From: Joerg Roedel <joerg.roedel@amd.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: tglx@linutronix.de, mingo@redhat.com,
linux-kernel@vger.kernel.org, iommu@lists.linux-foundation.org,
bhavna.sarathy@amd.com, Sebastian.Biemueller@amd.com,
robert.richter@amd.com, joro@8bytes.org
Subject: Re: [PATCH 14/34] AMD IOMMU: clue initialization code together
Date: Thu, 10 Jul 2008 14:37:47 +0200 [thread overview]
Message-ID: <20080710123747.GK14977@amd.com> (raw)
In-Reply-To: <20080709185512.6355159c.akpm@linux-foundation.org>
On Wed, Jul 09, 2008 at 06:55:12PM -0700, Andrew Morton wrote:
> On Thu, 26 Jun 2008 21:27:50 +0200 Joerg Roedel <joerg.roedel@amd.com> wrote:
>
> > This patch puts the AMD IOMMU ACPI table parsing and hardware initialization
> > functions together to the main intialization routine.
> >
> > index 555fcc9..c792ddc 100644
> > --- a/arch/x86/kernel/amd_iommu_init.c
> > +++ b/arch/x86/kernel/amd_iommu_init.c
> > @@ -643,3 +643,129 @@ static int __init init_memory_definitions(struct acpi_table_header *table)
> > return 0;
> > }
> >
> > +int __init amd_iommu_init(void)
> > +{
> > + int i, ret = 0;
> > +
> > +
> > + if (amd_iommu_disable) {
> > + printk(KERN_INFO "AMD IOMMU disabled by kernel command line\n");
> > + return 0;
> > + }
> > +
> > + /*
> > + * First parse ACPI tables to find the largest Bus/Dev/Func
> > + * we need to handle. Upon this information the shared data
> > + * structures for the IOMMUs in the system will be allocated
> > + */
> > + if (acpi_table_parse("IVRS", find_last_devid_acpi) != 0)
> > + return -ENODEV;
> > +
> > + dev_table_size = TBL_SIZE(DEV_TABLE_ENTRY_SIZE);
> > + alias_table_size = TBL_SIZE(ALIAS_TABLE_ENTRY_SIZE);
> > + rlookup_table_size = TBL_SIZE(RLOOKUP_TABLE_ENTRY_SIZE);
> > +
> > + ret = -ENOMEM;
> > +
> > + /* Device table - directly used by all IOMMUs */
> > + amd_iommu_dev_table = (void *)__get_free_pages(GFP_KERNEL,
> > + get_order(dev_table_size));
>
> Is there a reason for using __get_free_pages() in this manner everywhere?
> kmalloc() or kzalloc() are unsuitable?
Yes. The device table has to be page aligned and has a maximum size of
2 MB.
> > + if (amd_iommu_dev_table == NULL)
> > + goto out;
> > +
> > + /*
> > + * Alias table - map PCI Bus/Dev/Func to Bus/Dev/Func the
> > + * IOMMU see for that device
> > + */
> > + amd_iommu_alias_table = (void *)__get_free_pages(GFP_KERNEL,
> > + get_order(alias_table_size));
Hmm, maybe this table will fit into a kmalloc allocation. Its maximum
size is 128 kb.
> > + if (amd_iommu_alias_table == NULL)
> > + goto free;
> > +
> > + /* IOMMU rlookup table - find the IOMMU for a specific device */
> > + amd_iommu_rlookup_table = (void *)__get_free_pages(GFP_KERNEL,
> > + get_order(rlookup_table_size));
> > + if (amd_iommu_rlookup_table == NULL)
> > + goto free;
> > +
> > + /*
> > + * Protection Domain table - maps devices to protection domains
> > + * This table has the same size as the rlookup_table
> > + */
> > + amd_iommu_pd_table = (void *)__get_free_pages(GFP_KERNEL,
> > + get_order(rlookup_table_size));
The rlookup and the pd table have a maximum size of 1 MB.
> > + if (amd_iommu_pd_table == NULL)
> > + goto free;
> > +
> > + amd_iommu_pd_alloc_bitmap = (void *)__get_free_pages(GFP_KERNEL,
> > + get_order(MAX_DOMAIN_ID/8));
> > + if (amd_iommu_pd_alloc_bitmap == NULL)
> > + goto free;
> > +
> > + /*
> > + * memory is allocated now; initialize the device table with all zeroes
> > + * and let all alias entries point to itself
> > + */
> > + memset(amd_iommu_dev_table, 0, dev_table_size);
> > + for (i = 0; i < amd_iommu_last_bdf; ++i)
> > + amd_iommu_alias_table[i] = i;
> > +
> > + memset(amd_iommu_pd_table, 0, rlookup_table_size);
> > + memset(amd_iommu_pd_alloc_bitmap, 0, MAX_DOMAIN_ID / 8);
>
> Use of __GFP_ZERO in here would relieve us of a lot of memsets.
>
> > + /*
> > + * never allocate domain 0 because its used as the non-allocated and
> > + * error value placeholder
> > + */
> > + amd_iommu_pd_alloc_bitmap[0] = 1;
> > +
> > + /*
> > + * now the data structures are allocated and basically initialized
> > + * start the real acpi table scan
> > + */
> > + ret = -ENODEV;
> > + if (acpi_table_parse("IVRS", init_iommu_all) != 0)
> > + goto free;
> > +
> > + if (acpi_table_parse("IVRS", init_memory_definitions) != 0)
> > + goto free;
> > +
> > + printk(KERN_INFO "AMD IOMMU: aperture size is %d MB\n",
> > + (1 << (amd_iommu_aperture_order-20)));
> > +
> > + printk(KERN_INFO "AMD IOMMU: device isolation ");
> > + if (amd_iommu_isolate)
> > + printk("enabled\n");
> > + else
> > + printk("disabled\n");
> > +
> > +out:
> > + return ret;
> > +
> > +free:
> > + if (amd_iommu_pd_alloc_bitmap)
> > + free_pages((unsigned long)amd_iommu_pd_alloc_bitmap, 1);
> > +
> > + if (amd_iommu_pd_table)
> > + free_pages((unsigned long)amd_iommu_pd_table,
> > + get_order(rlookup_table_size));
> > +
> > + if (amd_iommu_rlookup_table)
> > + free_pages((unsigned long)amd_iommu_rlookup_table,
> > + get_order(rlookup_table_size));
> > +
> > + if (amd_iommu_alias_table)
> > + free_pages((unsigned long)amd_iommu_alias_table,
> > + get_order(alias_table_size));
> > +
> > + if (amd_iommu_dev_table)
> > + free_pages((unsigned long)amd_iommu_dev_table,
> > + get_order(dev_table_size));
>
> free_pages(0) is legal, so all these tests are unneeded, I expect.
>
> This remains true if you decide to switch to kzalloc().
Thanks for that hint. I will remove the checks.
> > + free_iommu_all();
> > +
> > + free_unity_maps();
> > +
> > + goto out;
> > +}
>
>
--
| AMD Saxony Limited Liability Company & Co. KG
Operating | Wilschdorfer Landstr. 101, 01109 Dresden, Germany
System | Register Court Dresden: HRA 4896
Research | General Partner authorized to represent:
Center | AMD Saxony LLC (Wilmington, Delaware, US)
| General Manager of AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy
next prev parent reply other threads:[~2008-07-10 12:39 UTC|newest]
Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-06-26 19:27 [PATCH 00/34] AMD IOMMU driver Joerg Roedel
2008-06-26 19:27 ` [PATCH 01/34] AMD IOMMU: add Kconfig entry Joerg Roedel
2008-06-27 14:25 ` Adrian Bunk
2008-06-27 14:47 ` Andi Kleen
2008-06-27 16:39 ` Muli Ben-Yehuda
2008-06-27 16:54 ` Joerg Roedel
2008-06-27 16:59 ` Muli Ben-Yehuda
2008-06-27 17:05 ` Joerg Roedel
2008-06-27 17:12 ` Muli Ben-Yehuda
2008-06-27 17:20 ` Joerg Roedel
2008-06-27 17:31 ` Muli Ben-Yehuda
2008-06-27 17:40 ` Joerg Roedel
2008-06-27 17:44 ` Muli Ben-Yehuda
2008-06-27 17:52 ` Joerg Roedel
2008-07-02 5:45 ` Amit Shah
2008-07-02 8:12 ` Alan Cox
2008-06-27 20:39 ` Duran, Leo
2008-06-27 22:29 ` Muli Ben-Yehuda
2008-06-27 22:47 ` Duran, Leo
2008-06-28 11:04 ` Joerg Roedel
2008-06-28 14:40 ` Duran, Leo
2008-06-28 16:27 ` Joerg Roedel
2008-06-28 14:58 ` Duran, Leo
2008-06-27 18:54 ` Andi Kleen
2008-06-28 10:52 ` Joerg Roedel
2008-06-27 16:40 ` Joerg Roedel
2008-07-08 12:45 ` Pavel Machek
2008-07-09 8:48 ` Ingo Molnar
2008-07-10 0:50 ` FUJITA Tomonori
2008-07-10 8:44 ` Ingo Molnar
2008-06-26 19:27 ` [PATCH 02/34] AMD IOMMU: add header file for driver data structures and defines Joerg Roedel
2008-06-29 15:07 ` FUJITA Tomonori
2008-06-29 15:14 ` Joerg Roedel
2008-06-29 23:11 ` FUJITA Tomonori
2008-06-30 12:22 ` Ingo Molnar
2008-07-10 1:38 ` Andrew Morton
2008-07-10 1:50 ` Arjan van de Ven
2008-07-10 2:36 ` Andrew Morton
2008-07-10 12:12 ` Joerg Roedel
2008-07-10 12:59 ` Andrew Morton
2008-06-26 19:27 ` [PATCH 03/34] AMD IOMMU: add defines and structures for ACPI scanning code Joerg Roedel
2008-07-10 1:41 ` Andrew Morton
2008-07-10 12:17 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 04/34] AMD IOMMU: add data structures to manage the IOMMUs in the system Joerg Roedel
2008-07-10 1:43 ` Andrew Morton
2008-07-10 12:25 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 05/34] AMD IOMMU: add functions to find last possible PCI device for IOMMU Joerg Roedel
2008-06-26 19:27 ` [PATCH 06/34] AMD IOMMU: add amd_iommu_init.c to Makefile Joerg Roedel
2008-06-26 19:27 ` [PATCH 07/34] AMD IOMMU: add functions for mapping/unmapping the MMIO space Joerg Roedel
2008-06-26 19:27 ` [PATCH 08/34] AMD IOMMU: add functions for programming IOMMU " Joerg Roedel
2008-06-26 19:27 ` [PATCH 09/34] AMD IOMMU: add command buffer (de-)allocation Joerg Roedel
2008-07-10 1:47 ` Andrew Morton
2008-06-26 19:27 ` [PATCH 10/34] AMD IOMMU: add device table initialization functions Joerg Roedel
2008-06-26 19:27 ` [PATCH 11/34] AMD IOMMU: add functions for IOMMU hardware initialization from ACPI Joerg Roedel
2008-07-10 1:49 ` Andrew Morton
2008-06-26 19:27 ` [PATCH 12/34] AMD IOMMU: add detect code for AMD IOMMU hardware Joerg Roedel
2008-07-10 1:51 ` Andrew Morton
2008-06-26 19:27 ` [PATCH 13/34] AMD IOMMU: add functions to parse IOMMU memory mapping requirements for devices Joerg Roedel
2008-07-10 1:51 ` Andrew Morton
2008-06-26 19:27 ` [PATCH 14/34] AMD IOMMU: clue initialization code together Joerg Roedel
2008-07-10 1:55 ` Andrew Morton
2008-07-10 12:37 ` Joerg Roedel [this message]
2008-07-10 13:03 ` Andrew Morton
2008-07-10 13:31 ` Joerg Roedel
2008-07-10 18:37 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 15/34] AMD IOMMU: add early detection code Joerg Roedel
2008-06-26 19:27 ` [PATCH 16/34] AMD IOMMU: add kernel command line parameters for AMD IOMMU Joerg Roedel
2008-07-10 1:56 ` Andrew Morton
2008-07-10 4:25 ` FUJITA Tomonori
2008-07-10 4:39 ` Andrew Morton
2008-07-10 6:26 ` Ingo Molnar
2008-07-10 20:42 ` Randy Dunlap
2008-07-14 23:56 ` FUJITA Tomonori
2008-07-10 7:04 ` Yinghai Lu
2008-07-10 12:41 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 17/34] AMD IOMMU: add generic defines and structures for mapping code Joerg Roedel
2008-07-10 2:01 ` Andrew Morton
2008-07-10 2:38 ` Andrew Morton
2008-07-10 4:25 ` FUJITA Tomonori
2008-07-10 12:44 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 18/34] AMD IOMMU: add amd_iommu.c to Makefile Joerg Roedel
2008-06-26 19:27 ` [PATCH 19/34] AMD IOMMU: add functions to send IOMMU commands Joerg Roedel
2008-07-10 2:04 ` Andrew Morton
2008-07-10 12:53 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 20/34] AMD IOMMU: add functions to initialize unity mappings Joerg Roedel
2008-06-26 19:27 ` [PATCH 21/34] AMD IOMMU: add address allocation and deallocation functions Joerg Roedel
2008-06-29 15:07 ` FUJITA Tomonori
2008-06-29 15:17 ` Joerg Roedel
2008-06-29 23:11 ` FUJITA Tomonori
2008-06-26 19:27 ` [PATCH 22/34] AMD IOMMU: add domain " Joerg Roedel
2008-07-10 2:14 ` Andrew Morton
2008-07-10 12:54 ` Joerg Roedel
2008-06-26 19:27 ` [PATCH 23/34] AMD IOMMU: add functions to find IOMMU device resources Joerg Roedel
2008-07-10 2:18 ` Andrew Morton
2008-07-10 16:46 ` Joerg Roedel
2008-07-10 23:59 ` Andrew Morton
2008-06-26 19:28 ` [PATCH 24/34] AMD IOMMU: add generic dma_ops mapping functions Joerg Roedel
2008-06-26 19:28 ` [PATCH 25/34] AMD IOMMU: add dma_ops mapping functions for single mappings Joerg Roedel
2008-07-10 2:26 ` Andrew Morton
2008-07-10 13:20 ` Joerg Roedel
2008-06-26 19:28 ` [PATCH 26/34] AMD IOMMU: add mapping functions for scatter gather lists Joerg Roedel
2008-06-29 15:07 ` FUJITA Tomonori
2008-06-30 13:25 ` Joerg Roedel
2008-06-26 19:28 ` [PATCH 27/34] AMD IOMMU: add mapping functions for coherent mappings Joerg Roedel
2008-06-26 19:28 ` [PATCH 28/34] AMD IOMMU: add pre-allocation of protection domains Joerg Roedel
2008-06-26 19:28 ` [PATCH 29/34] AMD IOMMU: add dma_ops initialization function Joerg Roedel
2008-06-26 19:28 ` [PATCH 30/34] AMD IOMMU: add amd_iommu.h to export functions to the generic x86 dma code Joerg Roedel
2008-06-26 19:28 ` [PATCH 31/34] AMD IOMMU: initialize dma_ops from IOMMU initialization and enable IOMMUs Joerg Roedel
2008-06-26 19:28 ` [PATCH 32/34] AMD_IOMMU: call detect and initialization functions from dma code Joerg Roedel
2008-06-26 19:28 ` [PATCH 33/34] AMD IOMMU: add MAINTAINERS entry Joerg Roedel
2008-06-26 19:28 ` [PATCH 34/34] AMD IOMMU: add documentation for kernel parameters Joerg Roedel
2008-06-29 15:07 ` FUJITA Tomonori
2008-06-30 12:25 ` Ingo Molnar
2008-06-26 20:37 ` [PATCH 00/34] AMD IOMMU driver Rafael J. Wysocki
2008-06-26 20:37 ` Joerg Roedel
2008-06-26 21:02 ` Rafael J. Wysocki
2008-06-27 8:18 ` Ingo Molnar
2008-06-27 10:07 ` Ingo Molnar
2008-06-27 10:15 ` Joerg Roedel
2008-06-27 10:59 ` Joerg Roedel
2008-07-11 10:22 ` Eric W. Biederman
2008-07-11 14:11 ` Joerg Roedel
2008-07-11 16:23 ` Duran, Leo
2008-07-11 17:20 ` Eric W. Biederman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20080710123747.GK14977@amd.com \
--to=joerg.roedel@amd.com \
--cc=Sebastian.Biemueller@amd.com \
--cc=akpm@linux-foundation.org \
--cc=bhavna.sarathy@amd.com \
--cc=iommu@lists.linux-foundation.org \
--cc=joro@8bytes.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=robert.richter@amd.com \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox