| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20140821141729.GT1660@lahna.fi.intel.com>
Date: Thu, 21 Aug 2014 17:17:58 +0300
From: Mika Westerberg <mika.westerberg@...ux.intel.com>
To: Jiang Liu <jiang.liu@...ux.intel.com>
Cc: Benjamin Herrenschmidt <benh@...nel.crashing.org>,
Thomas Gleixner <tglx@...utronix.de>,
Grant Likely <grant.likely@...aro.org>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Randy Dunlap <rdunlap@...radead.org>,
Yinghai Lu <yinghai@...nel.org>, x86@...nel.org,
Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Tony Luck <tony.luck@...el.com>,
Joerg Roedel <joro@...tes.org>,
Paul Gortmaker <paul.gortmaker@...driver.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org,
linux-acpi@...r.kernel.org, Ingo Molnar <mingo@...nel.org>
Subject: Re: [Patch V4 29/42] x86, irq: introduce two helper functions to
support irqdomain map operation
On Mon, Jun 09, 2014 at 04:19:58PM +0800, Jiang Liu wrote:
> Currently there are multiple entries to program IOAPIC pins, such as
> io_apic_setup_irq_pin_once(), io_apic_set_pci_routing() and
> setup_IO_APIC_irq_extra() etc.
>
> This patch introduces two functions to help consolidate the code to
> program IOAPIC pins. Function mp_set_pin_attr() is used to optionally
> set trigger, polarity and NUMA node property for an IOAPIC pin.
> If mp_set_pin_attr() is not invoked for a pin, the default configuration
> from BIOS will be used.
>
> Function mp_irqdomain_map() is an common implementation of irqdomain map()
> operation. It figures out attribures for pin and then actually programs
> the IOAPIC pin. We hope this will be the only entrance for programming
> IOAPIC pin.
>
> And the flow will:
> 1) caller such as xxx_pci_irq_enable figures out pin attributes.
> 2) Invoke mp_set_pin_attr() to set attributes for a pin. If the pin has
> already bin programmed, mp_set_pin_attr() will aslo detects attribute
> confictions.
> 3) Invoke mp_map_pin_to_irq()
> 3.1) If IRQ has already been assigned, return irq_find_mapping()
> 3.2) Else irq_create_mapping()
> ->irq_domain_associate()
> ->mp_irqdomain_map()
> ->io_apic_setup_irq_pin()
>
> So every pin will only programmed once by mp_irqdomain_map(), so we
> could kill io_apic_setup_irq_pin_once(), io_apic_set_pci_routing() and
> setup_IO_APIC_irq_extra() etc.
>
> Signed-off-by: Jiang Liu <jiang.liu@...ux.intel.com>
> ---
> arch/x86/include/asm/io_apic.h | 5 ++
> arch/x86/kernel/apic/io_apic.c | 99 +++++++++++++++++++++++++++++++++++++++-
> 2 files changed, 103 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h
> index 3e4bea3a52b1..c53587868590 100644
> --- a/arch/x86/include/asm/io_apic.h
> +++ b/arch/x86/include/asm/io_apic.h
> @@ -173,7 +173,9 @@ enum ioapic_domain_type {
> };
>
> struct device_node;
> +struct irq_domain;
> struct irq_domain_ops;
> +
> struct ioapic_domain_cfg {
> enum ioapic_domain_type type;
> const struct irq_domain_ops *ops;
> @@ -192,6 +194,9 @@ extern u32 mp_pin_to_gsi(int ioapic, int pin);
> extern int mp_map_gsi_to_irq(u32 gsi, unsigned int flags);
> extern void __init mp_register_ioapic(int id, u32 address, u32 gsi_base,
> struct ioapic_domain_cfg *cfg);
> +extern int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
> + irq_hw_number_t hwirq);
> +extern int mp_set_gsi_attr(u32 gsi, int trigger, int polarity, int node);
> extern void __init pre_init_apic_IRQ0(void);
>
> extern void mp_save_irq(struct mpc_intsrc *m);
> diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
> index 9c5f70699a80..61aae90f7e23 100644
> --- a/arch/x86/kernel/apic/io_apic.c
> +++ b/arch/x86/kernel/apic/io_apic.c
> @@ -87,6 +87,14 @@ static DEFINE_RAW_SPINLOCK(vector_lock);
> static DEFINE_MUTEX(ioapic_mutex);
> static unsigned int ioapic_dynirq_base;
>
> +struct mp_pin_info {
> + int trigger;
> + int polarity;
> + int node;
> + int set;
> + u32 count;
> +};
> +
> static struct ioapic {
> /*
> * # of IRQ routing registers
> @@ -102,6 +110,7 @@ static struct ioapic {
> struct mp_ioapic_gsi gsi_config;
> struct ioapic_domain_cfg irqdomain_cfg;
> struct irq_domain *irqdomain;
> + struct mp_pin_info *pin_info;
> DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1);
> } ioapics[MAX_IO_APICS];
>
> @@ -147,6 +156,11 @@ static inline int mp_init_irq_at_boot(int ioapic, int irq)
> return ioapic == 0 || (irq >= 0 && irq < nr_legacy_irqs());
> }
>
> +static inline struct mp_pin_info *mp_pin_info(int ioapic_idx, int pin)
> +{
> + return ioapics[ioapic_idx].pin_info + pin;
> +}
> +
> static inline struct irq_domain *mp_ioapic_irqdomain(int ioapic)
> {
> return ioapics[ioapic].irqdomain;
> @@ -1006,6 +1020,7 @@ static int mp_map_pin_to_irq(u32 gsi, int idx, int ioapic, int pin,
> {
> int irq;
> struct irq_domain *domain = mp_ioapic_irqdomain(ioapic);
> + struct mp_pin_info *info = mp_pin_info(ioapic, pin);
>
> /*
> * Don't use irqdomain to manage ISA IRQs because there may be
> @@ -1034,6 +1049,13 @@ static int mp_map_pin_to_irq(u32 gsi, int idx, int ioapic, int pin,
> irq = irq_find_mapping(domain, pin);
> if (irq <= 0 && (flags & IOAPIC_MAP_ALLOC))
> irq = alloc_irq_from_domain(domain, gsi, pin);
> +
> + if (flags & IOAPIC_MAP_ALLOC) {
> + if (irq > 0)
> + info->count++;
> + else if (info->count == 0)
> + info->set = 0;
> + }
> mutex_unlock(&ioapic_mutex);
>
> return irq > 0 ? irq : -1;
> @@ -2923,18 +2945,27 @@ out:
>
> static int mp_irqdomain_create(int ioapic)
> {
> + size_t size;
> int hwirqs = mp_ioapic_pin_count(ioapic);
> struct ioapic *ip = &ioapics[ioapic];
> struct ioapic_domain_cfg *cfg = &ip->irqdomain_cfg;
> struct mp_ioapic_gsi *gsi_cfg = mp_ioapic_gsi_routing(ioapic);
>
> + size = sizeof(struct mp_pin_info) * mp_ioapic_pin_count(ioapic);
> + ip->pin_info = kzalloc(size, GFP_KERNEL);
> + if (!ip->pin_info)
> + return -ENOMEM;
> +
> if (cfg->type == IOAPIC_DOMAIN_INVALID)
> return 0;
>
> ip->irqdomain = irq_domain_add_linear(cfg->dev, hwirqs, cfg->ops,
> (void *)(long)ioapic);
> - if(!ip->irqdomain)
> + if(!ip->irqdomain) {
> + kfree(ip->pin_info);
> + ip->pin_info = NULL;
> return -ENOMEM;
> + }
>
> if (cfg->type == IOAPIC_DOMAIN_LEGACY ||
> cfg->type == IOAPIC_DOMAIN_STRICT)
> @@ -3898,6 +3929,72 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base,
> nr_ioapics++;
> }
>
> +int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
> + irq_hw_number_t hwirq)
> +{
> + int ioapic = (int)(long)domain->host_data;
> + struct mp_pin_info *info = mp_pin_info(ioapic, hwirq);
> + struct io_apic_irq_attr attr;
> +
> + /*
> + * Skip the timer IRQ if there's a quirk handler installed and if it
> + * returns 1:
> + */
> + if (apic->multi_timer_check &&
> + apic->multi_timer_check(ioapic, virq))
> + return 0;
> +
> + /* Get default attribute if not set by caller yet */
> + if (!info->set) {
> + u32 gsi = mp_pin_to_gsi(ioapic, hwirq);
> +
> + if (acpi_get_override_irq(gsi, &info->trigger,
> + &info->polarity) < 0) {
Sadly this seems to break LPSS ACPI enumerated devices.
Before this change /proc/interrupts say:
0: 14 0 0 0 IO-APIC-edge timer
1: 10 0 0 0 IO-APIC-edge i8042
4: 79 0 0 0 IO-APIC-edge serial
5: 52 0 0 0 IO-APIC-fasteoi mmc0
6: 0 0 0 0 IO-APIC-fasteoi dw_dmac
7: 0 0 0 0 IO-APIC-fasteoi INT3432:00, INT3433:00
8: 1 0 0 0 IO-APIC-edge rtc0
9: 174 0 0 0 IO-APIC-fasteoi acpi
57: 476 0 0 0 PCI-MSI-edge xhci_hcd
58: 17 0 0 0 PCI-MSI-edge snd_hda_intel
After the change it looks like this:
0: 14 0 0 0 IO-APIC-edge timer
1: 10 0 0 0 IO-APIC-edge i8042
4: 64 0 0 0 IO-APIC-edge serial
5: 0 0 0 0 IO-APIC-edge mmc0
6: 0 0 0 0 IO-APIC-edge dw_dmac
7: 0 0 0 0 IO-APIC-edge INT3432:00, INT3433:00
8: 1 0 0 0 IO-APIC-edge rtc0
9: 173 0 0 0 IO-APIC-fasteoi acpi
41: 471 0 0 0 PCI-MSI-edge xhci_hcd
42: 17 0 0 0 PCI-MSI-edge snd_hda_intel
Notice the interrupt triggering of IRQs 5, 6, and 7 has changed from level to
edge. I also see this printed on the console:
[ 1.676685] Failed to set pin attr for GSI6
[ 1.691708] Failed to set pin attr for GSI7
[ 1.706750] Failed to set pin attr for GSI7
[ 1.721768] Failed to set pin attr for GSI13
[ 1.736765] Failed to set pin attr for GSI5
[ 1.838342] Failed to set pin attr for GSI6
Any ideas how to get that fixed?
We used to handle this in drivers/acpi/resource.c:acpi_dev_get_irqresource() so
that only IRQ() and IRQNoFlags() ACPI resources resort calling
acpi_get_override_irq(). Now that doesn't help anymore.
> + /*
> + * PCI interrupts are always polarity one level
> + * triggered.
> + */
> + info->trigger = 1;
> + info->polarity = 1;
> + }
> + info->node = NUMA_NO_NODE;
> + info->set = 1;
> + }
> + set_io_apic_irq_attr(&attr, ioapic, hwirq, info->trigger,
> + info->polarity);
> +
> + return io_apic_setup_irq_pin(virq, info->node, &attr);
> +}
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists