[on hold] Istep 8 6 #32
Conversation
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
|
Not build-tested. It seems it doesn't touch the hardware but calculates coefficients for later use. |
|
@miczyg1 which values are written to |
src/soc/ibm/power9/isteps/istep8.c
Outdated
| */ | ||
| if ((eps_r[0] > eps_r[1]) || (eps_r[1] > eps_r[2]) || (eps_w[0] > eps_w[1])) | ||
| printk(BIOS_WARNING, "Invalid relationship between base epsilon values\n"); | ||
| } |
There was a problem hiding this comment.
@krystian-hebel the attributes ATTR_PROC_EPS_READ_CYCLES_T{0,1,2,} are written right here. Their values are as follows:
ATTR_PROC_EPS_READ_CYCLES_T0 = eps_r[0];
ATTR_PROC_EPS_READ_CYCLES_T1 = eps_r[1];
ATTR_PROC_EPS_READ_CYCLES_T2 = eps_r[2];
ATTR_PROC_EPS_WRITE_CYCLES_T1 = eps_r[0];
ATTR_PROC_EPS_WRITE_CYCLES_T2 = eps_r[1];
|
@miczyg1 @krystian-hebel What is the status here? |
|
@miczyg1 @krystian-hebel Could you update what is the status if this MR? I am also wondering on the code structure. In this MR we have istep8.h/istep8.c, while in the others MRs we have a separate file for each minor istep (like here: #33) Moreover, we already have some istep 8.x implementation here: https://github.com/3mdeb/coreboot/pull/39/files The directory structure is also different, e.g. |
…) function name Co-authored-by: krystian-hebel <40995177+krystian-hebel@users.noreply.github.com>
…eps_table_type variable Co-authored-by: krystian-hebel <40995177+krystian-hebel@users.noreply.github.com>
Co-authored-by: krystian-hebel <40995177+krystian-hebel@users.noreply.github.com>
variables Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
Removing ./isteps directory and istep8.c file Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
src/soc/ibm/power9 Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
previously static variables Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
global Signed-off-by: Mateusz Kowalski <mateusz.kowalski@3mdeb.com>
|
@miczyg1 @krystian-hebel Please take a look at that. @mmkow Please confirm if you believe it is finished |
|
@mmkow Is it ready for review or not? |
|
No, it still needs some work. I've changed its status to wip. |
|
@mmkow Ok. Make sure to ping me and @krystian-hebel once ready. |
| * CHIP_IS_NODE = MODE1 = default | ||
| * CHIP_IS_GROUP = MODE2 | ||
| */ | ||
| uint8_t pump_mode = CHIP_IS_GROUP; /* ATTR_PROC_FABRIC_PUMP_MODE MODE2 from talos.xml */ |
There was a problem hiding this comment.
This should be read from here: https://github.com/open-power/hostboot/blob/master/src/include/usr/initservice/mboxRegs.H#L163
Value from talos.xml is written by SBE, but it may change for different platforms or even different versions of SBE.
| if (!p9_calculate_frequencies(&core_floor_ratio, &core_ceiling_ratio, | ||
| &freq_fbc, &freq_core_ceiling)) { |
There was a problem hiding this comment.
freq_fbc is uninitialized here, it is later compared in p9_calculate_frequencies() and never written to.
| uint32_t freq_core_floor = 4800; | ||
| /* According to src/usr/targeting/common/xmltohb/attribute_types.xml | ||
| ATTR_FREQ_CORE_NOMINAL_MHZ maps to NOMINAL_FREQ_MHZ present in talos.xml */ | ||
| uint32_t freq_core_nom = 4800; | ||
| uint8_t async_safe_mode = FABRIC_ASYNC_PERFORMANCE_MODE; | ||
| *freq_ceiling = 4800; |
There was a problem hiding this comment.
https://github.com/3mdeb/openpower-coreboot-docs/blob/main/logs/scom_dumps/istep_6_12_host_voltage_config.log#L40 and surrounding lines. Obviously the default values are not the same as actual.
No description provided.