On 03/17/2017 03:00 AM, David Gibson wrote:
> On Thu, Mar 16, 2017 at 02:52:17PM +0100, Cédric Le Goater wrote:
>> On 03/15/2017 07:16 AM, David Gibson wrote:
>>> On Wed, Mar 08, 2017 at 11:52:49AM +0100, Cédric Le Goater wrote:
>>>> From: Benjamin Herrenschmidt <b...@kernel.crashing.org>
>>>>
>>>> The PSI (Processor Service Interface) Controller is one of the engines
>>>> of the "Bridge" unit which connects the different interfaces to the
>>>> Power Processor.
>>>>
>>>> This adds just enough of the PSI bridge to handle various on-chip and
>>>> the one external interrupt. The rest of PSI has to do with the link to
>>>> the IBM FSP service processor which we don't plan to emulate (not used
>>>> on OpenPower machines).
>>>>
>>>> Signed-off-by: Benjamin Herrenschmidt <b...@kernel.crashing.org>
>>>> [clg: - updated for qemu-2.9
>>>> - changed the XSCOM interface to fit new model
>>>> - QOMified the model
>>>> - reworked set_xive and worked around a skiboot bug
>>>> - removed the 'psi_mmio_to_xscom' mapping array ]
>>>> Signed-off-by: Cédric Le Goater <c...@kaod.org>
>>>> ---
>>>> hw/ppc/Makefile.objs | 2 +-
>>>> hw/ppc/pnv.c | 35 ++-
>>>> hw/ppc/pnv_psi.c | 583
>>>> +++++++++++++++++++++++++++++++++++++++++++++
>>>> include/hw/ppc/pnv.h | 8 +
>>>> include/hw/ppc/pnv_psi.h | 61 +++++
>>>> include/hw/ppc/pnv_xscom.h | 3 +
>>>> 6 files changed, 685 insertions(+), 7 deletions(-)
>>>> create mode 100644 hw/ppc/pnv_psi.c
>>>> create mode 100644 include/hw/ppc/pnv_psi.h
>>>>
>>>> diff --git a/hw/ppc/Makefile.objs b/hw/ppc/Makefile.objs
>>>> index 001293423c8d..dc19ee17fa57 100644
>>>> --- a/hw/ppc/Makefile.objs
>>>> +++ b/hw/ppc/Makefile.objs
>>>> @@ -6,7 +6,7 @@ obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o
>>>> spapr_rtas.o
>>>> obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o spapr_rng.o
>>>> obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o
>>>> # IBM PowerNV
>>>> -obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o
>>>> +obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o
>>>> ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)
>>>> obj-y += spapr_pci_vfio.o
>>>> endif
>>>> diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c
>>>> index 0ae11cc3a2ca..85b00bf235c6 100644
>>>> --- a/hw/ppc/pnv.c
>>>> +++ b/hw/ppc/pnv.c
>>>> @@ -356,15 +356,22 @@ static void ppc_powernv_reset(void)
>>>> * have a CPLD that will collect the SerIRQ and shoot them as a
>>>> * single level interrupt to the P8 chip. So let's setup a hook
>>>> * for doing just that.
>>>> - *
>>>> - * Note: The actual interrupt input isn't emulated yet, this will
>>>> - * come with the PSI bridge model.
>>>> */
>>>> static void pnv_lpc_isa_irq_handler_cpld(void *opaque, int n, int level)
>>>> {
>>>> - /* We don't yet emulate the PSI bridge which provides the external
>>>> - * interrupt, so just drop interrupts on the floor
>>>> - */
>>>> + PnvMachineState *pnv = POWERNV_MACHINE(qdev_get_machine());
>>>> + uint32_t old_state = pnv->cpld_irqstate;
>>>> + PnvChip *chip = opaque;
>>>> +
>>>> + if (level) {
>>>> + pnv->cpld_irqstate |= 1u << n;
>>>> + } else {
>>>> + pnv->cpld_irqstate &= ~(1u << n);
>>>> + }
>>>> + if (pnv->cpld_irqstate != old_state) {
>>>> + pnv_psi_irq_set(&chip->psi, PSIHB_IRQ_EXTERNAL,
>>>> + pnv->cpld_irqstate != 0);
>>>> + }
>>>> }
>>>>
>>>> static void pnv_lpc_isa_irq_handler(void *opaque, int n, int level)
>>>> @@ -702,6 +709,11 @@ static void pnv_chip_init(Object *obj)
>>>>
>>>> object_initialize(&chip->lpc, sizeof(chip->lpc), TYPE_PNV_LPC);
>>>> object_property_add_child(obj, "lpc", OBJECT(&chip->lpc), NULL);
>>>> +
>>>> + object_initialize(&chip->psi, sizeof(chip->psi), TYPE_PNV_PSI);
>>>> + object_property_add_child(obj, "psi", OBJECT(&chip->psi), NULL);
>>>> + object_property_add_const_link(OBJECT(&chip->psi), "xics",
>>>> + OBJECT(qdev_get_machine()),
>>>> &error_abort);
>>>> }
>>>>
>>>> static void pnv_chip_icp_realize(PnvChip *chip, Error **errp)
>>>> @@ -722,6 +734,8 @@ static void pnv_chip_realize(DeviceState *dev, Error
>>>> **errp)
>>>> char *typename = pnv_core_typename(pcc->cpu_model);
>>>> size_t typesize = object_type_get_instance_size(typename);
>>>> int i, core_hwid;
>>>> + MachineState *machine = MACHINE(qdev_get_machine());
>>>> + PnvMachineState *pnv = POWERNV_MACHINE(machine);
>>>>
>>>> if (!object_class_by_name(typename)) {
>>>> error_setg(errp, "Unable to find PowerNV CPU Core '%s'",
>>>> typename);
>>>> @@ -797,6 +811,15 @@ static void pnv_chip_realize(DeviceState *dev, Error
>>>> **errp)
>>>> }
>>>> g_free(typename);
>>>>
>>>> +
>>>> + /* Processor Service Interface (PSI) Host Bridge */
>>>> + object_property_set_bool(OBJECT(&chip->psi), true, "realized",
>>>> + &error_fatal);
>>>> + pnv_xscom_add_subregion(chip, PNV_XSCOM_PSI_BASE,
>>>> &chip->psi.xscom_regs);
>>>> +
>>>> + /* link in the PSI ICS */
>>>> + QLIST_INSERT_HEAD(&pnv->ics, &chip->psi.ics, list);
>>>> +
>>>> /* Create LPC controller */
>>>> object_property_set_bool(OBJECT(&chip->lpc), true, "realized",
>>>> &error_fatal);
>>>> diff --git a/hw/ppc/pnv_psi.c b/hw/ppc/pnv_psi.c
>>>> new file mode 100644
>>>> index 000000000000..6ba688aac075
>>>> --- /dev/null
>>>> +++ b/hw/ppc/pnv_psi.c
>>>> @@ -0,0 +1,583 @@
>>>> +/*
>>>> + * QEMU PowerNV PowerPC PSI interface
>>>> + *
>>>> + * Copyright (c) 2016, IBM Corporation
>>>> + *
>>>> + * This library is free software; you can redistribute it and/or
>>>> + * modify it under the terms of the GNU Lesser General Public
>>>> + * License as published by the Free Software Foundation; either
>>>> + * version 2 of the License, or (at your option) any later version.
>>>> + *
>>>> + * This library is distributed in the hope that it will be useful,
>>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
>>>> + * Lesser General Public License for more details.
>>>> + *
>>>> + * You should have received a copy of the GNU Lesser General Public
>>>> + * License along with this library; if not, see
>>>> <http://www.gnu.org/licenses/>.
>>>> + */
>>>> +
>>>> +#include "qemu/osdep.h"
>>>> +#include "hw/hw.h"
>>>> +#include "target/ppc/cpu.h"
>>>> +#include "qemu/log.h"
>>>> +#include "qapi/error.h"
>>>> +
>>>> +#include "exec/address-spaces.h"
>>>> +
>>>> +#include "hw/ppc/fdt.h"
>>>> +#include "hw/ppc/pnv.h"
>>>> +#include "hw/ppc/pnv_xscom.h"
>>>> +#include "hw/ppc/pnv_psi.h"
>>>> +
>>>> +#include <libfdt.h>
>>>> +
>>>> +#define PSIHB_XSCOM_FIR_RW 0x00
>>>> +#define PSIHB_XSCOM_FIR_AND 0x01
>>>> +#define PSIHB_XSCOM_FIR_OR 0x02
>>>> +#define PSIHB_XSCOM_FIRMASK_RW 0x03
>>>> +#define PSIHB_XSCOM_FIRMASK_AND 0x04
>>>> +#define PSIHB_XSCOM_FIRMASK_OR 0x05
>>>> +#define PSIHB_XSCOM_FIRACT0 0x06
>>>> +#define PSIHB_XSCOM_FIRACT1 0x07
>>>> +#define PSIHB_XSCOM_BAR 0x0a
>>>> +#define PSIHB_BAR_EN 0x0000000000000001ull
>>>> +#define PSIHB_XSCOM_FSPBAR 0x0b
>>>> +#define PSIHB_XSCOM_CR 0x0e
>>>> +#define PSIHB_CR_FSP_CMD_ENABLE 0x8000000000000000ull
>>>> +#define PSIHB_CR_FSP_MMIO_ENABLE 0x4000000000000000ull
>>>> +#define PSIHB_CR_FSP_IRQ_ENABLE 0x1000000000000000ull
>>>> +#define PSIHB_CR_FSP_ERR_RSP_ENABLE 0x0800000000000000ull
>>>> +#define PSIHB_CR_PSI_LINK_ENABLE 0x0400000000000000ull
>>>> +#define PSIHB_CR_FSP_RESET 0x0200000000000000ull
>>>> +#define PSIHB_CR_PSIHB_RESET 0x0100000000000000ull
>>>> +#define PSIHB_CR_PSI_IRQ 0x0000800000000000ull
>>>> +#define PSIHB_CR_FSP_IRQ 0x0000400000000000ull
>>>> +#define PSIHB_CR_FSP_LINK_ACTIVE 0x0000200000000000ull
>>>> + /* and more ... */
>>>> +#define PSIHB_XSCOM_SEMR 0x0f
>>>> +#define PSIHB_XSCOM_XIVR_PSI 0x10
>>>> +#define PSIHB_XIVR_SERVER_SH 40
>>>> +#define PSIHB_XIVR_SERVER_MSK (0xffffull << PSIHB_XIVR_SERVER_SH)
>>>> +#define PSIHB_XIVR_PRIO_SH 32
>>>> +#define PSIHB_XIVR_PRIO_MSK (0xffull << PSIHB_XIVR_PRIO_SH)
>>>> +#define PSIHB_XIVR_SRC_SH 29
>>>> +#define PSIHB_XIVR_SRC_MSK (0x7ull << PSIHB_XIVR_SRC_SH)
>>>> +#define PSIHB_XIVR_PENDING 0x01000000ull
>>>> +#define PSIHB_XSCOM_SCR 0x12
>>>> +#define PSIHB_XSCOM_CCR 0x13
>>>> +#define PSIHB_XSCOM_DMA_UPADD 0x14
>>>> +#define PSIHB_XSCOM_IRQ_STAT 0x15
>>>> +#define PSIHB_IRQ_STAT_OCC 0x0000001000000000ull
>>>> +#define PSIHB_IRQ_STAT_FSI 0x0000000800000000ull
>>>> +#define PSIHB_IRQ_STAT_LPCI2C 0x0000000400000000ull
>>>> +#define PSIHB_IRQ_STAT_LOCERR 0x0000000200000000ull
>>>> +#define PSIHB_IRQ_STAT_EXT 0x0000000100000000ull
>>>> +#define PSIHB_XSCOM_XIVR_OCC 0x16
>>>> +#define PSIHB_XSCOM_XIVR_FSI 0x17
>>>> +#define PSIHB_XSCOM_XIVR_LPCI2C 0x18
>>>> +#define PSIHB_XSCOM_XIVR_LOCERR 0x19
>>>> +#define PSIHB_XSCOM_XIVR_EXT 0x1a
>>>> +#define PSIHB_XSCOM_IRSN 0x1b
>>>> +#define PSIHB_IRSN_COMP_SH 45
>>>> +#define PSIHB_IRSN_COMP_MSK (0x7ffffull << PSIHB_IRSN_COMP_SH)
>>>> +#define PSIHB_IRSN_IRQ_MUX 0x0000000800000000ull
>>>> +#define PSIHB_IRSN_IRQ_RESET 0x0000000400000000ull
>>>> +#define PSIHB_IRSN_DOWNSTREAM_EN 0x0000000200000000ull
>>>> +#define PSIHB_IRSN_UPSTREAM_EN 0x0000000100000000ull
>>>> +#define PSIHB_IRSN_COMPMASK_SH 13
>>>> +#define PSIHB_IRSN_COMPMASK_MSK (0x7ffffull <<
>>>> PSIHB_IRSN_COMPMASK_SH)
>>>> +
>>>> +/*
>>>> + * These are the values of the registers when accessed through the
>>>> + * MMIO region. The relation is xscom = (mmio + 0x50) >> 3
>>>> + */
>>>> +#define PSIHB_MMIO_BAR 0x00
>>>> +#define PSIHB_MMIO_FSPBAR 0x08
>>>> +#define PSIHB_MMIO_CR 0x20
>>>> +#define PSIHB_MMIO_SEMR 0x28
>>>> +#define PSIHB_MMIO_XIVR_PSI 0x30
>>>> +#define PSIHB_MMIO_SCR 0x40
>>>> +#define PSIHB_MMIO_CCR 0x48
>>>> +#define PSIHB_MMIO_DMA_UPADD 0x50
>>>> +#define PSIHB_MMIO_IRQ_STAT 0x58
>>>> +#define PSIHB_MMIO_XIVR_OCC 0x60
>>>> +#define PSIHB_MMIO_XIVR_FSI 0x68
>>>> +#define PSIHB_MMIO_XIVR_LPCI2C 0x70
>>>> +#define PSIHB_MMIO_XIVR_LOCERR 0x78
>>>> +#define PSIHB_MMIO_XIVR_EXT 0x80
>>>> +#define PSIHB_MMIO_IRSN 0x88
>>>> +#define PSIHB_MMIO_MAX 0x100
>>>> +
>>>> +static void pnv_psi_set_bar(PnvPsi *psi, uint64_t bar)
>>>> +{
>>>> + MemoryRegion *sysmem = get_system_memory();
>>>> + uint64_t old = psi->regs[PSIHB_XSCOM_BAR];
>>>> +
>>>> + psi->regs[PSIHB_XSCOM_BAR] = bar & 0x0003fffffff00001;
>>>> +
>>>> + /* Update MR, always remove it first */
>>>> + if (old & PSIHB_BAR_EN) {
>>>> + memory_region_del_subregion(sysmem, &psi->regs_mr);
>>>> + }
>>>> + /* Then add it back if needed */
>>>> + if (bar & PSIHB_BAR_EN) {
>>>> + uint64_t addr = bar & 0x0003fffffff00000;
>>>> + memory_region_add_subregion(sysmem, addr, &psi->regs_mr);
>>>> + }
>>>> +}
>>>> +
>>>> +static void pnv_psi_update_fsp_mr(PnvPsi *psi)
>>>> +{
>>>> + /* XXX Update FSP MR if/when we support FSP BAR */
>>>> +}
>>>> +
>>>> +static void pnv_psi_set_cr(PnvPsi *psi, uint64_t cr)
>>>> +{
>>>> + uint64_t old = psi->regs[PSIHB_XSCOM_CR];
>>>> +
>>>> + psi->regs[PSIHB_XSCOM_CR] = cr & 0x0003ffff00000000;
>>>> +
>>>> + /* Check some bit changes */
>>>> + if ((old ^ psi->regs[PSIHB_XSCOM_CR]) & PSIHB_CR_FSP_MMIO_ENABLE) {
>>>> + pnv_psi_update_fsp_mr(psi);
>>>> + }
>>>> +}
>>>> +
>>>> +static void pnv_psi_set_irsn(PnvPsi *psi, uint64_t val)
>>>> +{
>>>> + uint32_t offset;
>>>> + ICSState *ics = &psi->ics;
>>>> +
>>>> + /* In this model we ignore the up/down enable bits for now
>>>> + * as SW doesn't use them (other than setting them at boot).
>>>> + * We ignore IRQ_MUX, its meaning isn't clear and we don't use
>>>> + * it and finally we ignore reset (XXX fix that ?)
>>>> + */
>>>> + psi->regs[PSIHB_XSCOM_IRSN] = val & (PSIHB_IRSN_COMP_MSK |
>>>> + PSIHB_IRSN_IRQ_MUX |
>>>> + PSIHB_IRSN_DOWNSTREAM_EN |
>>>> + PSIHB_IRSN_DOWNSTREAM_EN |
>>>> + PSIHB_IRSN_DOWNSTREAM_EN);
>>>> +
>>>> + /* We ignore the compare mask as well, our ICS emulation is too
>>>> + * simplistic to make any use if it, and we extract the offset
>>>> + * from the compare value
>>>> + */
>>>> + offset = (val & PSIHB_IRSN_COMP_MSK) >> PSIHB_IRSN_COMP_SH;
>>>> + ics->offset = offset;
>>>> +}
>>>> +
>>>> +static bool pnv_psi_irq_bits(PnvPsi *psi, PnvPsiIrq irq,
>>>> + uint32_t *out_xivr_reg,
>>>> + uint32_t *out_stat_reg,
>>>> + uint64_t *out_stat_bit)
>>>
>>> Your PnvPsiIrq values are arbitrary and contiguous AFACT. Why not
>>> just have a lookup table for this info, instead of a giant switch
>>> statement?
>>
>> Well, I agree but at the same time, we are not gaining much in terms
>> of lines by using an array,
>
> Hmm.. seems like an ~ x5 line reduction to me...
>
> And, IMO, easier to read.
OK. Will do. Hope you like it :)
>> and we have to check for boundaries which
>> the switch provide. The question would be different if we had more
>> parameters. So let's keep it that way.
>>
>>>> +{
>>>> + switch (irq) {
>>>> + case PSIHB_IRQ_PSI:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_PSI;
>>>> + *out_stat_reg = PSIHB_XSCOM_CR;
>>>> + *out_stat_bit = PSIHB_CR_PSI_IRQ;
>>>> + break;
>>>> + case PSIHB_IRQ_FSP:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_PSI;
>>>> + *out_stat_reg = PSIHB_XSCOM_CR;
>>>> + *out_stat_bit = PSIHB_CR_FSP_IRQ;
>>>> + break;
>>>> + case PSIHB_IRQ_OCC:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_OCC;
>>>> + *out_stat_reg = PSIHB_XSCOM_IRQ_STAT;
>>>> + *out_stat_bit = PSIHB_IRQ_STAT_OCC;
>>>> + break;
>>>> + case PSIHB_IRQ_FSI:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_FSI;
>>>> + *out_stat_reg = PSIHB_XSCOM_IRQ_STAT;
>>>> + *out_stat_bit = PSIHB_IRQ_STAT_FSI;
>>>> + break;
>>>> + case PSIHB_IRQ_LPC_I2C:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_LPCI2C;
>>>> + *out_stat_reg = PSIHB_XSCOM_IRQ_STAT;
>>>> + *out_stat_bit = PSIHB_IRQ_STAT_LPCI2C;
>>>> + break;
>>>> + case PSIHB_IRQ_LOCAL_ERR:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_LOCERR;
>>>> + *out_stat_reg = PSIHB_XSCOM_IRQ_STAT;
>>>> + *out_stat_bit = PSIHB_IRQ_STAT_LOCERR;
>>>> + break;
>>>> + case PSIHB_IRQ_EXTERNAL:
>>>> + *out_xivr_reg = PSIHB_XSCOM_XIVR_EXT;
>>>> + *out_stat_reg = PSIHB_XSCOM_IRQ_STAT;
>>>> + *out_stat_bit = PSIHB_IRQ_STAT_EXT;
>>>> + break;
>>>> + default:
>>>> + return false;
>>>> + }
>>>> + return true;
>>>> +}
>>>> +
>>>> +void pnv_psi_irq_set(PnvPsi *psi, PnvPsiIrq irq, bool state)
>>>> +{
>>>> + ICSState *ics = &psi->ics;
>>>> + uint32_t xivr_reg;
>>>> + uint32_t stat_reg;
>>>> + uint64_t stat_bit;
>>>> + uint32_t src;
>>>> + bool masked;
>>>> +
>>>> + if (!pnv_psi_irq_bits(psi, irq, &xivr_reg, &stat_reg, &stat_bit)) {
>>>> + qemu_log_mask(LOG_GUEST_ERROR, "PSI: Unsupported irq %d\n", irq);
>>>> + return;
>>>> + }
>>>> +
>>>> + src = (psi->regs[xivr_reg] & PSIHB_XIVR_SRC_MSK) >> PSIHB_XIVR_SRC_SH;
>>>> + masked = (psi->regs[xivr_reg] & PSIHB_XIVR_PRIO_MSK) ==
>>>> PSIHB_XIVR_PRIO_MSK;
>>>> + if (state) {
>>>> + psi->regs[stat_reg] |= stat_bit;
>>>> + /* XXX optimization: check mask here. That means re-evaluating
>>>> + * when unmasking, thus TODO
>>>> + */
>>>> + qemu_irq_raise(ics->qirqs[src]);
>>>> + } else {
>>>> + psi->regs[stat_reg] &= ~stat_bit;
>>>> +
>>>> + /* FSP and PSI are muxed so don't lower if either still set */
>>>> + if (stat_reg != PSIHB_XSCOM_CR ||
>>>> + !(psi->regs[stat_reg] & (PSIHB_CR_PSI_IRQ |
>>>> PSIHB_CR_FSP_IRQ))) {
>>>> + qemu_irq_lower(ics->qirqs[src]);
>>>> + } else {
>>>> + state = true;
>>>> + }
>>>> + }
>>>
>>> It might be cleaner to just revaluate the irq level from scratch here,
>>> and set the level, rather than doing this complicated dance to work
>>> out if it has changed.
>>
>> OK. I need to take a closer look at that.
>>
>>>> +
>>>> + /* XXX Note about the emulation of the pending bit: This isn't
>>>> + * entirely correct. The pending bit should be cleared when the
>>>> + * EOI has been received. However, we don't have callbacks on
>>>> + * EOI (especially not under KVM) so no way to emulate that
>>>> + * properly, so instead we just set that bit as the logical
>>>> + * "output" of the XIVR (ie pending & !masked)
>>>> + * XXX TODO: Also update it on set_xivr
>>>> + */
>>>> + if (state && !masked) {
>>>> + psi->regs[xivr_reg] |= PSIHB_XIVR_PENDING;
>>>> + } else {
>>>> + psi->regs[xivr_reg] &= ~PSIHB_XIVR_PENDING;
>>>> + }
>>>> +}
>>>> +
>>>> +static void pnv_psi_set_xivr(PnvPsi *psi, uint32_t reg, uint64_t val)
>>>> +{
>>>> + ICSState *ics = &psi->ics;
>>>> + uint16_t server;
>>>> + uint8_t prio;
>>>> + uint8_t src;
>>>> + int icp_index;
>>>> +
>>>> + psi->regs[reg] = (psi->regs[reg] & PSIHB_XIVR_PENDING) |
>>>> + (val & (PSIHB_XIVR_SERVER_MSK |
>>>> + PSIHB_XIVR_PRIO_MSK |
>>>> + PSIHB_XIVR_SRC_MSK));
>>>> + val = psi->regs[reg];
>>>> + server = (val & PSIHB_XIVR_SERVER_MSK) >> PSIHB_XIVR_SERVER_SH;
>>>> + prio = (val & PSIHB_XIVR_PRIO_MSK) >> PSIHB_XIVR_PRIO_SH;
>>>> + src = (val & PSIHB_XIVR_SRC_MSK) >> PSIHB_XIVR_SRC_SH;
>>>> + if (src > PSIHB_IRQ_EXTERNAL) {
>>>> + /* XXX Generate error ? */
>>>> + return;
>>>> + }
>>>> +
>>>> + /*
>>>> + * Linux fills the irq xivr with the hw processor id plus the
>>>> + * link bits. shift back to get something valid.
>>>> + */
>>>> + server >>= 2;
>>>> +
>>>> + /*
>>>> + * When skiboot initializes PSIHB, it fills the xives with
>>>> + * server=0, prio=0xff, but we don't have a CPU with a pir=0. So
>>>> + * skip that case.
>>>> + */
>>>> + if (prio != 0xff) {
>>>> + icp_index = xics_get_cpu_index_by_pir(server);
>>>> + assert(icp_index != -1);
>>>> + } else {
>>>> + if (server) {
>>>> + qemu_log_mask(LOG_GUEST_ERROR, "PSI: bogus server %d for IRQ
>>>> %d\n",
>>>> + server, src);
>>>> + }
>>>> + icp_index = server;
>>>> + }
>>>
>>> This logic doesn't seem like it belongs here. You've received a
>>> server number, seems like you should just pass it on to the xics code,
>>> and if there can be an error, have that detect it.
>>
>> I have removed all of that in a private patch. This is tracking
>> an issue in skiboot which did not clear correctly the PSI xive.
>> It is partially resolved in recent version but I still see some
>> warnings whe the guest reboots.
>
> Ok. Workarounds for firmware bugs are fine, but they need a detailed
> comment so other people have a hope of understanding why they're
> there. Including stating outright that it is a bug workaround, rather
> than expected firmware behaviour.
Yes. I will included a detailed comment in an extra patch if this is
still needed.
Thanks,
C.
>>
>>>> +
>>>> + /* Now because of source remapping, weird things can happen
>>>> + * if you change the source number dynamically, our simple ICS
>>>> + * doesn't deal with remapping. So we just poke a different
>>>> + * ICS entry based on what source number was written. This will
>>>> + * do for now but a more accurate implementation would instead
>>>> + * use a fixed server/prio and a remapper of the generated irq.
>>>> + */
>>>> + ics_simple_write_xive(ics, src, icp_index, prio, prio);
>>>> +}
>>>> +
>>>> +static uint64_t pnv_psi_reg_read(PnvPsi *psi, uint32_t offset, bool mmio)
>>>> +{
>>>> + uint64_t val = 0xffffffffffffffffull;
>>>> +
>>>> + switch (offset) {
>>>> + case PSIHB_XSCOM_FIR_RW:
>>>> + case PSIHB_XSCOM_FIRACT0:
>>>> + case PSIHB_XSCOM_FIRACT1:
>>>> + case PSIHB_XSCOM_BAR:
>>>> + case PSIHB_XSCOM_FSPBAR:
>>>> + case PSIHB_XSCOM_CR:
>>>> + case PSIHB_XSCOM_XIVR_PSI:
>>>> + case PSIHB_XSCOM_XIVR_OCC:
>>>> + case PSIHB_XSCOM_XIVR_FSI:
>>>> + case PSIHB_XSCOM_XIVR_LPCI2C:
>>>> + case PSIHB_XSCOM_XIVR_LOCERR:
>>>> + case PSIHB_XSCOM_XIVR_EXT:
>>>> + case PSIHB_XSCOM_IRQ_STAT:
>>>> + case PSIHB_XSCOM_SEMR:
>>>> + case PSIHB_XSCOM_DMA_UPADD:
>>>> + case PSIHB_XSCOM_IRSN:
>>>> + val = psi->regs[offset];
>>>> + break;
>>>> + default:
>>>> + qemu_log_mask(LOG_UNIMP, "PSI Unimplemented register: Ox%" PRIx32
>>>> "\n",
>>>> + offset);
>>>
>>> As noted elsewhere, tracepoints are more usual than qemu_log() these
>>> days. But either way, this really should have a distinguishable
>>> message from the one in the write path.
>>
>> OK. Will add a read/write statement.
>>
>>>> + }
>>>> + return val;
>>>> +}
>>>> +
>>>> +static void pnv_psi_reg_write(PnvPsi *psi, uint32_t offset, uint64_t val,
>>>> + bool mmio)
>>>> +{
>>>> + switch (offset) {
>>>> + case PSIHB_XSCOM_FIR_RW:
>>>> + case PSIHB_XSCOM_FIRACT0:
>>>> + case PSIHB_XSCOM_FIRACT1:
>>>> + case PSIHB_XSCOM_SEMR:
>>>> + case PSIHB_XSCOM_DMA_UPADD:
>>>> + psi->regs[offset] = val;
>>>> + break;
>>>> + case PSIHB_XSCOM_FIR_OR:
>>>> + psi->regs[PSIHB_XSCOM_FIR_RW] |= val;
>>>> + break;
>>>> + case PSIHB_XSCOM_FIR_AND:
>>>> + psi->regs[PSIHB_XSCOM_FIR_RW] &= val;
>>>> + break;
>>>> + case PSIHB_XSCOM_BAR:
>>>> + /* Only XSCOM can write this one */
>>>> + if (!mmio) {
>>>> + pnv_psi_set_bar(psi, val);
>>>> + }
>>>> + break;
>>>> + case PSIHB_XSCOM_FSPBAR:
>>>> + psi->regs[PSIHB_XSCOM_BAR] = val & 0x0003ffff00000000;
>>>
>>> Should that be PSIHB_XSCOM_FSPBAR?
>>
>> yes ...
>>
>>>> + pnv_psi_update_fsp_mr(psi);
>>>> + break;
>>>> + case PSIHB_XSCOM_CR:
>>>> + pnv_psi_set_cr(psi, val);
>>>> + break;
>>>> + case PSIHB_XSCOM_SCR:
>>>> + pnv_psi_set_cr(psi, psi->regs[PSIHB_XSCOM_CR] | val);
>>>> + break;
>>>> + case PSIHB_XSCOM_CCR:
>>>> + pnv_psi_set_cr(psi, psi->regs[PSIHB_XSCOM_CR] & ~val);
>>>> + break;
>>>> + case PSIHB_XSCOM_XIVR_PSI:
>>>> + case PSIHB_XSCOM_XIVR_OCC:
>>>> + case PSIHB_XSCOM_XIVR_FSI:
>>>> + case PSIHB_XSCOM_XIVR_LPCI2C:
>>>> + case PSIHB_XSCOM_XIVR_LOCERR:
>>>> + case PSIHB_XSCOM_XIVR_EXT:
>>>> + pnv_psi_set_xivr(psi, offset, val);
>>>> + break;
>>>> + case PSIHB_XSCOM_IRQ_STAT:
>>>> + /* Read only, should we generate an error ? */
>>>> + break;
>>>> + case PSIHB_XSCOM_IRSN:
>>>> + pnv_psi_set_irsn(psi, val);
>>>> + break;
>>>> + default:
>>>> + qemu_log_mask(LOG_UNIMP, "PSI Unimplemented register: Ox%" PRIx32
>>>> "\n",
>>>> + offset);
>>>> + }
>>>> +}
>>>> +
>>>> +static inline uint32_t psi_mmio_to_xscom(hwaddr addr)
>>>> +{
>>>> + return (addr >> 3) + PSIHB_XSCOM_BAR;
>>>> +}
>>>> +
>>>> +static uint64_t pnv_psi_mmio_read(void *opaque, hwaddr addr, unsigned
>>>> size)
>>>> +{
>>>> + return pnv_psi_reg_read(opaque, psi_mmio_to_xscom(addr), true);
>>>> +}
>>>> +
>>>> +static void pnv_psi_mmio_write(void *opaque, hwaddr addr,
>>>> + uint64_t val, unsigned size)
>>>> +{
>>>> + pnv_psi_reg_write(opaque, psi_mmio_to_xscom(addr), val, true);
>>>> +}
>>>> +
>>>> +static const MemoryRegionOps psi_mmio_ops = {
>>>> + .read = pnv_psi_mmio_read,
>>>> + .write = pnv_psi_mmio_write,
>>>> + .endianness = DEVICE_BIG_ENDIAN,
>>>> + .valid = {
>>>> + .min_access_size = 8,
>>>> + .max_access_size = 8,
>>>> + },
>>>> + .impl = {
>>>> + .min_access_size = 8,
>>>> + .max_access_size = 8,
>>>> + },
>>>> +};
>>>> +
>>>> +static uint64_t pnv_psi_xscom_read(void *opaque, hwaddr addr, unsigned
>>>> size)
>>>> +{
>>>> + PnvPsi *psi = PNV_PSI(opaque);
>>>> + uint32_t offset = addr >> 3;
>>>> +
>>>> + return pnv_psi_reg_read(psi, offset, false);
>>>> +}
>>>> +
>>>> +static void pnv_psi_xscom_write(void *opaque, hwaddr addr,
>>>> + uint64_t val, unsigned size)
>>>> +{
>>>> + PnvPsi *psi = PNV_PSI(opaque);
>>>> + uint32_t offset = addr >> 3;
>>>> +
>>>> + pnv_psi_reg_write(psi, offset, val, false);
>>>> +}
>>>> +
>>>> +static const MemoryRegionOps pnv_psi_xscom_ops = {
>>>> + .read = pnv_psi_xscom_read,
>>>> + .write = pnv_psi_xscom_write,
>>>> + .valid.min_access_size = 8,
>>>> + .valid.max_access_size = 8,
>>>> + .impl.min_access_size = 8,
>>>> + .impl.max_access_size = 8,
>>>
>>> Consistent nesting format of the two MemoryRegionOps would be good.
>>
>> OK.
>>
>>>> + .endianness = DEVICE_BIG_ENDIAN,
>>>> +};
>>>> +
>>>> +static void pnv_psi_init(Object *obj)
>>>> +{
>>>> + PnvPsi *psi = PNV_PSI(obj);
>>>> +
>>>> + object_initialize(&psi->ics, sizeof(psi->ics), TYPE_ICS_SIMPLE);
>>>> + qdev_set_parent_bus(DEVICE(&psi->ics), sysbus_get_default());
>>>> + object_property_add_child(obj, "ics-psi", OBJECT(&psi->ics), NULL);
>>>> +}
>>>> +
>>>> +static void pnv_psi_realize(DeviceState *dev, Error **errp)
>>>> +{
>>>> + PnvPsi *psi = PNV_PSI(dev);
>>>> + ICSState *ics = &psi->ics;
>>>> + Object *obj;
>>>> + Error *err = NULL, *local_err = NULL;
>>>> + unsigned int i;
>>>> +
>>>> + /* Initialize MMIO region */
>>>> + memory_region_init_io(&psi->regs_mr, OBJECT(dev), &psi_mmio_ops, psi,
>>>> + "psihb", PNV_PSIHB_BAR_SIZE);
>>>> +
>>>> + /* Default BAR. Use object properties ? */
>>>> + pnv_psi_set_bar(psi, PNV_PSIHB_BAR | PSIHB_BAR_EN);
>>>> +
>>>> + /* Default sources in XIVR */
>>>> + psi->regs[PSIHB_XSCOM_XIVR_PSI] = PSIHB_XIVR_PRIO_MSK |
>>>> + (0ull << PSIHB_XIVR_SRC_SH);
>>>> + psi->regs[PSIHB_XSCOM_XIVR_OCC] = PSIHB_XIVR_PRIO_MSK |
>>>> + (1ull << PSIHB_XIVR_SRC_SH);
>>>> + psi->regs[PSIHB_XSCOM_XIVR_FSI] = PSIHB_XIVR_PRIO_MSK |
>>>> + (2ull << PSIHB_XIVR_SRC_SH);
>>>> + psi->regs[PSIHB_XSCOM_XIVR_LPCI2C] = PSIHB_XIVR_PRIO_MSK |
>>>> + (3ull << PSIHB_XIVR_SRC_SH);
>>>> + psi->regs[PSIHB_XSCOM_XIVR_LOCERR] = PSIHB_XIVR_PRIO_MSK |
>>>> + (4ull << PSIHB_XIVR_SRC_SH);
>>>> + psi->regs[PSIHB_XSCOM_XIVR_EXT] = PSIHB_XIVR_PRIO_MSK |
>>>> + (5ull << PSIHB_XIVR_SRC_SH);
>>>> +
>>>> + /* get XICSFabric from chip */
>>>> + obj = object_property_get_link(OBJECT(dev), "xics", &err);
>>>> + if (!obj) {
>>>> + error_setg(errp, "%s: required link 'xics' not found: %s",
>>>> + __func__, error_get_pretty(err));
>>>> + return;
>>>> + }
>>>> +
>>>> + /*
>>>> + * PSI interrupt control source
>>>> + */
>>>> + object_property_set_int(OBJECT(ics), PSI_NUM_INTERRUPTS, "nr-irqs",
>>>> &err);
>>>> + object_property_add_const_link(OBJECT(ics), "xics", obj, &err);
>>>> + object_property_set_bool(OBJECT(ics), true, "realized", &local_err);
>>>> + error_propagate(&err, local_err);
>>>> + if (err) {
>>>> + error_propagate(errp, err);
>>>> + return;
>>>> + }
>>>> +
>>>> + for (i = 0; i < ics->nr_irqs; i++) {
>>>> + ics_set_irq_type(ics, i, true);
>>>> + }
>>>> +
>>>> + /* XScom region for PSI registers */
>>>> + pnv_xscom_region_init(&psi->xscom_regs, OBJECT(dev),
>>>> &pnv_psi_xscom_ops,
>>>> + psi, "xscom-psi", PNV_XSCOM_PSI_SIZE);
>>>> +}
>>>> +
>>>> +static int pnv_psi_populate(PnvXScomInterface *dev, void *fdt, int
>>>> xscom_offset)
>>>> +{
>>>> + const char compat[] = "ibm,power8-psihb-x\0ibm,psihb-x";
>>>> + char *name;
>>>> + int offset;
>>>> + uint32_t lpc_pcba = PNV_XSCOM_PSI_BASE;
>>>> + uint32_t reg[] = {
>>>> + cpu_to_be32(lpc_pcba),
>>>> + cpu_to_be32(PNV_XSCOM_PSI_SIZE)
>>>> + };
>>>> +
>>>> + name = g_strdup_printf("psihb@%x", lpc_pcba);
>>>> + offset = fdt_add_subnode(fdt, xscom_offset, name);
>>>> + _FDT(offset);
>>>> + g_free(name);
>>>> +
>>>> + _FDT((fdt_setprop(fdt, offset, "reg", reg, sizeof(reg))));
>>>> +
>>>> + _FDT((fdt_setprop_cell(fdt, offset, "#address-cells", 2)));
>>>> + _FDT((fdt_setprop_cell(fdt, offset, "#size-cells", 1)));
>>>> + _FDT((fdt_setprop(fdt, offset, "compatible", compat,
>>>> + sizeof(compat))));
>>>> + return 0;
>>>> +}
>>>> +
>>>> +
>>>> +static void pnv_psi_class_init(ObjectClass *klass, void *data)
>>>> +{
>>>> + DeviceClass *dc = DEVICE_CLASS(klass);
>>>> + PnvXScomInterfaceClass *xdc = PNV_XSCOM_INTERFACE_CLASS(klass);
>>>> +
>>>> + xdc->populate = pnv_psi_populate;
>>>> +
>>>> + dc->realize = pnv_psi_realize;
>>>> +}
>>>> +
>>>> +static const TypeInfo pnv_psi_info = {
>>>> + .name = TYPE_PNV_PSI,
>>>> + .parent = TYPE_DEVICE,
>>>
>>> Since the PSI has an MMIO presence, it probably should be a
>>> SysBusDevice, rather than a raw descendent of TYPE_DEVICE.
>>
>> Yes indeed.
>>
>> So I will resend the patchset with just the XICS part. I want to
>> take a look at pnv_psi_irq_set() first.
>>
>> Thanks,
>>
>> C.
>>
>>>> + .instance_size = sizeof(PnvPsi),
>>>> + .instance_init = pnv_psi_init,
>>>> + .class_init = pnv_psi_class_init,
>>>> + .interfaces = (InterfaceInfo[]) {
>>>> + { TYPE_PNV_XSCOM_INTERFACE },
>>>> + { }
>>>> + }
>>>> +};
>>>> +
>>>> +static void pnv_psi_register_types(void)
>>>> +{
>>>> + type_register_static(&pnv_psi_info);
>>>> +}
>>>> +
>>>> +type_init(pnv_psi_register_types)
>>>> diff --git a/include/hw/ppc/pnv.h b/include/hw/ppc/pnv.h
>>>> index f11215ea31f2..f93ec32603b7 100644
>>>> --- a/include/hw/ppc/pnv.h
>>>> +++ b/include/hw/ppc/pnv.h
>>>> @@ -23,6 +23,7 @@
>>>> #include "hw/sysbus.h"
>>>> #include "hw/ppc/pnv_lpc.h"
>>>> #include "hw/ppc/xics.h"
>>>> +#include "hw/ppc/pnv_psi.h"
>>>>
>>>> #define TYPE_PNV_CHIP "powernv-chip"
>>>> #define PNV_CHIP(obj) OBJECT_CHECK(PnvChip, (obj), TYPE_PNV_CHIP)
>>>> @@ -58,6 +59,7 @@ typedef struct PnvChip {
>>>> MemoryRegion icp_mmio;
>>>>
>>>> PnvLpcController lpc;
>>>> + PnvPsi psi;
>>>> } PnvChip;
>>>>
>>>> typedef struct PnvChipClass {
>>>> @@ -119,6 +121,8 @@ typedef struct PnvMachineState {
>>>> ICPState *icps;
>>>> uint32_t nr_servers;
>>>> QLIST_HEAD(, ICSState) ics;
>>>> +
>>>> + uint32_t cpld_irqstate;
>>>> } PnvMachineState;
>>>>
>>>> #define PNV_FDT_ADDR 0x01000000
>>>> @@ -150,4 +154,8 @@ typedef struct PnvMachineState {
>>>> #define PNV_ICP_BASE(chip) 0x0003ffff80000000ull
>>>> #define PNV_ICP_SIZE 0x0000000000100000ull
>>>>
>>>> +#define PNV_PSIHB_BAR 0x0003fffe80000000ull
>>>> +#define PNV_PSIHB_BAR_SIZE 0x0000000000100000ull
>>>> +
>>>> +
>>>> #endif /* _PPC_PNV_H */
>>>> diff --git a/include/hw/ppc/pnv_psi.h b/include/hw/ppc/pnv_psi.h
>>>> new file mode 100644
>>>> index 000000000000..ac3c5f8362e3
>>>> --- /dev/null
>>>> +++ b/include/hw/ppc/pnv_psi.h
>>>> @@ -0,0 +1,61 @@
>>>> +/*
>>>> + * QEMU PowerPC PowerNV PSI controller
>>>> + *
>>>> + * Copyright (c) 2016, IBM Corporation.
>>>> + *
>>>> + * This library is free software; you can redistribute it and/or
>>>> + * modify it under the terms of the GNU Lesser General Public
>>>> + * License as published by the Free Software Foundation; either
>>>> + * version 2 of the License, or (at your option) any later version.
>>>> + *
>>>> + * This library is distributed in the hope that it will be useful,
>>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
>>>> + * Lesser General Public License for more details.
>>>> + *
>>>> + * You should have received a copy of the GNU Lesser General Public
>>>> + * License along with this library; if not, see
>>>> <http://www.gnu.org/licenses/>.
>>>> + */
>>>> +#ifndef _PPC_PNV_PSI_H
>>>> +#define _PPC_PNV_PSI_H
>>>> +
>>>> +#define TYPE_PNV_PSI "pnv-psi"
>>>> +#define PNV_PSI(obj) \
>>>> + OBJECT_CHECK(PnvPsi, (obj), TYPE_PNV_PSI)
>>>> +
>>>> +#define PSIHB_XSCOM_MAX 0x20
>>>> +
>>>> +typedef struct XICSState XICSState;
>>>> +
>>>> +typedef struct PnvPsi {
>>>> + DeviceState parent;
>>>> +
>>>> + MemoryRegion regs_mr;
>>>> +
>>>> + /* FSP region not supported */
>>>> + /* MemoryRegion fsp_mr; */
>>>> +
>>>> + /* Interrupt generation */
>>>> + ICSState ics;
>>>> +
>>>> + /* Registers */
>>>> + uint64_t regs[PSIHB_XSCOM_MAX];
>>>> +
>>>> + MemoryRegion xscom_regs;
>>>> +} PnvPsi;
>>>> +
>>>> +typedef enum PnvPsiIrq {
>>>> + PSIHB_IRQ_PSI, /* internal use only */
>>>> + PSIHB_IRQ_FSP, /* internal use only */
>>>> + PSIHB_IRQ_OCC,
>>>> + PSIHB_IRQ_FSI,
>>>> + PSIHB_IRQ_LPC_I2C,
>>>> + PSIHB_IRQ_LOCAL_ERR,
>>>> + PSIHB_IRQ_EXTERNAL,
>>>> +} PnvPsiIrq;
>>>> +
>>>> +#define PSI_NUM_INTERRUPTS 6
>>>> +
>>>> +extern void pnv_psi_irq_set(PnvPsi *psi, PnvPsiIrq irq, bool state);
>>>> +
>>>> +#endif /* _PPC_PNV_PSI_H */
>>>> diff --git a/include/hw/ppc/pnv_xscom.h b/include/hw/ppc/pnv_xscom.h
>>>> index 0faa1847bf13..2938abd74955 100644
>>>> --- a/include/hw/ppc/pnv_xscom.h
>>>> +++ b/include/hw/ppc/pnv_xscom.h
>>>> @@ -60,6 +60,9 @@ typedef struct PnvXScomInterfaceClass {
>>>> #define PNV_XSCOM_LPC_BASE 0xb0020
>>>> #define PNV_XSCOM_LPC_SIZE 0x4
>>>>
>>>> +#define PNV_XSCOM_PSI_BASE 0x2010900
>>>> +#define PNV_XSCOM_PSI_SIZE 0x20
>>>> +
>>>> extern void pnv_xscom_realize(PnvChip *chip, Error **errp);
>>>> extern int pnv_xscom_populate(PnvChip *chip, void *fdt, int offset);
>>>>
>>>
>>
>
