On 15.06.2023 20:17, Luca Weiss wrote:
> From: Craig Tatlor <[email protected]>
>
> The qfprom actually has size 0x3000, so adjust the reg.
>
> Note that the non-ECC-corrected qfprom can be found at 0xfc4b8000
> (-0x4000). The current reg points to the ECC-corrected qfprom block
> which should have equivalent values at all offsets compared to the
> non-corrected version.
>
> [[email protected]: extract to standalone patch and adjust for review
> comments]
>
> Fixes: c59ffb519357 ("arm: dts: msm8974: Add thermal zones, tsens and qfprom nodes")
> Signed-off-by: Craig Tatlor <[email protected]>
> Signed-off-by: Luca Weiss <[email protected]>
> ---
Not sure of the actual size of the region, maybe Bjorn can help..
Downstream 3.10 suggests 0x60F0, 0x20F0 after adjusting for the ECC offset
Konrad
> Changes in v2:
> - Keep base offset but expand reg from 0x1000 to 0x3000 (Konrad)
> - Link to v1: https://lore.kernel.org/r/[email protected]
> ---
> arch/arm/boot/dts/qcom-msm8974.dtsi | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi b/arch/arm/boot/dts/qcom-msm8974.dtsi
> index 7ed0d925a4e9..3156fe25967f 100644
> --- a/arch/arm/boot/dts/qcom-msm8974.dtsi
> +++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
> @@ -1194,7 +1194,7 @@ restart@fc4ab000 {
>
> qfprom: qfprom@fc4bc000 {
> compatible = "qcom,msm8974-qfprom", "qcom,qfprom";
> - reg = <0xfc4bc000 0x1000>;
> + reg = <0xfc4bc000 0x3000>;
> #address-cells = <1>;
> #size-cells = <1>;
>
>
> ---
> base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> change-id: 20230130-msm8974-qfprom-619c0e8f26eb
>
> Best regards,
On Thu, Jun 15, 2023 at 08:20:41PM +0200, Konrad Dybcio wrote:
> On 15.06.2023 20:17, Luca Weiss wrote:
> > From: Craig Tatlor <[email protected]>
> >
> > The qfprom actually has size 0x3000, so adjust the reg.
> >
> > Note that the non-ECC-corrected qfprom can be found at 0xfc4b8000
> > (-0x4000). The current reg points to the ECC-corrected qfprom block
> > which should have equivalent values at all offsets compared to the
> > non-corrected version.
> >
> > [[email protected]: extract to standalone patch and adjust for review
> > comments]
> >
> > Fixes: c59ffb519357 ("arm: dts: msm8974: Add thermal zones, tsens and qfprom nodes")
> > Signed-off-by: Craig Tatlor <[email protected]>
> > Signed-off-by: Luca Weiss <[email protected]>
> > ---
> Not sure of the actual size of the region, maybe Bjorn can help..
>
> Downstream 3.10 suggests 0x60F0, 0x20F0 after adjusting for the ECC offset
>
There is indeed 0x3000 bytes until the next region, but afaict the
corrected ECC values only cover the first 0x800 bytes thereof.
Can you please let me know if this patch fixes a problem, or just
makes the numbers look better?
Regards,
Bjorn
> Konrad
> > Changes in v2:
> > - Keep base offset but expand reg from 0x1000 to 0x3000 (Konrad)
> > - Link to v1: https://lore.kernel.org/r/[email protected]
> > ---
> > arch/arm/boot/dts/qcom-msm8974.dtsi | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi b/arch/arm/boot/dts/qcom-msm8974.dtsi
> > index 7ed0d925a4e9..3156fe25967f 100644
> > --- a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > +++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
> > @@ -1194,7 +1194,7 @@ restart@fc4ab000 {
> >
> > qfprom: qfprom@fc4bc000 {
> > compatible = "qcom,msm8974-qfprom", "qcom,qfprom";
> > - reg = <0xfc4bc000 0x1000>;
> > + reg = <0xfc4bc000 0x3000>;
> > #address-cells = <1>;
> > #size-cells = <1>;
> >
> >
> > ---
> > base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> > change-id: 20230130-msm8974-qfprom-619c0e8f26eb
> >
> > Best regards,
Hi Bjorn,
On Montag, 31. Juli 2023 23:45:21 CEST Bjorn Andersson wrote:
> On Thu, Jun 15, 2023 at 08:20:41PM +0200, Konrad Dybcio wrote:
> > On 15.06.2023 20:17, Luca Weiss wrote:
> > > From: Craig Tatlor <[email protected]>
> > >
> > > The qfprom actually has size 0x3000, so adjust the reg.
> > >
> > > Note that the non-ECC-corrected qfprom can be found at 0xfc4b8000
> > > (-0x4000). The current reg points to the ECC-corrected qfprom block
> > > which should have equivalent values at all offsets compared to the
> > > non-corrected version.
> > >
> > > [[email protected]: extract to standalone patch and adjust for review
> > > comments]
> > >
> > > Fixes: c59ffb519357 ("arm: dts: msm8974: Add thermal zones, tsens and
> > > qfprom nodes") Signed-off-by: Craig Tatlor <[email protected]>
> > > Signed-off-by: Luca Weiss <[email protected]>
> > > ---
> >
> > Not sure of the actual size of the region, maybe Bjorn can help..
> >
> > Downstream 3.10 suggests 0x60F0, 0x20F0 after adjusting for the ECC offset
>
> There is indeed 0x3000 bytes until the next region, but afaict the
> corrected ECC values only cover the first 0x800 bytes thereof.
>
> Can you please let me know if this patch fixes a problem, or just
> makes the numbers look better?
Initially this patch came from a different direction, to make space to use the
PVS bits for cpufreq. Since Konrad said in earlier revisions that I should
always use the +0x4000 space for the ECC-corrected variant I've switched to
that.
If you think it's not useful to have the qfprom size reflect the actual size,
we can also drop this patch since I don't think it's actually necessary for
anything that I have lying around in some branches.
I think I've just sent the current patch to make sure the hardware description
(dts) is as accurate as possible, but of course since any info on Qualcomm is
very restricted it could also be a bit wrong.
Regards
Luca
>
> Regards,
> Bjorn
>
> > Konrad
> >
> > > Changes in v2:
> > > - Keep base offset but expand reg from 0x1000 to 0x3000 (Konrad)
> > > - Link to v1:
> > > https://lore.kernel.org/r/20230130-msm8974-qfprom-v1-1-975aa0e5e083@z3n
> > > tu.xyz ---
> > >
> > > arch/arm/boot/dts/qcom-msm8974.dtsi | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > b/arch/arm/boot/dts/qcom-msm8974.dtsi index 7ed0d925a4e9..3156fe25967f
> > > 100644
> > > --- a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > +++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > @@ -1194,7 +1194,7 @@ restart@fc4ab000 {
> > >
> > > qfprom: qfprom@fc4bc000 {
> > >
> > > compatible = "qcom,msm8974-qfprom",
"qcom,qfprom";
> > >
> > > - reg = <0xfc4bc000 0x1000>;
> > > + reg = <0xfc4bc000 0x3000>;
> > >
> > > #address-cells = <1>;
> > > #size-cells = <1>;
> > >
> > > ---
> > > base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> > > change-id: 20230130-msm8974-qfprom-619c0e8f26eb
> > >
> > > Best regards,
On Sonntag, 6. August 2023 12:47:51 CEST Luca Weiss wrote:
> Hi Bjorn,
>
> On Montag, 31. Juli 2023 23:45:21 CEST Bjorn Andersson wrote:
> > On Thu, Jun 15, 2023 at 08:20:41PM +0200, Konrad Dybcio wrote:
> > > On 15.06.2023 20:17, Luca Weiss wrote:
> > > > From: Craig Tatlor <[email protected]>
> > > >
> > > > The qfprom actually has size 0x3000, so adjust the reg.
> > > >
> > > > Note that the non-ECC-corrected qfprom can be found at 0xfc4b8000
> > > > (-0x4000). The current reg points to the ECC-corrected qfprom block
> > > > which should have equivalent values at all offsets compared to the
> > > > non-corrected version.
> > > >
> > > > [[email protected]: extract to standalone patch and adjust for review
> > > > comments]
> > > >
> > > > Fixes: c59ffb519357 ("arm: dts: msm8974: Add thermal zones, tsens and
> > > > qfprom nodes") Signed-off-by: Craig Tatlor <[email protected]>
> > > > Signed-off-by: Luca Weiss <[email protected]>
> > > > ---
> > >
> > > Not sure of the actual size of the region, maybe Bjorn can help..
> > >
> > > Downstream 3.10 suggests 0x60F0, 0x20F0 after adjusting for the ECC
> > > offset
> >
> > There is indeed 0x3000 bytes until the next region, but afaict the
> > corrected ECC values only cover the first 0x800 bytes thereof.
> >
> > Can you please let me know if this patch fixes a problem, or just
> > makes the numbers look better?
>
> Initially this patch came from a different direction, to make space to use
> the PVS bits for cpufreq. Since Konrad said in earlier revisions that I
> should always use the +0x4000 space for the ECC-corrected variant I've
> switched to that.
>
> If you think it's not useful to have the qfprom size reflect the actual
> size, we can also drop this patch since I don't think it's actually
> necessary for anything that I have lying around in some branches.
>
> I think I've just sent the current patch to make sure the hardware
> description (dts) is as accurate as possible, but of course since any info
> on Qualcomm is very restricted it could also be a bit wrong.
Hi Bjorn,
this patch is still lying in my inbox. Do you think it's correct or incorrect
- so should we drop it?
Regards
Luca
>
> Regards
> Luca
>
> > Regards,
> > Bjorn
> >
> > > Konrad
> > >
> > > > Changes in v2:
> > > > - Keep base offset but expand reg from 0x1000 to 0x3000 (Konrad)
> > > > - Link to v1:
> > > > https://lore.kernel.org/r/20230130-msm8974-qfprom-v1-1-975aa0e5e083@z3
> > > > n
> > > > tu.xyz ---
> > > >
> > > > arch/arm/boot/dts/qcom-msm8974.dtsi | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > b/arch/arm/boot/dts/qcom-msm8974.dtsi index 7ed0d925a4e9..3156fe25967f
> > > > 100644
> > > > --- a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > +++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > @@ -1194,7 +1194,7 @@ restart@fc4ab000 {
> > > >
> > > > qfprom: qfprom@fc4bc000 {
> > > >
> > > > compatible = "qcom,msm8974-qfprom",
>
> "qcom,qfprom";
>
> > > > - reg = <0xfc4bc000 0x1000>;
> > > > + reg = <0xfc4bc000 0x3000>;
> > > >
> > > > #address-cells = <1>;
> > > > #size-cells = <1>;
> > > >
> > > > ---
> > > > base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> > > > change-id: 20230130-msm8974-qfprom-619c0e8f26eb
> > > >
> > > > Best regards,
On Fri, 22 Sept 2023 at 19:57, Luca Weiss <[email protected]> wrote:
>
> On Sonntag, 6. August 2023 12:47:51 CEST Luca Weiss wrote:
> > Hi Bjorn,
> >
> > On Montag, 31. Juli 2023 23:45:21 CEST Bjorn Andersson wrote:
> > > On Thu, Jun 15, 2023 at 08:20:41PM +0200, Konrad Dybcio wrote:
> > > > On 15.06.2023 20:17, Luca Weiss wrote:
> > > > > From: Craig Tatlor <[email protected]>
> > > > >
> > > > > The qfprom actually has size 0x3000, so adjust the reg.
> > > > >
> > > > > Note that the non-ECC-corrected qfprom can be found at 0xfc4b8000
> > > > > (-0x4000). The current reg points to the ECC-corrected qfprom block
> > > > > which should have equivalent values at all offsets compared to the
> > > > > non-corrected version.
> > > > >
> > > > > [[email protected]: extract to standalone patch and adjust for review
> > > > > comments]
> > > > >
> > > > > Fixes: c59ffb519357 ("arm: dts: msm8974: Add thermal zones, tsens and
> > > > > qfprom nodes") Signed-off-by: Craig Tatlor <[email protected]>
> > > > > Signed-off-by: Luca Weiss <[email protected]>
> > > > > ---
> > > >
> > > > Not sure of the actual size of the region, maybe Bjorn can help..
> > > >
> > > > Downstream 3.10 suggests 0x60F0, 0x20F0 after adjusting for the ECC
> > > > offset
> > >
> > > There is indeed 0x3000 bytes until the next region, but afaict the
> > > corrected ECC values only cover the first 0x800 bytes thereof.
> > >
> > > Can you please let me know if this patch fixes a problem, or just
> > > makes the numbers look better?
> >
> > Initially this patch came from a different direction, to make space to use
> > the PVS bits for cpufreq. Since Konrad said in earlier revisions that I
> > should always use the +0x4000 space for the ECC-corrected variant I've
> > switched to that.
> >
> > If you think it's not useful to have the qfprom size reflect the actual
> > size, we can also drop this patch since I don't think it's actually
> > necessary for anything that I have lying around in some branches.
> >
> > I think I've just sent the current patch to make sure the hardware
> > description (dts) is as accurate as possible, but of course since any info
> > on Qualcomm is very restricted it could also be a bit wrong.
>
> Hi Bjorn,
>
> this patch is still lying in my inbox. Do you think it's correct or incorrect
> - so should we drop it?
There are JTAG and coresight fuses at 0xfc4be024. So, I think, the
regions should be extended to 0x20f0 or 0x2100. BTW: could you please
also fix msm8974 and apq8084 in a similar way?
>
> Regards
> Luca
>
> >
> > Regards
> > Luca
> >
> > > Regards,
> > > Bjorn
> > >
> > > > Konrad
> > > >
> > > > > Changes in v2:
> > > > > - Keep base offset but expand reg from 0x1000 to 0x3000 (Konrad)
> > > > > - Link to v1:
> > > > > https://lore.kernel.org/r/20230130-msm8974-qfprom-v1-1-975aa0e5e083@z3
> > > > > n
> > > > > tu.xyz ---
> > > > >
> > > > > arch/arm/boot/dts/qcom-msm8974.dtsi | 2 +-
> > > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > > >
> > > > > diff --git a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > > b/arch/arm/boot/dts/qcom-msm8974.dtsi index 7ed0d925a4e9..3156fe25967f
> > > > > 100644
> > > > > --- a/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > > +++ b/arch/arm/boot/dts/qcom-msm8974.dtsi
> > > > > @@ -1194,7 +1194,7 @@ restart@fc4ab000 {
> > > > >
> > > > > qfprom: qfprom@fc4bc000 {
> > > > >
> > > > > compatible = "qcom,msm8974-qfprom",
> >
> > "qcom,qfprom";
> >
> > > > > - reg = <0xfc4bc000 0x1000>;
> > > > > + reg = <0xfc4bc000 0x3000>;
> > > > >
> > > > > #address-cells = <1>;
> > > > > #size-cells = <1>;
> > > > >
> > > > > ---
> > > > > base-commit: 858fd168a95c5b9669aac8db6c14a9aeab446375
> > > > > change-id: 20230130-msm8974-qfprom-619c0e8f26eb
> > > > >
> > > > > Best regards,
>
>
>
>
--
With best wishes
Dmitry