Some extreme test environment may keep IC temperature very low or very high
during system boot stage. For stability concern, we add thermal voltage
compenstation if needed no matter svs bank phase is in init02 or mon mode.
Signed-off-by: Roger Lu <[email protected]>
Reviewed-by: AngeloGioacchino Del Regno <[email protected]>
---
drivers/soc/mediatek/mtk-svs.c | 17 +++++++++--------
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
index 299f580847bd..e104866d1ab5 100644
--- a/drivers/soc/mediatek/mtk-svs.c
+++ b/drivers/soc/mediatek/mtk-svs.c
@@ -558,7 +558,7 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
}
/* Get thermal effect */
- if (svsb->phase == SVSB_PHASE_MON) {
+ if (!IS_ERR_OR_NULL(svsb->tzd)) {
ret = thermal_zone_get_temp(svsb->tzd, &tzone_temp);
if (ret || (svsb->temp > SVSB_TEMP_UPPER_BOUND &&
svsb->temp < SVSB_TEMP_LOWER_BOUND)) {
@@ -573,7 +573,8 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
temp_voffset += svsb->tzone_ltemp_voffset;
/* 2-line bank update all opp volts when running mon mode */
- if (svsb->type == SVSB_HIGH || svsb->type == SVSB_LOW) {
+ if (svsb->phase == SVSB_PHASE_MON && (svsb->type == SVSB_HIGH ||
+ svsb->type == SVSB_LOW)) {
opp_start = 0;
opp_stop = svsb->opp_count;
}
@@ -589,11 +590,6 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
/* do nothing */
goto unlock_mutex;
case SVSB_PHASE_INIT02:
- svsb_volt = max(svsb->volt[i], svsb->vmin);
- opp_volt = svs_bank_volt_to_opp_volt(svsb_volt,
- svsb->volt_step,
- svsb->volt_base);
- break;
case SVSB_PHASE_MON:
svsb_volt = max(svsb->volt[i] + temp_voffset, svsb->vmin);
opp_volt = svs_bank_volt_to_opp_volt(svsb_volt,
@@ -1683,7 +1679,7 @@ static int svs_bank_resource_setup(struct svs_platform *svsp)
}
}
- if (svsb->mode_support & SVSB_MODE_MON) {
+ if (!IS_ERR_OR_NULL(svsb->tzone_name)) {
svsb->tzd = thermal_zone_get_zone_by_name(svsb->tzone_name);
if (IS_ERR(svsb->tzd)) {
dev_err(svsb->dev, "cannot get \"%s\" thermal zone\n",
@@ -2127,6 +2123,7 @@ static struct svs_bank svs_mt8192_banks[] = {
.type = SVSB_LOW,
.set_freq_pct = svs_set_bank_freq_pct_v3,
.get_volts = svs_get_bank_volts_v3,
+ .tzone_name = "gpu1",
.volt_flags = SVSB_REMOVE_DVTFIXED_VOLT,
.mode_support = SVSB_MODE_INIT02,
.opp_count = MAX_OPP_ENTRIES,
@@ -2144,6 +2141,10 @@ static struct svs_bank svs_mt8192_banks[] = {
.core_sel = 0x0fff0100,
.int_st = BIT(0),
.ctl0 = 0x00540003,
+ .tzone_htemp = 85000,
+ .tzone_htemp_voffset = 0,
+ .tzone_ltemp = 25000,
+ .tzone_ltemp_voffset = 7,
},
{
.sw_id = SVSB_GPU,
--
2.18.0
Hi Roger,
I have some doubts, please see below.
On 02/02/2023 13:41, Roger Lu wrote:
> Some extreme test environment may keep IC temperature very low or very high
> during system boot stage. For stability concern, we add thermal voltage
> compenstation if needed no matter svs bank phase is in init02 or mon mode.
>
> Signed-off-by: Roger Lu <[email protected]>
> Reviewed-by: AngeloGioacchino Del Regno <[email protected]>
> ---
> drivers/soc/mediatek/mtk-svs.c | 17 +++++++++--------
> 1 file changed, 9 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
> index 299f580847bd..e104866d1ab5 100644
> --- a/drivers/soc/mediatek/mtk-svs.c
> +++ b/drivers/soc/mediatek/mtk-svs.c
> @@ -558,7 +558,7 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
> }
>
> /* Get thermal effect */
> - if (svsb->phase == SVSB_PHASE_MON) {
> + if (!IS_ERR_OR_NULL(svsb->tzd)) {
> ret = thermal_zone_get_temp(svsb->tzd, &tzone_temp);
> if (ret || (svsb->temp > SVSB_TEMP_UPPER_BOUND &&
> svsb->temp < SVSB_TEMP_LOWER_BOUND)) {
> @@ -573,7 +573,8 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
> temp_voffset += svsb->tzone_ltemp_voffset;
>
> /* 2-line bank update all opp volts when running mon mode */
> - if (svsb->type == SVSB_HIGH || svsb->type == SVSB_LOW) {
> + if (svsb->phase == SVSB_PHASE_MON && (svsb->type == SVSB_HIGH ||
> + svsb->type == SVSB_LOW)) {
> opp_start = 0;
> opp_stop = svsb->opp_count;
> }
> @@ -589,11 +590,6 @@ static int svs_adjust_pm_opp_volts(struct svs_bank *svsb)
> /* do nothing */
> goto unlock_mutex;
> case SVSB_PHASE_INIT02:
> - svsb_volt = max(svsb->volt[i], svsb->vmin);
> - opp_volt = svs_bank_volt_to_opp_volt(svsb_volt,
> - svsb->volt_step,
> - svsb->volt_base);
> - break;
> case SVSB_PHASE_MON:
> svsb_volt = max(svsb->volt[i] + temp_voffset, svsb->vmin);
> opp_volt = svs_bank_volt_to_opp_volt(svsb_volt,
> @@ -1683,7 +1679,7 @@ static int svs_bank_resource_setup(struct svs_platform *svsp)
> }
> }
>
> - if (svsb->mode_support & SVSB_MODE_MON) {
> + if (!IS_ERR_OR_NULL(svsb->tzone_name)) {
> svsb->tzd = thermal_zone_get_zone_by_name(svsb->tzone_name);
> if (IS_ERR(svsb->tzd)) {
> dev_err(svsb->dev, "cannot get \"%s\" thermal zone\n",
> @@ -2127,6 +2123,7 @@ static struct svs_bank svs_mt8192_banks[] = {
> .type = SVSB_LOW,
> .set_freq_pct = svs_set_bank_freq_pct_v3,
> .get_volts = svs_get_bank_volts_v3,
> + .tzone_name = "gpu1",
> .volt_flags = SVSB_REMOVE_DVTFIXED_VOLT,
> .mode_support = SVSB_MODE_INIT02,
> .opp_count = MAX_OPP_ENTRIES,
> @@ -2144,6 +2141,10 @@ static struct svs_bank svs_mt8192_banks[] = {
> .core_sel = 0x0fff0100,
> .int_st = BIT(0),
> .ctl0 = 0x00540003,
> + .tzone_htemp = 85000,
> + .tzone_htemp_voffset = 0,
> + .tzone_ltemp = 25000,
> + .tzone_ltemp_voffset = 7,
Which is the exact same tzone then in the other bank. Which brings me to a good
point:
Is the tzone bank specific or the same for all banks?
At least for mt8192 they are not. I suppose with this change to the code mt8183
could take advantage of this on all it's banks as well. In that case, can we
start to restructure the struct svs_bank to only have the tzone values declared
once?
Background is that I'm very unhappy with the svs_bank data strucutre. It seems
like a "throw it all in here". It should be structured for functional parts of
the banks. Maybe using structs, maybe unions where possible. In any case having
a flat struct of over 50 members isn't really what we want.
Regards,
Matthias
> },
> {
> .sw_id = SVSB_GPU,
Hi Matthias Sir,
Sorry for the late reply.
... [snip] ...
> > @@ -2127,6 +2123,7 @@ static struct svs_bank svs_mt8192_banks[] = {
> > .type = SVSB_LOW,
> > .set_freq_pct = svs_set_bank_freq_pct_v3,
> > .get_volts = svs_get_bank_volts_v3,
> > + .tzone_name = "gpu1",
> > .volt_flags = SVSB_REMOVE_DVTFIXED_VOLT,
> > .mode_support = SVSB_MODE_INIT02,
> > .opp_count = MAX_OPP_ENTRIES,
> > @@ -2144,6 +2141,10 @@ static struct svs_bank svs_mt8192_banks[] = {
> > .core_sel = 0x0fff0100,
> > .int_st = BIT(0),
> > .ctl0 = 0x00540003,
> > + .tzone_htemp = 85000,
> > + .tzone_htemp_voffset = 0,
> > + .tzone_ltemp = 25000,
> > + .tzone_ltemp_voffset = 7,
>
> Which is the exact same tzone then in the other bank. Which brings me to a
> good
> point:
> Is the tzone bank specific or the same for all banks?
Thermal zone (tzone) isn't for all SVS banks. In other words, tzone is specific
for corresponding DVFS domain like SVS GPU tzone is for GPU DVFS domain. Let's
take MT8183 SVS and MT8192 SVS as examples.
MT8192 SVS applies 2-line HW design (High/low 2 banks optimize the same DVFS
domain). So, SVS GPU High/low bank uses the same GPU tzone.
MT8183 SVS applies 1-line HW design (1 bank optimizes 1 DVFS domain)
Therefore, SVS CPU/GPU/CCI bank use different tzone because they are different
DVFS domain.
> At least for mt8192 they are not. I suppose with this change to the code
> mt8183
> could take advantage of this on all it's banks as well.
> In that case, can we
> start to restructure the struct svs_bank to only have the tzone values
> declared
> once?
Since tzone isn't for all banks, we cannot declare it once for all IC supports
from this point of view.
>
> Background is that I'm very unhappy with the svs_bank data strucutre. It
> seems
> like a "throw it all in here". It should be structured for functional parts
> of
> the banks. Maybe using structs, maybe unions where possible. In any case
> having
> a flat struct of over 50 members isn't really what we want.
My apology. We'll structure svs_bank for functional parts of them.
>
> Regards,
> Matthias
On 11/02/2023 12:12, Roger Lu (陸瑞傑) wrote:
> Hi Matthias Sir,
>
> Sorry for the late reply.
>
> ... [snip] ...
>
>>> @@ -2127,6 +2123,7 @@ static struct svs_bank svs_mt8192_banks[] = {
>>> .type = SVSB_LOW,
>>> .set_freq_pct = svs_set_bank_freq_pct_v3,
>>> .get_volts = svs_get_bank_volts_v3,
>>> + .tzone_name = "gpu1",
>>> .volt_flags = SVSB_REMOVE_DVTFIXED_VOLT,
>>> .mode_support = SVSB_MODE_INIT02,
>>> .opp_count = MAX_OPP_ENTRIES,
>>> @@ -2144,6 +2141,10 @@ static struct svs_bank svs_mt8192_banks[] = {
>>> .core_sel = 0x0fff0100,
>>> .int_st = BIT(0),
>>> .ctl0 = 0x00540003,
>>> + .tzone_htemp = 85000,
>>> + .tzone_htemp_voffset = 0,
>>> + .tzone_ltemp = 25000,
>>> + .tzone_ltemp_voffset = 7,
>>
>> Which is the exact same tzone then in the other bank. Which brings me to a
>> good
>> point:
>> Is the tzone bank specific or the same for all banks?
>
> Thermal zone (tzone) isn't for all SVS banks. In other words, tzone is specific
> for corresponding DVFS domain like SVS GPU tzone is for GPU DVFS domain. Let's
> take MT8183 SVS and MT8192 SVS as examples.
>
> MT8192 SVS applies 2-line HW design (High/low 2 banks optimize the same DVFS
> domain). So, SVS GPU High/low bank uses the same GPU tzone.
>
> MT8183 SVS applies 1-line HW design (1 bank optimizes 1 DVFS domain)
> Therefore, SVS CPU/GPU/CCI bank use different tzone because they are different
> DVFS domain.
>
>> At least for mt8192 they are not. I suppose with this change to the code
>> mt8183
>> could take advantage of this on all it's banks as well.
>> In that case, can we
>> start to restructure the struct svs_bank to only have the tzone values
>> declared
>> once?
>
> Since tzone isn't for all banks, we cannot declare it once for all IC supports
> from this point of view.
>
Thanks for clarification, applied now.
>>
>> Background is that I'm very unhappy with the svs_bank data strucutre. It
>> seems
>> like a "throw it all in here". It should be structured for functional parts
>> of
>> the banks. Maybe using structs, maybe unions where possible. In any case
>> having
>> a flat struct of over 50 members isn't really what we want.
>
> My apology. We'll structure svs_bank for functional parts of them.
>
>>
>> Regards,
>> Matthias