Note: This is a beta release of Red Hat Bugzilla 5.0. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Also email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback here.
Bug 454417 - Inconsistent documentation regarding pci_alloc_consistent
Summary: Inconsistent documentation regarding pci_alloc_consistent
Alias: None
Product: Red Hat Enterprise Linux 4
Classification: Red Hat
Component: kernel
Version: 4.8
Hardware: All
OS: Linux
Target Milestone: rc
: ---
Assignee: Prarit Bhargava
QA Contact: Martin Jenner
Depends On:
Blocks: 461304
TreeView+ depends on / blocked
Reported: 2008-07-08 12:01 UTC by Prarit Bhargava
Modified: 2009-05-18 19:37 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Last Closed: 2009-05-18 19:37:12 UTC
Target Upstream Version:

Attachments (Terms of Use)
RHEL4 fix for this issue (deleted)
2008-07-08 12:01 UTC, Prarit Bhargava
no flags Details | Diff
RHEL4 fix for this issue [v2] (deleted)
2008-07-08 12:06 UTC, Prarit Bhargava
no flags Details | Diff
RHEL4 fix for this issue [v3] (deleted)
2008-07-08 12:10 UTC, Prarit Bhargava
no flags Details | Diff

System ID Priority Status Summary Last Updated
Red Hat Product Errata RHSA-2009:1024 normal SHIPPED_LIVE Important: Red Hat Enterprise Linux 4.8 kernel security and bug fix update 2009-05-18 14:57:26 UTC

Description Prarit Bhargava 2008-07-08 12:01:41 UTC
Description of problem:

William Reich noticed the following inconsistency between the DMA-API.txt file
and the DMA-mapping.txt file.  The DMA-mapping.txt file was backported into RHEL4.

Version-Release number of selected component (if applicable): 77.EL

Additional info:

pci_alloc_consistent() calls dma_alloc_coherent()

dma_alloc_coherent() executes the following code:

     memory = dma_alloc_pages(dev, gfp, get_order(size));
                int high, mmu;
                bus = virt_to_bus(memory);
                high = (bus + size) >= dma_mask; <<< by bumping the dma_mask,
                                                     you are simply increasing
                                                     this value.  Therefore,
                                                     high = 0 (always).
                mmu = high;
                if (force_iommu && !(gfp & GFP_DMA)) 
                        mmu = 1;
                if (no_iommu || dma_mask < 0xffffffffUL) { 
                        if (high) {
                                if (!(gfp & GFP_DMA)) { 
                                        gfp |= GFP_DMA; 
                                        free_pages((unsigned long)memory,
                                        goto again;
                                goto free;
                        mmu = 0; 
                memset(memory, 0, size); 
                if (!mmu) { 
                        *dma_handle = virt_to_bus(memory);
                        return memory;  <<< we always return here because mmu=0

In the case that that high evaluates to 1, we end up at another method of
getting a dma area,

        *dma_handle = dma_map_area(dev, bus, size, PCI_DMA_BIDIRECTIONAL, 0);

The call to dma_map_area *does not guarantee* any alignment.  It simply queries
the iommu for an empty area and returns pointers to the area (virtual and a dma

Comment 1 Prarit Bhargava 2008-07-08 12:01:42 UTC
Created attachment 311255 [details]
RHEL4 fix for this issue

Comment 2 Prarit Bhargava 2008-07-08 12:06:03 UTC
Created attachment 311256 [details]
RHEL4 fix for this issue [v2]

Updated patch.

Comment 3 Prarit Bhargava 2008-07-08 12:10:29 UTC
Created attachment 311259 [details]
RHEL4 fix for this issue [v3]

Updated patch.

Comment 4 Prarit Bhargava 2008-07-08 12:17:59 UTC
William (Reich) -- I submitted this patch internally to our kernel list.  That
means that this patch will be reviewed and eventually included in RHEL4.


Comment 5 RHEL Product and Program Management 2008-07-08 12:54:33 UTC
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux maintenance release.  Product Management has requested
further review of this request by Red Hat Engineering, for potential
inclusion in a Red Hat Enterprise Linux Update release for currently deployed
products.  This request is not yet committed for inclusion in an Update

Comment 6 William Reich 2008-07-11 19:47:25 UTC

This was stimulated by work on report #298811

Comment 7 RHEL Product and Program Management 2008-09-03 13:16:42 UTC
Updating PM score.

Comment 8 Vivek Goyal 2008-09-09 21:43:36 UTC
Committed in 78.8.EL . RPMS are available at

Comment 11 errata-xmlrpc 2009-05-18 19:37:12 UTC
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

Note You need to log in before you can comment on or make changes to this bug.