. I was distracted by
> the stub in pata_cs5536.c - that's what my text editor showed me for the
> defintion of rdmsr()
>
> Hi Damien,
>
> I see that you applied this patch. If it has not been pushed can we hand
> edit the commit message or I can send a v2? Or maybe we can live with it...
Send me an updated commit message and I will amend.
>
> Thanks,
> John
>
>>
>>>>>
>>>>> Signed-off-by: John Garry <[email protected]>
>>>
>>> Reviewed-by: Sergey Shtylyov <[email protected]>
>>
>> [...]
>>
>> MBR, Sergey
>> .
>
--
Damien Le Moal
Western Digital Research
On 6/29/22 4:51 AM, Damien Le Moal wrote:
>> x86_64 allmodconfig build with W=1 gives these warnings:
>>
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>> u32 reg, dummy;
>> ^~~~~
>> cc1: all warnings being treated as errors
>>
>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>
>> Signed-off-by: John Garry <[email protected]>
>
> Applied to for-5.20. Thanks !
Why not to 5.19? The warnings are fatal as can be seen from the commit log...
MBR, Sergey
On 6/29/22 22:32, Sergey Shtylyov wrote:
> On 6/29/22 4:51 AM, Damien Le Moal wrote:
>
>>> x86_64 allmodconfig build with W=1 gives these warnings:
>>>
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_piomode’:
>>> drivers/ata/pata_cs5535.c:93:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> drivers/ata/pata_cs5535.c: In function ‘cs5535_set_dmamode’:
>>> drivers/ata/pata_cs5535.c:132:11: error: variable ‘dummy’ set but not used [-Werror=unused-but-set-variable]
>>> u32 reg, dummy;
>>> ^~~~~
>>> cc1: all warnings being treated as errors
>>>
>>> Mark variables 'dummy' as "maybe unused" to satisfy when rdmsr() is
>>> stubbed, which is the same as what we already do in pata_cs5536.c .
>>>
>>> Signed-off-by: John Garry <[email protected]>
>>
>> Applied to for-5.20. Thanks !
>
> Why not to 5.19? The warnings are fatal as can be seen from the commit log...
This is only a compile warning and nobody complained, not even the 0-day
bot/linux-next builds, and for years... Most definitely not an urgent fix.
But yeah, sure, can do 5.19 fixes too.
>
> MBR, Sergey
>
--
Damien Le Moal
Western Digital Research