Bug 14121 - Wireless Cards Fail to Intitialize - ath5k 32-bit DMA not available
Summary: Wireless Cards Fail to Intitialize - ath5k 32-bit DMA not available
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Networking
Classification: Unclassified
Component: Wireless (show other bugs)
Hardware: All Linux
: P1 high
Assignee: networking_wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-09-04 16:23 UTC by Jim Gifford
Modified: 2010-06-07 19:47 UTC (History)
5 users (show)

See Also:
Kernel Version: 2.6.30.x
Subsystem:
Regression: No
Bisected commit-id:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Description Jim Gifford 2009-09-04 16:23:56 UTC 
Worked perfectly under 2.6.29.x series, but with the 2.6.30.x series, when I plugin in my ath5k card into my Powerbook, I get the following messages and the card is not usable.

[90539.033008] pcmcia_socket pcmcia_socket0: pccard: CardBus card inserted into slot 0
[90539.033538] pci 0000:01:00.0: reg 10 32bit mmio: [0x000000-0x00ffff]
[90539.096778] cfg80211: Calling CRDA to update world regulatory domain
[90539.245672] ath5k 0000:01:00.0: enabling device (0000 -> 0002)
[90539.246150] ath5k 0000:01:00.0: 32-bit DMA not available
[90539.246402] ath5k: probe of 0000:01:00.0 failed with error -5
Comment 1 Nick Kossifidis 2009-09-08 11:55:42 UTC 
/* XXX 32-bit addressing only */
ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
if (ret) {
        dev_err(&pdev->dev, "32-bit DMA not available\n");
        goto err_dis;
}

Seems that pci_set_dma_mask returns an error for some reason. Doesn't seem like an ath5k bug.
Comment 2 Roger Blofeld 2009-11-14 15:08:55 UTC 
I think this is due to commit 4ae0ff606e848fa4957ebf8f97e5db5fdeec27be. There is a missing "set_dma_ops(dev, &dma_direct_ops)" somewhere, but I don't know where to put it.

I have the same probe failed problem with a cardbus sata_sil device on my g4 powerbook, and adding that code to the sata_sil.c code fixed the problem for me.
Comment 3 John W. Linville 2009-11-16 15:16:02 UTC 
Forwarded link to Kumar Gala, the author of the commit referenced above -- hopefully he has some insight. (galak@kernel.crashing.org doesn't seem to work for Cc list...)
Comment 4 Dominik Brodowski 2010-02-23 00:28:19 UTC 
This seems to be fixed in recent kernels (2.6.33-rcX) by means of 2d1c861871d767153538a77c498752b36d4bb4b8 (PCI/cardbus: Add a fixup hook and fix powerpc). Could you confirm this, please?
Comment 5 John W. Linville 2010-03-02 16:23:38 UTC 
Jim, please respond re: comment 4...thanks!
Comment 6 John W. Linville 2010-06-07 19:47:01 UTC 
Closed due to lack of response...
Note You need to log in before you can comment on or make changes to this bug.