Message-ID: <a5cdc7f8-b340-d372-2971-0d24b01de217@gmail.com>
Date: Sun, 2 Jul 2023 19:13:14 +0700
From: Bagas Sanjaya <bagasdotme@...il.com>
To: Johannes Berg <johannes.berg@...el.com>,
 Gregory Greenman <gregory.greenman@...el.com>,
 Niklāvs Koļesņikovs <pinkflames.linux@...il.com>,
 Nate Watterson <nwatters@...eaurora.org>, Robin Murphy
 <robin.murphy@....com>, Joerg Roedel <joro@...tes.org>,
 Will Deacon <will@...nel.org>,
 Linux IO Memory Management Unit <iommu@...ts.linux.dev>
Cc: Linux Wireless <linux-wireless@...r.kernel.org>,
 Linux Networking <netdev@...r.kernel.org>
Subject: Fwd: iwlwifi causes dma-iommu.c:693 __iommu_dma_unmap since commit
 19898ce9cf8a

Hi,

I notice a regression report on Bugzilla [1]: Quoting from it:

> Since commit 19898ce9cf8a the iwlwifi has generated three possibly identical kernel stack traces for me. Because I only use the Bluetooth but not the Wi-Fi functionality, this is not a big deal for me but I thought such an issue is worth reporting nontheless.
> 
> All three traces point at **drivers/iommu/dma-iommu.c:693 __iommu_dma_unmap+0x150/0x160**.
> 
> I'm attaching to this bug report the three stack traces along with other possibly relevant dmesg parts. Sorry in advance for not cutting at the cut here markers which resulted in considerably longer text but I suspected that the PCI, ACPI, memory and possibly iwlwifi related messages may be of importance, too. If I should cut the stack traces out and attach them as three distinct files (and diff to see if there's any change between them) let me know. I can provide a full (but redacted) dmesg output of a git master build, if required as well.
> 
> I did try booting a much more recent git master build with *iommu.passthrough=0 iommu.strict=0* on the kernel command line but that did not seem to make any difference.
> 
> ```
> 19898ce9cf8a33e0ac35cb4c7f68de297cc93cb2 is the first bad commit
> commit 19898ce9cf8a33e0ac35cb4c7f68de297cc93cb2
> Author: Johannes Berg <johannes.berg@...el.com>
> Date:   Wed Jun 21 13:12:07 2023 +0300
> 
>     wifi: iwlwifi: split 22000.c into multiple files
>     
>     Split the configuration list in 22000.c into four new files,
>     per new device family, so we don't have this huge unusable
>     file. Yes, this duplicates a few small things, but that's
>     still much better than what we have now.
>     
>     Signed-off-by: Johannes Berg <johannes.berg@...el.com>
>     Signed-off-by: Gregory Greenman <gregory.greenman@...el.com>
>     Link: https://lore.kernel.org/r/20230621130443.7543603b2ee7.Ia8dd54216d341ef1ddc0531f2c9aa30d30536a5d@changeid
>     Signed-off-by: Johannes Berg <johannes.berg@...el.com>
> 
>  drivers/net/wireless/intel/iwlwifi/Makefile     |   1 +
>  drivers/net/wireless/intel/iwlwifi/cfg/22000.c  | 939 +-----------------------
>  drivers/net/wireless/intel/iwlwifi/cfg/ax210.c  | 452 ++++++++++++
>  drivers/net/wireless/intel/iwlwifi/cfg/bz.c     | 523 +++++++++++++
>  drivers/net/wireless/intel/iwlwifi/cfg/sc.c     | 214 ++++++
>  drivers/net/wireless/intel/iwlwifi/iwl-config.h |   2 +
>  drivers/net/wireless/intel/iwlwifi/pcie/drv.c   |   3 +
>  7 files changed, 1206 insertions(+), 928 deletions(-)
>  create mode 100644 drivers/net/wireless/intel/iwlwifi/cfg/ax210.c
>  create mode 100644 drivers/net/wireless/intel/iwlwifi/cfg/bz.c
>  create mode 100644 drivers/net/wireless/intel/iwlwifi/cfg/sc.c
> ```
> 

See Bugzilla for the full thread and attached dmesg.

Anyway, I'm adding it to regzbot to ensure that it doesn't fall through
cracks unnoticed:

#regzbot introduced: 19898ce9cf8a33 https://bugzilla.kernel.org/show_bug.cgi?id=217622
#regzbot title: dma-iommu.c:693 __iommu_dma_unmap bug trace due to 22000.c split

Thanks.

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=217622

-- 
An old man doll... just what I always wanted! - Clara

Powered by blists - more mailing lists