Patch Set 19: Code-Review+2

so we took a deeper dive into GPIO discrepancies between vendor bios and coreboot.
we compared outputs of inteltool -G on vendor bios and with coreboot running this patch stack.
there are certain differences:

sdiff -w 100 -s tiogapass_coreboot_inteltool_gpio.txt tiogapass_vendorbios_inteltool_gpio.txt
  0x0080: 0x00000000 (HOSTSW_OWN_GPP_A)	      |	0x0080: 0x00f7ffff (HOSTSW_OWN_GPP_A)
  0x0084: 0x00000000 (HOSTSW_OWN_GPP_B)	      |	0x0084: 0x00ffffff (HOSTSW_OWN_GPP_B)
  0x0088: 0x00000000 (HOSTSW_OWN_GPP_F)	      |	0x0088: 0x00ffffff (HOSTSW_OWN_GPP_F)
  0x0060: 0x00000000 (PADCFGLOCK_GPP_C_0)     |	0x0060: 0x00000200 (PADCFGLOCK_GPP_C_0)
  0x0064: 0x00000000 (PADCFGLOCKTX_GPP_C_0)   |	0x0064: 0x00000200 (PADCFGLOCKTX_GPP_C_0)
  0x0070: 0x00000000 (PADCFGLOCK_GPP_E_0)     |	0x0070: 0x00000003 (PADCFGLOCK_GPP_E_0)
  0x0080: 0x00000000 (HOSTSW_OWN_GPP_C_0)     |	0x0080: 0x002fbb04 (HOSTSW_OWN_GPP_C_0)
  0x0084: 0x00000000 (HOSTSW_OWN_GPP_D_0)     |	0x0084: 0x00fffffe (HOSTSW_OWN_GPP_D_0)
  0x0088: 0x00000000 (HOSTSW_OWN_GPP_E_0)     |	0x0088: 0x00001fff (HOSTSW_OWN_GPP_E_0)
  0x0108: 0x00000000 (GPI_IS_GPP_E_0)	      |	0x0108: 0x00000003 (GPI_IS_GPP_E_0)
  0x0128: 0x00000003 (GPI_GPE_STS_GPP_E_0)    |	0x0128: 0x00000000 (GPI_GPE_STS_GPP_E_0)
  0x0140: 0x00c00000 (GPI_SMI_STS_GPP_C_0)    |	0x0140: 0x00400000 (GPI_SMI_STS_GPP_C_0)
  0x0144: 0x00000001 (GPI_SMI_STS_GPP_D_0)    |	0x0144: 0x00000000 (GPI_SMI_STS_GPP_D_0)
  0x0148: 0x00000003 (GPI_SMI_STS_GPP_E_0)    |	0x0148: 0x00000000 (GPI_SMI_STS_GPP_E_0)
  0x0154: 0x00000000 (GPI_SMI_EN_GPP_D_0)     |	0x0154: 0x00000001 (GPI_SMI_EN_GPP_D_0)
  0x0560: 0x0000004444000100 GPP_D20      GPIO|	0x0560: 0x0000004444000102 GPP_D20      GPIO 
  0x0080: 0x00000000 (HOSTSW_OWN_GPD_0)	      |	0x0080: 0x00000ffe (HOSTSW_OWN_GPD_0)
  0x0080: 0x00000000 (HOSTSW_OWN_GPP_I_0)     |	0x0080: 0x000007ff (HOSTSW_OWN_GPP_I_0)
  0x0400: 0x0000001844000902 GPP_I0       LAN_|	0x0400: 0x0000001844000900 GPP_I0       LAN_T
  0x0080: 0x00000000 (HOSTSW_OWN_GPP_J)	      |	0x0080: 0x00ffffff (HOSTSW_OWN_GPP_J)
  0x0084: 0x00000000 (HOSTSW_OWN_GPP_K)	      |	0x0084: 0x000007ff (HOSTSW_OWN_GPP_K)
  0x0440: 0x0000002044000600 GPP_J8       LAN |	0x0440: 0x0000002044000602 GPP_J8       LAN_I
  0x0450: 0x0000002244000600 GPP_J10      LAN |	0x0450: 0x0000002244000602 GPP_J10      LAN_I
  0x0460: 0x0000002444000600 GPP_J12      LAN |	0x0460: 0x0000002444000602 GPP_J12      LAN_I
  0x0470: 0x0000002644000600 GPP_J14      LAN |	0x0470: 0x0000002644000602 GPP_J14      LAN_I
  0x04c0: 0x0000003044000102 GPP_K0       GPI |	0x04c0: 0x0000003044000100 GPP_K0       GPIO 
  0x0060: 0x00000000 (PADCFGLOCK_GPP_G)	      |	0x0060: 0x00400000 (PADCFGLOCK_GPP_G)
  0x0064: 0x00000000 (PADCFGLOCKTX_GPP_G)     |	0x0064: 0x00400000 (PADCFGLOCKTX_GPP_G)
  0x0080: 0x00000000 (HOSTSW_OWN_GPP_G)	      |	0x0080: 0x00efffff (HOSTSW_OWN_GPP_G)
  0x0084: 0x00000000 (HOSTSW_OWN_GPP_H)	      |	0x0084: 0x00fc93df (HOSTSW_OWN_GPP_H)
  0x0088: 0x00000000 (HOSTSW_OWN_GPP_L)	      |	0x0088: 0x000ffffc (HOSTSW_OWN_GPP_L)
  0x0570: 0x0000004f44000100 GPP_H22      GPIO|	0x0570: 0x0000004f44000102 GPP_H22      GPIO 
  0x0578: 0x0000005044000100 GPP_H23      GPIO|	0x0578: 0x0000005044000102 GPP_H23      GPIO

So it looks like the difference is mostly in pad ownership. Common gpio code seems to have rather convoluted logic w.r.t setting pad ownership. We may need to add additional macros if we want to replicate pad config with 100% fidelity. But I think this may not be needed. At any rate I think we have good confidence this patch correctly defines all the offsets and community numbers, so I think it can land

Ok, thanks. I will try to resolve problem with pad ownership.

View Change

1 comment:

• Commit Message:

  • Patch Set #15, Line 26:

    Waiting for test results

    Done

To view, visit change 39425. To unsubscribe, or for help writing mail filters, visit settings.