Re: [PATCH v3 3/4] iommu/arm-smmu: Add support to handle Qcom's wait-for-safe logic

Re: [PATCH v3 3/4] iommu/arm-smmu: Add support to handle Qcom's wait-for-safe logic

[Marc Gonzalez]

On 12/06/2019 09:15, Vivek Gautam wrote:

> This change is inspired by the downstream change from Patrick Daly
> to address performance issues with display and camera by handling
> this wait-for-safe within separte io-pagetable ops to do TLB
> maintenance. So a big thanks to him for the change.
> 
> Without this change the UFS reads are pretty slow:
> $ time dd if=/dev/sda of=/dev/zero bs=1048576 count=10 conv=sync
> 10+0 records in
> 10+0 records out
> 10485760 bytes (10.0MB) copied, 22.394903 seconds, 457.2KB/s
> real    0m 22.39s
> user    0m 0.00s
> sys     0m 0.01s
> 
> With this change they are back to rock!
> $ time dd if=/dev/sda of=/dev/zero bs=1048576 count=300 conv=sync
> 300+0 records in
> 300+0 records out
> 314572800 bytes (300.0MB) copied, 1.030541 seconds, 291.1MB/s
> real    0m 1.03s
> user    0m 0.00s
> sys     0m 0.54s

This issue does not affect msm8998, I presume?

Nevertheless, I see much lower performance on msm8998:

# dd if=/dev/sde of=/dev/null bs=1M status=progress
3892314112 bytes (3.9 GB, 3.6 GiB) copied, 50.0042 s, 77.8 MB/s

80 MB/s on msm8998 -- vs -- 300 MB/s on sdm845

Do you have the interconnect patches on sdm845 that allow boosting
the clock/bandwidth for specific HW blocks?

Regards.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v3 3/4] iommu/arm-smmu: Add support to handle Qcom's wait-for-safe logic

[Vivek Gautam]

On 6/14/2019 6:45 PM, Marc Gonzalez wrote:
> On 12/06/2019 09:15, Vivek Gautam wrote:
>
>> This change is inspired by the downstream change from Patrick Daly
>> to address performance issues with display and camera by handling
>> this wait-for-safe within separte io-pagetable ops to do TLB
>> maintenance. So a big thanks to him for the change.
>>
>> Without this change the UFS reads are pretty slow:
>> $ time dd if=/dev/sda of=/dev/zero bs=1048576 count=10 conv=sync
>> 10+0 records in
>> 10+0 records out
>> 10485760 bytes (10.0MB) copied, 22.394903 seconds, 457.2KB/s
>> real    0m 22.39s
>> user    0m 0.00s
>> sys     0m 0.01s
>>
>> With this change they are back to rock!
>> $ time dd if=/dev/sda of=/dev/zero bs=1048576 count=300 conv=sync
>> 300+0 records in
>> 300+0 records out
>> 314572800 bytes (300.0MB) copied, 1.030541 seconds, 291.1MB/s
>> real    0m 1.03s
>> user    0m 0.00s
>> sys     0m 0.54s
> This issue does not affect msm8998, I presume?
>
> Nevertheless, I see much lower performance on msm8998:
>
> # dd if=/dev/sde of=/dev/null bs=1M status=progress
> 3892314112 bytes (3.9 GB, 3.6 GiB) copied, 50.0042 s, 77.8 MB/s
>
> 80 MB/s on msm8998 -- vs -- 300 MB/s on sdm845
>
> Do you have the interconnect patches on sdm845 that allow boosting
> the clock/bandwidth for specific HW blocks?

Umm, No. This is the upstream 5.2-rc4 plus 4-6 patches to enable display 
and fix splash screen.
Is this the performance for UFS? The numbers i posted were for UFS.

Thanks
Vivek

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v3 3/4] iommu/arm-smmu: Add support to handle Qcom's wait-for-safe logic

[Marc Gonzalez]

On 17/06/2019 11:50, Vivek Gautam wrote:

> On 6/14/2019 6:45 PM, Marc Gonzalez wrote:
>
>> # dd if=/dev/sde of=/dev/null bs=1M status=progress
>> 3892314112 bytes (3.9 GB, 3.6 GiB) copied, 50.0042 s, 77.8 MB/s
>>
>> 80 MB/s on msm8998 -- vs -- 300 MB/s on sdm845
>>
>> Do you have the interconnect patches on sdm845 that allow boosting
>> the clock/bandwidth for specific HW blocks?
> 
> Umm, No. This is the upstream 5.2-rc4 plus 4-6 patches to enable display 
> and fix splash screen.
> Is this the performance for UFS? The numbers I posted were for UFS.

Correct, the numbers I provided were for msm8998 UFS...

Basically, it looks like sdm845 UFS is 4x faster than msm8998 UFS
using upstream. Which is surprising (may depend on specific Flash
chip in use though).

Would be good if somebody with both boards could post numbers.
I'll try to post "fresh" numbers when I can.

Regards.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel