UbiquitousOS
/
OpenCloudOS-Kernel
mirror of https://gitee.com/OpenCloudOS/OpenCloudOS-Kernel.git
11
0
Fork 0
Code Issues Projects Releases Wiki Activity
OpenCloudOS-Kernel/arch/powerpc/platforms/pseries/msi.c

527 lines
12 KiB
C
Raw Normal View History

[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
/*
* Copyright 2006 Jake Moilanen <moilanen@austin.ibm.com>, IBM Corp.
* Copyright 2006-2007 Michael Ellerman, IBM Corp.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; version 2 of the
* License.
*
*/
#include <linux/device.h>
#include <linux/irq.h>
#include <linux/msi.h>
#include <asm/rtas.h>
#include <asm/hw_irq.h>
#include <asm/ppc-pci.h>
powerpc/pseries: Switch pseries drivers to use machine_xxx_initcall() A lot of the code in platforms/pseries is using non-machine initcalls. That means if a kernel built with pseries support runs on another platform, for example powernv, the initcalls will still run. Most of these cases are OK, though sometimes only due to luck. Some were having more effect: * hcall_inst_init - Checking FW_FEATURE_LPAR which is set on ps3 & celleb. * mobility_sysfs_init - created sysfs files unconditionally - but no effect due to ENOSYS from rtas_ibm_suspend_me() * apo_pm_init - created sysfs, allows write - nothing checks the value written to though * alloc_dispatch_log_kmem_cache - creating kmem_cache on non-pseries machines Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2014-07-16 10:02:43 +08:00
#include <asm/machdep.h>
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
static int query_token, change_token;
#define RTAS_QUERY_FN 0
#define RTAS_CHANGE_FN 1
#define RTAS_RESET_FN 2
#define RTAS_CHANGE_MSI_FN 3
#define RTAS_CHANGE_MSIX_FN 4
powerpc/pseries: Force 32 bit MSIs for devices that require it The following patch implements a new PAPR change which allows the OS to force the use of 32 bit MSIs, regardless of what the PCI capabilities indicate. This is required for some devices that advertise support for 64 bit MSIs but don't actually support them. Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-03 19:30:59 +08:00
#define RTAS_CHANGE_32MSI_FN 5
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
/* RTAS Helpers */
static int rtas_change_msi(struct pci_dn *pdn, u32 func, u32 num_irqs)
{
u32 addr, seq_num, rtas_ret[3];
unsigned long buid;
int rc;
addr = rtas_config_addr(pdn->busno, pdn->devfn, 0);
buid = pdn->phb->buid;
seq_num = 1;
do {
powerpc/pseries: Force 32 bit MSIs for devices that require it The following patch implements a new PAPR change which allows the OS to force the use of 32 bit MSIs, regardless of what the PCI capabilities indicate. This is required for some devices that advertise support for 64 bit MSIs but don't actually support them. Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-03 19:30:59 +08:00
if (func == RTAS_CHANGE_MSI_FN || func == RTAS_CHANGE_MSIX_FN ||
func == RTAS_CHANGE_32MSI_FN)
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
rc = rtas_call(change_token, 6, 4, rtas_ret, addr,
BUID_HI(buid), BUID_LO(buid),
func, num_irqs, seq_num);
else
rc = rtas_call(change_token, 6, 3, rtas_ret, addr,
BUID_HI(buid), BUID_LO(buid),
func, num_irqs, seq_num);
seq_num = rtas_ret[1];
} while (rtas_busy_delay(rc));
[POWERPC] Simplify rtas_change_msi() error semantics Currently rtas_change_msi() returns either the error code from RTAS, or if the RTAS call succeeded the number of irqs that were configured by RTAS. This makes checking the return value more complicated than it needs to be. Instead, have rtas_change_msi() check that the number of irqs configured by RTAS is equal to what we requested - and return an error otherwise. This makes the return semantics match the usual 0 for success, something else for error. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:50 +08:00
/*
powerpc/pseries: Return the number of MSIs we could allocate If we can't allocate the requested number of MSIs, we can still tell the generic code how many we were able to allocate. That can then be passed onto the driver, allowing it to request that many in future, and probably succeeed. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:33 +08:00
* If the RTAS call succeeded, return the number of irqs allocated.
* If not, make sure we return a negative error code.
[POWERPC] Simplify rtas_change_msi() error semantics Currently rtas_change_msi() returns either the error code from RTAS, or if the RTAS call succeeded the number of irqs that were configured by RTAS. This makes checking the return value more complicated than it needs to be. Instead, have rtas_change_msi() check that the number of irqs configured by RTAS is equal to what we requested - and return an error otherwise. This makes the return semantics match the usual 0 for success, something else for error. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:50 +08:00
*/
powerpc/pseries: Return the number of MSIs we could allocate If we can't allocate the requested number of MSIs, we can still tell the generic code how many we were able to allocate. That can then be passed onto the driver, allowing it to request that many in future, and probably succeeed. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:33 +08:00
if (rc == 0)
rc = rtas_ret[0];
else if (rc > 0)
rc = -rc;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
[POWERPC] Simplify rtas_change_msi() error semantics Currently rtas_change_msi() returns either the error code from RTAS, or if the RTAS call succeeded the number of irqs that were configured by RTAS. This makes checking the return value more complicated than it needs to be. Instead, have rtas_change_msi() check that the number of irqs configured by RTAS is equal to what we requested - and return an error otherwise. This makes the return semantics match the usual 0 for success, something else for error. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:50 +08:00
pr_debug("rtas_msi: ibm,change_msi(func=%d,num=%d), got %d rc = %d\n",
func, num_irqs, rtas_ret[0], rc);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
return rc;
}
static void rtas_disable_msi(struct pci_dev *pdev)
{
struct pci_dn *pdn;
powerpc: Make radeon 32-bit MSI quirk work on powernv This moves the quirk itself to pci_64.c as to get built on all ppc64 platforms (the only ones with a pci_dn), factors the two implementations of get_pdn() into a single pci_get_dn() and use the quirk to do 32-bit MSIs on IODA based powernv platforms. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 06:58:21 +08:00
pdn = pci_get_pdn(pdev);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
if (!pdn)
return;
powerpc/pseries: Disable MSI using new interface if possible On upcoming hardware, we have a PCI adapter with two functions, one of which uses MSI and the other uses MSI-X. This adapter, when MSI is disabled using the "old" firmware interface (RTAS_CHANGE_FN), still signals an MSI-X interrupt and triggers an EEH. We are working with the vendor to ensure that the hardware is not at fault, but if we use the "new" interface (RTAS_CHANGE_MSI_FN) to disable MSI, we also automatically disable MSI-X and the adapter does not appear to signal any stray MSI-X interrupt. Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Acked-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2011-03-03 23:41:02 +08:00
/*
* disabling MSI with the explicit interface also disables MSI-X
*/
if (rtas_change_msi(pdn, RTAS_CHANGE_MSI_FN, 0) != 0) {
/*
* may have failed because explicit interface is not
* present
*/
if (rtas_change_msi(pdn, RTAS_CHANGE_FN, 0) != 0) {
pr_debug("rtas_msi: Setting MSIs to 0 failed!\n");
}
}
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
static int rtas_query_irq_number(struct pci_dn *pdn, int offset)
{
u32 addr, rtas_ret[2];
unsigned long buid;
int rc;
addr = rtas_config_addr(pdn->busno, pdn->devfn, 0);
buid = pdn->phb->buid;
do {
rc = rtas_call(query_token, 4, 3, rtas_ret, addr,
BUID_HI(buid), BUID_LO(buid), offset);
} while (rtas_busy_delay(rc));
if (rc) {
pr_debug("rtas_msi: error (%d) querying source number\n", rc);
return rc;
}
return rtas_ret[0];
}
static void rtas_teardown_msi_irqs(struct pci_dev *pdev)
{
struct msi_desc *entry;
list_for_each_entry(entry, &pdev->msi_list, list) {
if (entry->irq == NO_IRQ)
continue;
powerpc: Convert to new irq_* function names Scripted with coccinelle. Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-03-25 23:45:20 +08:00
irq_set_msi_desc(entry->irq, NULL);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
irq_dispose_mapping(entry->irq);
}
rtas_disable_msi(pdev);
}
powerpc/pseries: Add support for ibm,req#msi-x Firmware encodes the number of MSI-X requested by a device in a different property than for MSI. Pull the property name out as a parameter and share the logic for both cases. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:31 +08:00
static int check_req(struct pci_dev *pdev, int nvec, char *prop_name)
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
{
struct device_node *dn;
struct pci_dn *pdn;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
const __be32 *p;
u32 req_msi;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
powerpc: Make radeon 32-bit MSI quirk work on powernv This moves the quirk itself to pci_64.c as to get built on all ppc64 platforms (the only ones with a pci_dn), factors the two implementations of get_pdn() into a single pci_get_dn() and use the quirk to do 32-bit MSIs on IODA based powernv platforms. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 06:58:21 +08:00
pdn = pci_get_pdn(pdev);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
if (!pdn)
return -ENODEV;
dn = pdn->node;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
p = of_get_property(dn, prop_name, NULL);
if (!p) {
powerpc/pseries: Add support for ibm,req#msi-x Firmware encodes the number of MSI-X requested by a device in a different property than for MSI. Pull the property name out as a parameter and share the logic for both cases. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:31 +08:00
pr_debug("rtas_msi: No %s on %s\n", prop_name, dn->full_name);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
return -ENOENT;
}
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
req_msi = be32_to_cpup(p);
if (req_msi < nvec) {
powerpc/pseries: Add support for ibm,req#msi-x Firmware encodes the number of MSI-X requested by a device in a different property than for MSI. Pull the property name out as a parameter and share the logic for both cases. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:31 +08:00
pr_debug("rtas_msi: %s requests < %d MSIs\n", prop_name, nvec);
powerpc/pseries: Return req#msi(-x) if request is larger If a driver asks for more MSIs than the devices "req#msi(-x)" property, we currently return -ENOSPC. This doesn't give the driver any chance to make a new request with a number that might work. So if "req#msi(-x)" is less than the request, return its value. To be 100% safe, make sure we return an error if req_msi == 0. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:18:49 +08:00
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
if (req_msi == 0) /* Be paranoid */
powerpc/pseries: Return req#msi(-x) if request is larger If a driver asks for more MSIs than the devices "req#msi(-x)" property, we currently return -ENOSPC. This doesn't give the driver any chance to make a new request with a number that might work. So if "req#msi(-x)" is less than the request, return its value. To be 100% safe, make sure we return an error if req_msi == 0. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:18:49 +08:00
return -ENOSPC;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
return req_msi;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
return 0;
}
powerpc/pseries: Add support for ibm,req#msi-x Firmware encodes the number of MSI-X requested by a device in a different property than for MSI. Pull the property name out as a parameter and share the logic for both cases. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:31 +08:00
static int check_req_msi(struct pci_dev *pdev, int nvec)
{
return check_req(pdev, nvec, "ibm,req#msi");
}
static int check_req_msix(struct pci_dev *pdev, int nvec)
{
return check_req(pdev, nvec, "ibm,req#msi-x");
}
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
/* Quota calculation */
static struct device_node *find_pe_total_msi(struct pci_dev *dev, int *total)
{
struct device_node *dn;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
const __be32 *p;
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
dn = of_node_get(pci_device_to_OF_node(dev));
while (dn) {
p = of_get_property(dn, "ibm,pe-total-#msi", NULL);
if (p) {
pr_debug("rtas_msi: found prop on dn %s\n",
dn->full_name);
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
*total = be32_to_cpup(p);
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
return dn;
}
dn = of_get_next_parent(dn);
}
return NULL;
}
static struct device_node *find_pe_dn(struct pci_dev *dev, int *total)
{
struct device_node *dn;
powerpc/eeh: Trace error based on PE from beginning There're 2 conditions to trigger EEH error detection: invalid value returned from reading I/O or config space. On each case, the function eeh_dn_check_failure will be called to initialize EEH event and put it into the poll for further processing. The patch changes the function for a little bit so that the EEH error will be traced based on PE instead of EEH device any more. Also, the function eeh_find_device_pe() has been removed since the eeh device is tracing the PE by struct eeh_dev::pe. Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-09-08 06:44:13 +08:00
struct eeh_dev *edev;
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
/* Found our PE and assume 8 at that point. */
dn = pci_device_to_OF_node(dev);
if (!dn)
return NULL;
powerpc/eeh: Trace error based on PE from beginning There're 2 conditions to trigger EEH error detection: invalid value returned from reading I/O or config space. On each case, the function eeh_dn_check_failure will be called to initialize EEH event and put it into the poll for further processing. The patch changes the function for a little bit so that the EEH error will be traced based on PE instead of EEH device any more. Also, the function eeh_find_device_pe() has been removed since the eeh device is tracing the PE by struct eeh_dev::pe. Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-09-08 06:44:13 +08:00
/* Get the top level device in the PE */
edev = of_node_to_eeh_dev(dn);
powerpc/pseries: Fix oops with MSIs when missing EEH PEs The new EEH code introduced a small regression, if the EEH PEs are missin (which happens currently in qemu for example), it will deref a NULL pointer in the MSI code. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-11-23 10:25:39 +08:00
if (edev->pe)
edev = list_first_entry(&edev->pe->edevs, struct eeh_dev, list);
powerpc/eeh: Trace error based on PE from beginning There're 2 conditions to trigger EEH error detection: invalid value returned from reading I/O or config space. On each case, the function eeh_dn_check_failure will be called to initialize EEH event and put it into the poll for further processing. The patch changes the function for a little bit so that the EEH error will be traced based on PE instead of EEH device any more. Also, the function eeh_find_device_pe() has been removed since the eeh device is tracing the PE by struct eeh_dev::pe. Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-09-08 06:44:13 +08:00
dn = eeh_dev_to_of_node(edev);
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
if (!dn)
return NULL;
/* We actually want the parent */
dn = of_get_parent(dn);
if (!dn)
return NULL;
/* Hardcode of 8 for old firmwares */
*total = 8;
pr_debug("rtas_msi: using PE dn %s\n", dn->full_name);
return dn;
}
struct msi_counts {
struct device_node *requestor;
int num_devices;
int request;
int quota;
int spare;
int over_quota;
};
static void *count_non_bridge_devices(struct device_node *dn, void *data)
{
struct msi_counts *counts = data;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
const __be32 *p;
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
u32 class;
pr_debug("rtas_msi: counting %s\n", dn->full_name);
p = of_get_property(dn, "class-code", NULL);
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
class = p ? be32_to_cpup(p) : 0;
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
if ((class >> 8) != PCI_CLASS_BRIDGE_PCI)
counts->num_devices++;
return NULL;
}
static void *count_spare_msis(struct device_node *dn, void *data)
{
struct msi_counts *counts = data;
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
const __be32 *p;
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
int req;
if (dn == counts->requestor)
req = counts->request;
else {
/* We don't know if a driver will try to use MSI or MSI-X,
* so we just have to punt and use the larger of the two. */
req = 0;
p = of_get_property(dn, "ibm,req#msi", NULL);
if (p)
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
req = be32_to_cpup(p);
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
p = of_get_property(dn, "ibm,req#msi-x", NULL);
if (p)
powerpc/pseries: Fix endian issues in MSI code The MSI code is miscalculating quotas in little endian mode. Add required byteswaps to fix this. Before we claimed a quota of 65536, after the patch we see the correct value of 256. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-12-12 12:59:40 +08:00
req = max(req, (int)be32_to_cpup(p));
powerpc/pseries: Implement a quota system for MSIs There are hardware limitations on the number of available MSIs, which firmware expresses using a property named "ibm,pe-total-#msi". This property tells us how many MSIs are available for devices below the point in the PCI tree where we find the property. For old firmwares which don't have the property, we assume there are 8 MSIs available per "partitionable endpoint" (PE). The PE can be found using existing EEH code, which uses the methods described in PAPR. For our purposes we want the parent of the node that's identified using this method. When a driver requests n MSIs for a device, we first establish where the "ibm,pe-total-#msi" property above that device is, or we find the PE if the property is not found. In both cases we call this node the "pe_dn". We then count all non-bridge devices below the pe_dn, to establish how many devices in total may need MSIs. The quota is then simply the total available divided by the number of devices, if the request is less than or equal to the quota, the request is fine and we're done. If the request is greater than the quota, we try to determine if there are any "spare" MSIs which we can give to this device. Spare MSIs are found by looking for other devices which can never use their full quota, because their "req#msi(-x)" property is less than the quota. If we find any spare, we divide the spares by the number of devices that could request more than their quota. This ensures the spare MSIs are spread evenly amongst all over-quota requestors. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-02-17 08:21:56 +08:00
}
if (req < counts->quota)
counts->spare += counts->quota - req;
else if (req > counts->quota)
counts->over_quota++;
return NULL;
}
static int msi_quota_for_device(struct pci_dev *dev, int request)
{
struct device_node *pe_dn;
struct msi_counts counts;
int total;
pr_debug("rtas_msi: calc quota for %s, request %d\n", pci_name(dev),
request);
pe_dn = find_pe_total_msi(dev, &total);
if (!pe_dn)
pe_dn = find_pe_dn(dev, &total);
if (!pe_dn) {
pr_err("rtas_msi: couldn't find PE for %s\n", pci_name(dev));
goto out;
}
pr_debug("rtas_msi: found PE %s\n", pe_dn->full_name);
memset(&counts, 0, sizeof(struct msi_counts));
/* Work out how many devices we have below this PE */
traverse_pci_devices(pe_dn, count_non_bridge_devices, &counts);
if (counts.num_devices == 0) {
pr_err("rtas_msi: found 0 devices under PE for %s\n",
pci_name(dev));
goto out;
}
counts.quota = total / counts.num_devices;
if (request <= counts.quota)
goto out;
/* else, we have some more calculating to do */
counts.requestor = pci_device_to_OF_node(dev);
counts.request = request;
traverse_pci_devices(pe_dn, count_spare_msis, &counts);
/* If the quota isn't an integer multiple of the total, we can
* use the remainder as spare MSIs for anyone that wants them. */
counts.spare += total % counts.num_devices;
/* Divide any spare by the number of over-quota requestors */
if (counts.over_quota)
counts.quota += counts.spare / counts.over_quota;
/* And finally clamp the request to the possibly adjusted quota */
request = min(counts.quota, request);
pr_debug("rtas_msi: request clamped to quota %d\n", request);
out:
of_node_put(pe_dn);
return request;
}
powerpc/pseries: Reject discontiguous/non-zero based MSI-X requests There's no way for us to express to firmware that we want a discontiguous, or non-zero based, range of MSI-X entries. So we must reject such requests. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-03-05 22:44:26 +08:00
static int check_msix_entries(struct pci_dev *pdev)
{
struct msi_desc *entry;
int expected;
/* There's no way for us to express to firmware that we want
* a discontiguous, or non-zero based, range of MSI-X entries.
* So we must reject such requests. */
expected = 0;
list_for_each_entry(entry, &pdev->msi_list, list) {
if (entry->msi_attrib.entry_nr != expected) {
pr_debug("rtas_msi: bad MSI-X entries.\n");
return -EINVAL;
}
expected++;
}
return 0;
}
powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support Recent commit e61133dda480062d221f09e4fc18f66763f8ecd0 added support for a new firmware feature to force an adapter to use 32 bit MSIs. However, this firmware is not available for all systems. The hack below allows devices needing 32 bit MSIs to work on these systems as well. It is careful to only enable this on Gen2 slots, which should limit this to configurations where this hack is needed and tested to work. [Small change to factor out the hack into a separate function -- BenH] Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 19:07:46 +08:00
static void rtas_hack_32bit_msi_gen2(struct pci_dev *pdev)
{
u32 addr_hi, addr_lo;
/*
* We should only get in here for IODA1 configs. This is based on the
* fact that we using RTAS for MSIs, we don't have the 32 bit MSI RTAS
* support, and we are in a PCIe Gen2 slot.
*/
dev_info(&pdev->dev,
"rtas_msi: No 32 bit MSI firmware support, forcing 32 bit MSI\n");
pci_read_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_HI, &addr_hi);
addr_lo = 0xffff0000 | ((addr_hi >> (48 - 32)) << 4);
pci_write_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_LO, addr_lo);
pci_write_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_HI, 0);
}
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
static int rtas_setup_msi_irqs(struct pci_dev *pdev, int nvec_in, int type)
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
{
struct pci_dn *pdn;
PCI/MSI/PPC: Remove arch_msi_check_device() Move MSI checks from arch_msi_check_device() to arch_setup_msi_irqs(). This makes the code more compact and allows removing arch_msi_check_device() from generic MSI code. Signed-off-by: Alexander Gordeev <agordeev@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Acked-by: Michael Ellerman <mpe@ellerman.id.au>
2014-09-08 02:57:53 +08:00
int hwirq, virq, i, quota, rc;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
struct msi_desc *entry;
[POWERPC] Read back MSI message in rtas_setup_msi_irqs() so restore works There are plans afoot to use pci_restore_msi_state() to restore MSI state after a device reset. In order for this to work for the RTAS MSI backend, we need to read back the MSI message from config space after it has been setup by firmware. This should be sufficient for restoring the MSI state after a device reset, however we will need to revisit this for suspend to disk if that is ever implemented on pseries. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-10-23 12:23:44 +08:00
struct msi_msg msg;
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
int nvec = nvec_in;
powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support Recent commit e61133dda480062d221f09e4fc18f66763f8ecd0 added support for a new firmware feature to force an adapter to use 32 bit MSIs. However, this firmware is not available for all systems. The hack below allows devices needing 32 bit MSIs to work on these systems as well. It is careful to only enable this on Gen2 slots, which should limit this to configurations where this hack is needed and tested to work. [Small change to factor out the hack into a separate function -- BenH] Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 19:07:46 +08:00
int use_32bit_msi_hack = 0;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
PCI/MSI/PPC: Remove arch_msi_check_device() Move MSI checks from arch_msi_check_device() to arch_setup_msi_irqs(). This makes the code more compact and allows removing arch_msi_check_device() from generic MSI code. Signed-off-by: Alexander Gordeev <agordeev@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Acked-by: Michael Ellerman <mpe@ellerman.id.au>
2014-09-08 02:57:53 +08:00
if (type == PCI_CAP_ID_MSIX)
rc = check_req_msix(pdev, nvec);
else
rc = check_req_msi(pdev, nvec);
if (rc)
return rc;
quota = msi_quota_for_device(pdev, nvec);
if (quota && quota < nvec)
return quota;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
powerpc/pseries: Reject discontiguous/non-zero based MSI-X requests There's no way for us to express to firmware that we want a discontiguous, or non-zero based, range of MSI-X entries. So we must reject such requests. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-03-05 22:44:26 +08:00
if (type == PCI_CAP_ID_MSIX && check_msix_entries(pdev))
return -EINVAL;
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
/*
* Firmware currently refuse any non power of two allocation
* so we round up if the quota will allow it.
*/
if (type == PCI_CAP_ID_MSIX) {
int m = roundup_pow_of_two(nvec);
PCI/MSI/PPC: Remove arch_msi_check_device() Move MSI checks from arch_msi_check_device() to arch_setup_msi_irqs(). This makes the code more compact and allows removing arch_msi_check_device() from generic MSI code. Signed-off-by: Alexander Gordeev <agordeev@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Acked-by: Michael Ellerman <mpe@ellerman.id.au>
2014-09-08 02:57:53 +08:00
quota = msi_quota_for_device(pdev, m);
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
if (quota >= m)
nvec = m;
}
PCI/MSI/PPC: Remove arch_msi_check_device() Move MSI checks from arch_msi_check_device() to arch_setup_msi_irqs(). This makes the code more compact and allows removing arch_msi_check_device() from generic MSI code. Signed-off-by: Alexander Gordeev <agordeev@redhat.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Acked-by: Michael Ellerman <mpe@ellerman.id.au>
2014-09-08 02:57:53 +08:00
pdn = pci_get_pdn(pdev);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
/*
* Try the new more explicit firmware interface, if that fails fall
* back to the old interface. The old interface is known to never
* return MSI-Xs.
*/
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
again:
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
if (type == PCI_CAP_ID_MSI) {
powerpc/pseries: Honor the generic "no_64bit_msi" flag Instead of the arch specific quirk which we are deprecating Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> CC: <stable@vger.kernel.org>
2014-10-07 13:12:55 +08:00
if (pdev->no_64bit_msi) {
powerpc/pseries: Force 32 bit MSIs for devices that require it The following patch implements a new PAPR change which allows the OS to force the use of 32 bit MSIs, regardless of what the PCI capabilities indicate. This is required for some devices that advertise support for 64 bit MSIs but don't actually support them. Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-03 19:30:59 +08:00
rc = rtas_change_msi(pdn, RTAS_CHANGE_32MSI_FN, nvec);
powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support Recent commit e61133dda480062d221f09e4fc18f66763f8ecd0 added support for a new firmware feature to force an adapter to use 32 bit MSIs. However, this firmware is not available for all systems. The hack below allows devices needing 32 bit MSIs to work on these systems as well. It is careful to only enable this on Gen2 slots, which should limit this to configurations where this hack is needed and tested to work. [Small change to factor out the hack into a separate function -- BenH] Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 19:07:46 +08:00
if (rc < 0) {
/*
* We only want to run the 32 bit MSI hack below if
* the max bus speed is Gen2 speed
*/
if (pdev->bus->max_bus_speed != PCIE_SPEED_5_0GT)
return rc;
use_32bit_msi_hack = 1;
}
} else
rc = -1;
if (rc < 0)
powerpc/pseries: Force 32 bit MSIs for devices that require it The following patch implements a new PAPR change which allows the OS to force the use of 32 bit MSIs, regardless of what the PCI capabilities indicate. This is required for some devices that advertise support for 64 bit MSIs but don't actually support them. Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-03 19:30:59 +08:00
rc = rtas_change_msi(pdn, RTAS_CHANGE_MSI_FN, nvec);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support Recent commit e61133dda480062d221f09e4fc18f66763f8ecd0 added support for a new firmware feature to force an adapter to use 32 bit MSIs. However, this firmware is not available for all systems. The hack below allows devices needing 32 bit MSIs to work on these systems as well. It is careful to only enable this on Gen2 slots, which should limit this to configurations where this hack is needed and tested to work. [Small change to factor out the hack into a separate function -- BenH] Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 19:07:46 +08:00
if (rc < 0) {
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
pr_debug("rtas_msi: trying the old firmware call.\n");
rc = rtas_change_msi(pdn, RTAS_CHANGE_FN, nvec);
}
powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support Recent commit e61133dda480062d221f09e4fc18f66763f8ecd0 added support for a new firmware feature to force an adapter to use 32 bit MSIs. However, this firmware is not available for all systems. The hack below allows devices needing 32 bit MSIs to work on these systems as well. It is careful to only enable this on Gen2 slots, which should limit this to configurations where this hack is needed and tested to work. [Small change to factor out the hack into a separate function -- BenH] Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2013-05-22 19:07:46 +08:00
if (use_32bit_msi_hack && rc > 0)
rtas_hack_32bit_msi_gen2(pdev);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
} else
rc = rtas_change_msi(pdn, RTAS_CHANGE_MSIX_FN, nvec);
powerpc/pseries: Return the number of MSIs we could allocate If we can't allocate the requested number of MSIs, we can still tell the generic code how many we were able to allocate. That can then be passed onto the driver, allowing it to request that many in future, and probably succeeed. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:33 +08:00
if (rc != nvec) {
powerpc/pseries: Round up MSI-X requests The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2012-06-05 00:47:03 +08:00
if (nvec != nvec_in) {
nvec = nvec_in;
goto again;
}
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
pr_debug("rtas_msi: rtas_change_msi() failed\n");
[POWERPC] Simplify error logic in rtas_setup_msi_irqs() rtas_setup_msi_irqs() doesn't need to call teardown() itself, the generic code will do this for us as long as we return a non-zero value. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:48 +08:00
return rc;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
i = 0;
list_for_each_entry(entry, &pdev->msi_list, list) {
powerpc/pseries: Fix MSI-X interrupt querying We need to increment i in the loop that queries what interrupts firmware gave us, otherwise we'll incorrectly use the first value over and over. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:31 +08:00
hwirq = rtas_query_irq_number(pdn, i++);
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
if (hwirq < 0) {
pr_debug("rtas_msi: error (%d) getting hwirq\n", rc);
[POWERPC] Simplify error logic in rtas_setup_msi_irqs() rtas_setup_msi_irqs() doesn't need to call teardown() itself, the generic code will do this for us as long as we return a non-zero value. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:48 +08:00
return hwirq;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
virq = irq_create_mapping(NULL, hwirq);
if (virq == NO_IRQ) {
pr_debug("rtas_msi: Failed mapping hwirq %d\n", hwirq);
[POWERPC] Simplify error logic in rtas_setup_msi_irqs() rtas_setup_msi_irqs() doesn't need to call teardown() itself, the generic code will do this for us as long as we return a non-zero value. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-09-20 14:36:48 +08:00
return -ENOSPC;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
dev_dbg(&pdev->dev, "rtas_msi: allocated virq %d\n", virq);
powerpc: Convert to new irq_* function names Scripted with coccinelle. Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-03-25 23:45:20 +08:00
irq_set_msi_desc(virq, entry);
[POWERPC] Read back MSI message in rtas_setup_msi_irqs() so restore works There are plans afoot to use pci_restore_msi_state() to restore MSI state after a device reset. In order for this to work for the RTAS MSI backend, we need to read back the MSI message from config space after it has been setup by firmware. This should be sufficient for restoring the MSI state after a device reset, however we will need to revisit this for suspend to disk if that is ever implemented on pseries. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-10-23 12:23:44 +08:00
/* Read config space back so we can restore after reset */
PCI/MSI: Rename __read_msi_msg() to __pci_read_msi_msg() Rename __read_msi_msg() to __pci_read_msi_msg() and kill unused read_msi_msg(). It's a preparation to separate generic MSI code from PCI core. Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com> Cc: Bjorn Helgaas <bhelgaas@google.com> Cc: Grant Likely <grant.likely@linaro.org> Cc: Marc Zyngier <marc.zyngier@arm.com> Cc: Yingjoe Chen <yingjoe.chen@mediatek.com> Cc: Yijing Wang <wangyijing@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-11-09 23:10:33 +08:00
__pci_read_msi_msg(entry, &msg);
[POWERPC] Read back MSI message in rtas_setup_msi_irqs() so restore works There are plans afoot to use pci_restore_msi_state() to restore MSI state after a device reset. In order for this to work for the RTAS MSI backend, we need to read back the MSI message from config space after it has been setup by firmware. This should be sufficient for restoring the MSI state after a device reset, however we will need to revisit this for suspend to disk if that is ever implemented on pseries. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-10-23 12:23:44 +08:00
entry->msg = msg;
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
}
return 0;
}
static void rtas_msi_pci_irq_fixup(struct pci_dev *pdev)
{
/* No LSI -> leave MSIs (if any) configured */
if (pdev->irq == NO_IRQ) {
dev_dbg(&pdev->dev, "rtas_msi: no LSI, nothing to do.\n");
return;
}
/* No MSI -> MSIs can't have been assigned by fw, leave LSI */
powerpc/pseries: Check for MSI-X also in rtas_msi_pci_irq_fixup() We also need to check that the device isn't using MSI-X in the irq fixup routine, otherwise we might leave MSI-Xs configured at boot. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2009-01-23 04:54:32 +08:00
if (check_req_msi(pdev, 1) && check_req_msix(pdev, 1)) {
dev_dbg(&pdev->dev, "rtas_msi: no req#msi/x, nothing to do.\n");
[POWERPC] RTAS MSI implementation Implement MSI support via RTAS (RTAS = run-time firmware on pSeries machines). For now we assumes that if the required RTAS tokens for MSI are present, then we want to use the RTAS MSI routines. When RTAS is managing MSIs for us, it will/may enable MSI on devices that support it by default. This is contrary to the Linux model where a device is in LSI mode until the driver requests MSIs. To remedy this we add a pci_irq_fixup call, which disables MSI if they've been assigned by firmware and the device also supports LSI. Devices that don't support LSI at all will be left as is, drivers are still expected to call pci_enable_msi() before using the device. At the moment there is no pci_irq_fixup on pSeries, so we can just set it unconditionally. If other platforms use the RTAS MSI backend they'll need to check that still holds. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 10:58:35 +08:00
return;
}
dev_dbg(&pdev->dev, "rtas_msi: disabling existing MSI.\n");
rtas_disable_msi(pdev);
}
static int rtas_msi_init(void)
{
query_token = rtas_token("ibm,query-interrupt-source-number");
change_token = rtas_token("ibm,change-msi");
if ((query_token == RTAS_UNKNOWN_SERVICE) ||
(change_token == RTAS_UNKNOWN_SERVICE)) {
pr_debug("rtas_msi: no RTAS tokens, no MSI support.\n");
return -1;
}
pr_debug("rtas_msi: Registering RTAS MSI callbacks.\n");
WARN_ON(ppc_md.setup_msi_irqs);
ppc_md.setup_msi_irqs = rtas_setup_msi_irqs;
ppc_md.teardown_msi_irqs = rtas_teardown_msi_irqs;
WARN_ON(ppc_md.pci_irq_fixup);
ppc_md.pci_irq_fixup = rtas_msi_pci_irq_fixup;
return 0;
}
powerpc/pseries: Switch pseries drivers to use machine_xxx_initcall() A lot of the code in platforms/pseries is using non-machine initcalls. That means if a kernel built with pseries support runs on another platform, for example powernv, the initcalls will still run. Most of these cases are OK, though sometimes only due to luck. Some were having more effect: * hcall_inst_init - Checking FW_FEATURE_LPAR which is set on ps3 & celleb. * mobility_sysfs_init - created sysfs files unconditionally - but no effect due to ENOSYS from rtas_ibm_suspend_me() * apo_pm_init - created sysfs, allows write - nothing checks the value written to though * alloc_dispatch_log_kmem_cache - creating kmem_cache on non-pseries machines Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2014-07-16 10:02:43 +08:00
machine_arch_initcall(pseries, rtas_msi_init);