Alertl for sbtsi #141
Alertl for sbtsi #141
Conversation
| @@ -301,10 +301,8 @@ static int sbtsi_i3c_probe(struct i3c_device *i3cdev) | |||
| struct regmap *regmap; | |||
There was a problem hiding this comment.
Commit message: please start with drivers:misc
| (i3cdev->desc->info.pid == 0x118) || (i3cdev->desc->info.pid == 0x010118) || | ||
| (i3cdev->desc->info.pid == 0x01000118) || (i3cdev->desc->info.pid == 0x01010118))) | ||
| { | ||
| if (!(I3C_PID_INSTANCE_ID(i3cdev->desc->info.pid) == 0 || |
There was a problem hiding this comment.
This is going to be good for both sp5 and sp7 based platforms . If not Need to look option to Valid PID bases check instead of hard coding.
static const u64 valid_sbtsi_pids[] = {
0x22400000001ULL,
// Add more if needed
};
There was a problem hiding this comment.
@ojayanth
PID: 0x22400000001 is constant till Turin irrespective of socket number or die number.
From Venice the PID is changed to include, socket and die number info.
As there is only one PID till Turin, my understanding is, currently we can have the changes as is.
| @@ -421,6 +421,10 @@ static const struct i3c_device_id sbtsi_i3c_id[] = { | |||
| I3C_DEVICE_EXTRA_INFO(0, 0x0101, 0x118, NULL), /* P1 - IOD1 - SBTSI */ | |||
There was a problem hiding this comment.
Please follow commit message align to other commits, in the drivers.
| @@ -421,6 +421,10 @@ static const struct i3c_device_id sbtsi_i3c_id[] = { | |||
| I3C_DEVICE_EXTRA_INFO(0, 0x0101, 0x118, NULL), /* P1 - IOD1 - SBTSI */ | |||
| I3C_DEVICE_EXTRA_INFO(0x112, 0, 0x1, NULL), | |||
| I3C_DEVICE_EXTRA_INFO(0, 0x0, 0x0, NULL), | |||
| I3C_DEVICE_EXTRA_INFO(0x112, 0x0, 0x118, NULL), /* Socket:0, IOD:0 */ | |||
There was a problem hiding this comment.
what is the diffrence between PX vs Socket?
There was a problem hiding this comment.
I am not sure what is P here? is it processor?
we are following terminology as used in PPR
From PPR:
Table 186: SBI Address Encoding
Socket ID / Socket number SB-RMI Address Comment
000b 3Ch IOD0 Socket 0 SB-RMI address
001b 38h IOD0 Socket 1 SB-RMI address
100b 34h IOD1 Socket 0 SB-RMI address
101b 35h IOD1 Socket 1 SB-RMI address
| @@ -299,6 +299,7 @@ static int sbtsi_i3c_probe(struct i3c_device *i3cdev) | |||
| .val_bits = 8, | |||
There was a problem hiding this comment.
Commit message - srtart with driver:misc:
| @@ -324,8 +325,29 @@ static int sbtsi_i3c_probe(struct i3c_device *i3cdev) | |||
| tsi_dev->regmap = regmap; | |||
| mutex_init(&tsi_dev->lock); | |||
|
|
|||
| /* Need to verify for the static address for i3cdev */ | |||
There was a problem hiding this comment.
Use Helper Function + Lookup Table approch to avoid duplication here and line 389
static const char *sbtsi_addr_to_label(u8 addr)
{
switch (addr) {
case 0x44: return "0.1";
case 0x45: return "1.1";
case 0x48: return "1.0";
case 0x4c: return "0.0";
default: return NULL;
}
}
- From Fam: 0x1A, Model: 0x50, SBTSI device instance ID is 0, additionally PID for each silicon is different. SBRMI and SBTSI device PID's difference is in Instance ID. Modify the probe condition on basis of PID for legacy platform and on instance ID from Fam:0x1A Model:0x50. Reviewed-by: Naveen Krishna Chatradhi <[email protected]> Signed-off-by: Akshay Gupta <[email protected]>
- SBRMI/TSI devices for new APML has PID as per below description.
[11:0] Additional ID: 0x118
[15:12] SB-TSI(0x0)
[31:16] [16]: DIE_ID, [25:24]: SOCKET ID
sock 0 die 0: 0x0
sock 0 die 1: 0x1
sock 1 die 0: 0x100
sock 1 die 1: 0x101
[32]: PID type select, it is 0 now to select fixed vendor value;
[47:33]: Manufacture ID, MIPI alliance has allocated the value for each company,
AMD should be 0x0112
Reviewed-by: Naveen Krishna Chatradhi <[email protected]>
Signed-off-by: Akshay Gupta <[email protected]>
Data name label under hwmon for i2c is sbtsi and for i3c is sbtsi_i3c. In multi socket, this makes difficult to distinguish between socket 0 and socket 1. Hwmon entry names are updated for i2c/i3c to sbtsi_X.Y, where X is socket index and Y is IOD index in the socket Reviewed-by: Naveen Krishna Chatradhi <[email protected]> Signed-off-by: Akshay Gupta <[email protected]>
sathyapk
force-pushed
the
alertl_for_sbtsi
branch
from
1f9426e to
9472f32
Compare
splitting the previous pr 135