| 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: <545AFFF1.5030509@linux.intel.com>
Date: Thu, 06 Nov 2014 12:58:25 +0800
From: Jiang Liu <jiang.liu@...ux.intel.com>
To: Bjorn Helgaas <bhelgaas@...gle.com>
CC: Benjamin Herrenschmidt <benh@...nel.crashing.org>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Randy Dunlap <rdunlap@...radead.org>,
Yinghai Lu <yinghai@...nel.org>,
Borislav Petkov <bp@...en8.de>,
Grant Likely <grant.likely@...aro.org>,
Marc Zyngier <marc.zyngier@....com>,
Yingjoe Chen <yingjoe.chen@...iatek.com>,
Matthias Brugger <matthias.bgg@...il.com>,
Yijing Wang <wangyijing@...wei.com>,
Alexander Gordeev <agordeev@...hat.com>,
Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Tony Luck <tony.luck@...el.com>,
Joerg Roedel <joro@...tes.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
x86@...nel.org, linux-kernel@...r.kernel.org,
linux-pci@...r.kernel.org, linux-acpi@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org
Subject: Re: [Patch Part2 v4 21/31] PCI/MSI: enhance PCI MSI core to support
hierarchy irqdomain
On 2014/11/6 7:09, Bjorn Helgaas wrote:
> On Tue, Nov 04, 2014 at 08:01:55PM +0800, Jiang Liu wrote:
>
> In your topic:
>
> PCI/MSI: enhance PCI MSI core to support hierarchy irqdomain
>
> There's no need to repeat "PCI MSI". Please run "git log --oneline
> drivers/pci/msi.c" and make your similar (capitalize the first word).
Hi Bjornm
I'm already very carefully with your education about commit
log messages, but still missed this one:(. Will be even more careful
next time.
>
>> Enhance PCI MSI core to support hierarchy irqdomain, so the common
>> code could be shared among architectures.
>>
>> Signed-off-by: Jiang Liu <jiang.liu@...ux.intel.com>
>> ---
>> drivers/pci/Kconfig | 4 ++
>> drivers/pci/msi.c | 126 +++++++++++++++++++++++++++++++++++++++++++++++++++
>> include/linux/msi.h | 11 +++++
>> 3 files changed, 141 insertions(+)
>>
>> diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
>> index b9db0f2ce11f..022e89745f86 100644
>> --- a/drivers/pci/Kconfig
>> +++ b/drivers/pci/Kconfig
>> @@ -16,6 +16,10 @@ config PCI_MSI
>>
>> If you don't know what to do here, say Y.
>>
>> +config PCI_MSI_IRQ_DOMAIN
>> + bool
>> + depends on PCI_MSI && IRQ_DOMAIN_HIERARCHY
>> +
>> config PCI_DEBUG
>> bool "PCI Debugging"
>> depends on PCI && DEBUG_KERNEL
>> diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c
>> index da181c59394b..7423ee16972f 100644
>> --- a/drivers/pci/msi.c
>> +++ b/drivers/pci/msi.c
>> @@ -19,6 +19,7 @@
>> #include <linux/errno.h>
>> #include <linux/io.h>
>> #include <linux/slab.h>
>> +#include <linux/irqdomain.h>
>>
>> #include "pci.h"
>>
>> @@ -1098,3 +1099,128 @@ int pci_enable_msix_range(struct pci_dev *dev, struct msix_entry *entries,
>> return nvec;
>> }
>> EXPORT_SYMBOL(pci_enable_msix_range);
>> +
>> +#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
>
> Space, not tab.
Will fix it in next version.
>
>> +static inline irq_hw_number_t
>> +msi_get_hwirq(struct pci_dev *pdev, struct msi_desc *msidesc)
>
> The convention in this file is "struct pci_dev *dev". And "struct msi_desc
> *desc" (or maybe "*entry"). Try to converge things, not diverge them.
Thanks for reminder. Adding another check item to my list before
sending out patches:)
>
>> +{
>> + return (irq_hw_number_t)msidesc->msi_attrib.entry_nr |
>> + PCI_DEVID(pdev->bus->number, pdev->devfn) << 11 |
>> + (pci_domain_nr(pdev->bus) & 0xFFFFFFFF) << 27;
>
> Where does this bit layout come from? Is this defined in the spec
> somewhere? A reference would help.
We need a unique number to identify every possible MSI source,
and this ID number is only used within the irqdomain subsystem.
So we used above algorithm to generate the ID number, there's
no specification for it.
>
>> +}
>> +
>> +static int msi_domain_alloc(struct irq_domain *domain, unsigned int virq,
>> + unsigned int nr_irqs, void *arg)
>> +{
>> + int i, ret;
>> + irq_hw_number_t hwirq = arch_msi_irq_domain_get_hwirq(arg);
>> +
>> + if (irq_find_mapping(domain, hwirq) > 0)
>> + return -EEXIST;
>> +
>> + ret = irq_domain_alloc_irqs_parent(domain, virq, nr_irqs, arg);
>> + if (ret >= 0)
>
> if (ret < 0)
> return ret;
>
> and un-indent the mainline code below. Then it's obvious that this is the
> normal case, not the error case.
Sure, I want to only use one return statement, but didn't realized that
syntax seems like error handling:)
>
>> + for (i = 0; i < nr_irqs; i++) {
>> + irq_domain_set_hwirq_and_chip(domain, virq + i,
>> + hwirq + i, &msi_chip, (void *)(long)i);
>> + __irq_set_handler(virq + i, handle_edge_irq, 0, "edge");
>> + }
>> +
>> + return ret;
>> +}
>> +
>> +static void msi_domain_free(struct irq_domain *domain, unsigned int virq,
>> + unsigned int nr_irqs)
>> +{
>> + int i;
>> +
>> + for (i = 0; i < nr_irqs; i++) {
>> + struct msi_desc *msidesc = irq_get_msi_desc(virq);
>> +
>> + if (msidesc)
>> + msidesc->irq = 0;
>> + }
>> + irq_domain_free_irqs_top(domain, virq, nr_irqs);
>> +}
>> +
>> +static int msi_domain_activate(struct irq_domain *domain,
>> + struct irq_data *irq_data)
>> +{
>> + int ret = 0;
>> + struct msi_msg msg;
>> +
>> + /*
>> + * irq_data->chip_data is MSI/MSIx offset.
>
> "MSI-X", as you wrote on the next line.
Sure.
>
>> + * MSI-X message is written per-IRQ, the offset is always 0.
>> + * MSI message denotes a contiguous group of IRQs, written for 0th IRQ.
>> + */
>> + if (!irq_data->chip_data) {
>
> if (irq_data->chip_data)
> return 0;
>
> and un-indent the mainline code below, and drop the "ret = 0" init above.
>
>> + ret = irq_chip_compose_msi_msg(irq_data, &msg);
>> + if (ret == 0)
>
> if (ret)
> return ret;
>
>> + write_msi_msg(irq_data->irq, &msg);
>> + }
>> +
>> + return ret;
> return 0;
>> +}
>> +
>> +static int msi_domain_deactivate(struct irq_domain *domain,
>> + struct irq_data *irq_data)
>> +{
>> + struct msi_msg msg;
>> +
>> + if (irq_data->chip_data) {
>> + memset(&msg, 0, sizeof(msg));
>> + write_msi_msg(irq_data->irq, &msg);
>> + }
>> +
>> + return 0;
>> +}
>> +
>> +static struct irq_domain_ops msi_domain_ops = {
>> + .alloc = msi_domain_alloc,
>> + .free = msi_domain_free,
>> + .activate = msi_domain_activate,
>> + .deactivate = msi_domain_deactivate,
>> +};
>> +
>> +struct irq_domain *msi_create_irq_domain(struct irq_domain *parent)
>> +{
>> + struct irq_domain *domain;
>> +
>> + domain = irq_domain_add_tree(NULL, &msi_domain_ops, NULL);
>> + if (domain)
>
> if (!domain)
> return NULL;
>
> and un-indent this:
>
>> + domain->parent = parent;
>> +
>> + return domain;
>> +}
>> +
>> +int msi_irq_domain_alloc_irqs(struct irq_domain *domain, int type,
>> + struct pci_dev *dev, void *arg)
>> +{
>> + int i, virq;
>> + struct msi_desc *msidesc;
>> + int node = dev_to_node(&dev->dev);
>> +
>> + list_for_each_entry(msidesc, &dev->msi_list, list) {
>> + arch_msi_irq_domain_set_hwirq(arg, msi_get_hwirq(dev, msidesc));
>> + virq = irq_domain_alloc_irqs(domain, msidesc->nvec_used,
>> + node, arg);
>> + if (virq < 0) {
>> + /* Special handling for pci_enable_msi_range(). */
>> + return (type == PCI_CAP_ID_MSI &&
>> + msidesc->nvec_used > 1) ? 1 : -ENOSPC;
>
> I think "if" would be easier to read than this ternary expression.
Sure.
>
>> + }
>> + for (i = 0; i < msidesc->nvec_used; i++)
>> + irq_set_msi_desc_off(virq + i, i, msidesc);
>> + }
>> +
>> + list_for_each_entry(msidesc, &dev->msi_list, list)
>> + if (msidesc->nvec_used == 1)
>> + dev_dbg(&dev->dev, "irq %d for MSI/MSI-X\n", virq);
>> + else
>> + dev_dbg(&dev->dev, "irq [%d-%d] for MSI/MSI-X\n",
>> + virq, virq + msidesc->nvec_used - 1);
>> +
>> + return 0;
>> +}
>> +#endif /* CONFIG_PCI_MSI_IRQ_DOMAIN */
>> diff --git a/include/linux/msi.h b/include/linux/msi.h
>> index 44f4746d033b..05dcd425f82b 100644
>> --- a/include/linux/msi.h
>> +++ b/include/linux/msi.h
>> @@ -75,4 +75,15 @@ struct msi_chip {
>> void (*teardown_irq)(struct msi_chip *chip, unsigned int irq);
>> };
>>
>> +#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
>
> Use a space here, not a tab.
Sure.
>
>> +extern struct irq_chip msi_chip;
>
> I don't think "msi_chip" is a good name. "Chip" only hints that it's a
> semiconductor integrated circuit; it doesn't say anything about what it
> does. I've suggested "msi_controller" elsewhere.
>
> Why does this need to be exported? And why should there be only one in a
> system?
I have changed the interfaces as below in next version, so we could
hide "msi_chip" private and support different irq_chip for different
irqdomains.
struct irq_domain *msi_create_irq_domain(struct device_node *of_node,
struct irq_chip *chip,
struct irq_domain *parent);
>
>> +extern struct irq_domain *msi_create_irq_domain(struct irq_domain *parent);
>> +extern int msi_irq_domain_alloc_irqs(struct irq_domain *domain, int type,
>> + struct pci_dev *dev, void *arg);
>> +
>> +extern irq_hw_number_t arch_msi_irq_domain_get_hwirq(void *arg);
>> +extern void arch_msi_irq_domain_set_hwirq(void *arg, irq_hw_number_t hwirq);
>
> Look at the rest of the file and notice that the existing code does not use
> "extern" on function declarations.
Sure.
>
>> +#endif /* CONFIG_PCI_MSI_IRQ_DOMAIN */
>
> Use a space here (not a tab), like the #endif just below.
Sure.
Thanks for your review and great comments!
Regards!
Gerry
>
>> #endif /* LINUX_MSI_H */
>> --
>> 1.7.10.4
>>
--
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