linux-kernel - Re: [PATCH] media: solo6x10: replace max(a, min(b, c)) by clamp(b, a, c)

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 for Android: free password hash cracker in your pocket

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

Date: Sun, 14 Jan 2024 12:04:01 +0100
From: Hans Verkuil <hverkuil@...all.nl>
To: Aurelien Jarno <aurelien@...el32.net>, linux-kernel@...r.kernel.org,
 Bluecherry Maintainers <maintainers@...echerrydvr.com>,
 Anton Sviridenko <anton@...p.bluecherry.net>,
 Andrey Utkin <andrey_utkin@...tmail.com>, Ismael Luceno
 <ismael@...ev.co.uk>, Mauro Carvalho Chehab <mchehab@...nel.org>,
 "open list:SOFTLOGIC 6x10 MPEG CODEC" <linux-media@...r.kernel.org>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
 Andy Shevchenko' <andriy.shevchenko@...ux.intel.com>,
 Andrew Morton' <akpm@...ux-foundation.org>,
 Matthew Wilcox <willy@...radead.org>, Christoph Hellwig'
 <hch@...radead.org>, "Jason A . Donenfeld" <Jason@...c4.com>,
 Jiri Slaby <jirislaby@...il.com>, stable@...r.kernel.org,
 David Laight <David.Laight@...LAB.COM>
Subject: Re: [PATCH] media: solo6x10: replace max(a, min(b, c)) by clamp(b, a,
 c)

Hi Aurelien,

On 13/01/2024 19:33, Aurelien Jarno wrote:
> This patch replaces max(a, min(b, c)) by clamp(b, a, c) in the solo6x10
> driver. This improves the readability and more importantly, for the
> solo6x10-p2m.c file, this reduces on my system (x86-64, gcc 13):
> - the preprocessed size from 121 MiB to 4.5 MiB;
> - the build CPU time from 46.8 s to 1.6 s;
> - the build memory from 2786 MiB to 98MiB.

Thank you for this patch.

Reviewed-by: Hans Verkuil <hverkuil-cisco@...all.nl>

I'll pick this up as a fix for v6.8.

Linus, if you prefer to pick this up directly, then that's fine as well.

I was aware for some time that something had changed elsewhere that caused
the solo driver compilation to take a lot of memory, but I didn't have the
time yet to investigate it. I now see that this was discussed in the thread
from the Closes tag. Note that it would have been nice if that was CCed to
linux-media, since it concerned a media driver.

The use of clamp() is definitely cleaner code, but it is rather insane
that max(a, min(b, c)) takes 2786 MiB over 98 MiB for clamp(). The driver
code looks perfectly innocuous, and you won't see the problem unless you compile
on a system or a VM with a lot less memory. (I only noticed it myself when
compiling in a VM.

Perhaps sparse or smatch should be extended with a check for nested min/max
and warn for that with a recommendation to use clamp instead? Although right
now these tools skip the solo driver since it takes too much memory :-)

I even found a min(max(max())) case in drivers/net/wireless/ath/ath11k/mac.c:

        nss = min(nss, max(max(ath11k_mac_max_ht_nss(ht_mcs_mask),
                               ath11k_mac_max_vht_nss(vht_mcs_mask)),
                           ath11k_mac_max_he_nss(he_mcs_mask)));

I wonder how that expands in cpp.

So while this patch is a nice cleanup, the underlying problem remains.

Regards,

	Hans

> 
> In fine, this allows this relatively simple C file to be built on a
> 32-bit system.
> 
> Reported-by: Jiri Slaby <jirislaby@...il.com>
> Closes: https://lore.kernel.org/lkml/18c6df0d-45ed-450c-9eda-95160a2bbb8e@gmail.com/
> Cc: stable@...r.kernel.org # v6.7+
> Suggested-by: David Laight <David.Laight@...LAB.COM>
> Signed-off-by: Aurelien Jarno <aurelien@...el32.net>
> ---
>  drivers/media/pci/solo6x10/solo6x10-offsets.h | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/media/pci/solo6x10/solo6x10-offsets.h b/drivers/media/pci/solo6x10/solo6x10-offsets.h
> index f414ee1316f2..fdbb817e6360 100644
> --- a/drivers/media/pci/solo6x10/solo6x10-offsets.h
> +++ b/drivers/media/pci/solo6x10/solo6x10-offsets.h
> @@ -57,16 +57,16 @@
>  #define SOLO_MP4E_EXT_ADDR(__solo) \
>  	(SOLO_EREF_EXT_ADDR(__solo) + SOLO_EREF_EXT_AREA(__solo))
>  #define SOLO_MP4E_EXT_SIZE(__solo) \
> -	max((__solo->nr_chans * 0x00080000),				\
> -	    min(((__solo->sdram_size - SOLO_MP4E_EXT_ADDR(__solo)) -	\
> -		 __SOLO_JPEG_MIN_SIZE(__solo)), 0x00ff0000))
> +	clamp(__solo->sdram_size - SOLO_MP4E_EXT_ADDR(__solo) -	\
> +	      __SOLO_JPEG_MIN_SIZE(__solo),			\
> +	      __solo->nr_chans * 0x00080000, 0x00ff0000)
>  
>  #define __SOLO_JPEG_MIN_SIZE(__solo)		(__solo->nr_chans * 0x00080000)
>  #define SOLO_JPEG_EXT_ADDR(__solo) \
>  		(SOLO_MP4E_EXT_ADDR(__solo) + SOLO_MP4E_EXT_SIZE(__solo))
>  #define SOLO_JPEG_EXT_SIZE(__solo) \
> -	max(__SOLO_JPEG_MIN_SIZE(__solo),				\
> -	    min((__solo->sdram_size - SOLO_JPEG_EXT_ADDR(__solo)), 0x00ff0000))
> +	clamp(__solo->sdram_size - SOLO_JPEG_EXT_ADDR(__solo),	\
> +	      __SOLO_JPEG_MIN_SIZE(__solo), 0x00ff0000)
>  
>  #define SOLO_SDRAM_END(__solo) \
>  	(SOLO_JPEG_EXT_ADDR(__solo) + SOLO_JPEG_EXT_SIZE(__solo))