linux-kernel - Re: [RFC PATCH v2 8/8] dax: Fix incorrect list of dcache aliasing architectures

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 31 Jan 2024 10:14:22 -0500
From: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To: Dan Williams <dan.j.williams@...el.com>,
 Dave Chinner <david@...morbit.com>
Cc: Vishal Verma <vishal.l.verma@...el.com>, Dave Jiang
 <dave.jiang@...el.com>, linux-kernel@...r.kernel.org,
 Andrew Morton <akpm@...ux-foundation.org>,
 Linus Torvalds <torvalds@...ux-foundation.org>, linux-mm@...ck.org,
 linux-arch@...r.kernel.org, Matthew Wilcox <willy@...radead.org>,
 Arnd Bergmann <arnd@...db.de>, Russell King <linux@...linux.org.uk>,
 nvdimm@...ts.linux.dev, linux-cxl@...r.kernel.org,
 linux-fsdevel@...r.kernel.org
Subject: Re: [RFC PATCH v2 8/8] dax: Fix incorrect list of dcache aliasing
 architectures

On 2024-01-30 22:13, Dan Williams wrote:
> Dave Chinner wrote:
>> On Tue, Jan 30, 2024 at 11:52:55AM -0500, Mathieu Desnoyers wrote:
>>> commit d92576f1167c ("dax: does not work correctly with virtual aliasing caches")
>>> prevents DAX from building on architectures with virtually aliased
>>> dcache with:
>>>
>>>    depends on !(ARM || MIPS || SPARC)
>>>
>>> This check is too broad (e.g. recent ARMv7 don't have virtually aliased
>>> dcaches), and also misses many other architectures with virtually
>>> aliased dcache.
>>>
>>> This is a regression introduced in the v5.13 Linux kernel where the
>>> dax mount option is removed for 32-bit ARMv7 boards which have no dcache
>>> aliasing, and therefore should work fine with FS_DAX.
>>>
>>> This was turned into the following implementation of dax_is_supported()
>>> by a preparatory change:
>>>
>>>          return !IS_ENABLED(CONFIG_ARM) &&
>>>                 !IS_ENABLED(CONFIG_MIPS) &&
>>>                 !IS_ENABLED(CONFIG_SPARC);
>>>
>>> Use dcache_is_aliasing() instead to figure out whether the environment
>>> has aliasing dcaches.
>>>
>>> Fixes: d92576f1167c ("dax: does not work correctly with virtual aliasing caches")
>>> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
>>> Cc: Andrew Morton <akpm@...ux-foundation.org>
>>> Cc: Linus Torvalds <torvalds@...ux-foundation.org>
>>> Cc: linux-mm@...ck.org
>>> Cc: linux-arch@...r.kernel.org
>>> Cc: Dan Williams <dan.j.williams@...el.com>
>>> Cc: Vishal Verma <vishal.l.verma@...el.com>
>>> Cc: Dave Jiang <dave.jiang@...el.com>
>>> Cc: Matthew Wilcox <willy@...radead.org>
>>> Cc: Arnd Bergmann <arnd@...db.de>
>>> Cc: Russell King <linux@...linux.org.uk>
>>> Cc: nvdimm@...ts.linux.dev
>>> Cc: linux-cxl@...r.kernel.org
>>> Cc: linux-fsdevel@...r.kernel.org
>>> ---
>>>   include/linux/dax.h | 5 ++---
>>>   1 file changed, 2 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/include/linux/dax.h b/include/linux/dax.h
>>> index cfc8cd4a3eae..f59e604662e4 100644
>>> --- a/include/linux/dax.h
>>> +++ b/include/linux/dax.h
>>> @@ -5,6 +5,7 @@
>>>   #include <linux/fs.h>
>>>   #include <linux/mm.h>
>>>   #include <linux/radix-tree.h>
>>> +#include <linux/cacheinfo.h>
>>>   
>>>   typedef unsigned long dax_entry_t;
>>>   
>>> @@ -80,9 +81,7 @@ static inline bool daxdev_mapping_supported(struct vm_area_struct *vma,
>>>   }
>>>   static inline bool dax_is_supported(void)
>>>   {
>>> -	return !IS_ENABLED(CONFIG_ARM) &&
>>> -	       !IS_ENABLED(CONFIG_MIPS) &&
>>> -	       !IS_ENABLED(CONFIG_SPARC);
>>> +	return !dcache_is_aliasing();
>>
>> Yeah, if this is just a one liner should go into
>> fs_dax_get_by_bdev(), similar to the blk_queue_dax() check at the
>> start of the function.
>>
>> I also noticed that device mapper uses fs_dax_get_by_bdev() to
>> determine if it can support DAX, but this patch set does not address
>> that case. Hence it really seems to me like fs_dax_get_by_bdev() is
>> the right place to put this check.
> 
> Oh, good catch. Yes, I agree this can definitely be pushed down, but
> then I think it should be pushed down all the way to make alloc_dax()
> fail. That will need some additional fixups like:
> 
> diff --git a/drivers/md/dm.c b/drivers/md/dm.c
> index 8dcabf84d866..a35e60e62440 100644
> --- a/drivers/md/dm.c
> +++ b/drivers/md/dm.c
> @@ -2126,12 +2126,12 @@ static struct mapped_device *alloc_dev(int minor)
>                  md->dax_dev = alloc_dax(md, &dm_dax_ops);
>                  if (IS_ERR(md->dax_dev)) {
>                          md->dax_dev = NULL;
> -                       goto bad;
> +               } else {
> +                       set_dax_nocache(md->dax_dev);
> +                       set_dax_nomc(md->dax_dev);
> +                       if (dax_add_host(md->dax_dev, md->disk))
> +                               goto bad;
>                  }
> -               set_dax_nocache(md->dax_dev);
> -               set_dax_nomc(md->dax_dev);
> -               if (dax_add_host(md->dax_dev, md->disk))
> -                       goto bad;
>          }
>   
>          format_dev_t(md->name, MKDEV(_major, minor));
> 
> ...to make it not fatal to fail to register the dax_dev.

I've had a quick look at other users of alloc_dax() and
alloc_dax_region(), and so far I figure that all of those
really want to bail out on alloc_dax failure. Is dm.c the
only special-case we need to fix to make it non-fatal ?

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
https://www.efficios.com