[PATCH hotfix 6.11 v2 0/3] minmax: reduce egregious min/max macro expansion

[PATCH hotfix 6.11 v2 0/3] minmax: reduce egregious min/max macro expansion

[Lorenzo Stoakes]

Avoid nested min()/max() which results in egregious macro expansion.

This issue was introduced by commit 867046cc7027 ("minmax: relax check to
allow comparison between unsigned arguments and signed constants") [0].

Work has been done to address the issue of egregious min()/max() macro
expansion in commit 22f546873149 ("minmax: improve macro expansion and type
checking") and related, however it appears that some issues remain on more
tightly constrained systems.

Adjust a few known-bad cases of deeply nested macros to avoid doing so to
mitigate this. Porting the patch first proposed in [1] to Linus's tree.

Note that we use clamp_t() rather than clamp() when we otherwise could use
the latter because it isn't certain than the clamp() won't expand into a
more egregiously huge form, certainly prior to the recent fixes.

Running an allmodconfig build using the methodology described in [2] we
observe a 35 MiB reduction in generated code.

The difference is much more significant prior to recent minmax fixes which
were not backported. As per [1] prior these the reduction is more like 200
MiB.

This resolves an issue with slackware 15.0 32-bit compilation as reported
by Richard Narron.

Presumably the min/max fixups would be difficult to backport, this series
should be easier and fix's Richard's problem in 5.15.

[0]:https://lore.kernel.org/all/b97faef60ad24922b530241c5d7c933c@AcuMS.aculab.com/
[1]:https://lore.kernel.org/lkml/5882b96e-1287-4390-8174-3316d39038ef@lucifer.local/
[2]:https://lore.kernel.org/linux-mm/36aa2cad-1db1-4abf-8dd2-fb20484aabc3@lucifer.local/

v2:
* Split into separate patches to make backporting easier.
* Corrected type in sDIGIT_FITTING() argument.
* Added reviewed-by tags.

v1:
https://lore.kernel.org/all/20240911153457.1005227-1-lorenzo.stoakes@oracle.com/

Lorenzo Stoakes (3):
  minmax: reduce min/max macro expansion in mvpp2 driver
  minmax: reduce min/max macro expansion in skbuff
  minmax: reduce min/max macro expansion in atomisp driver

 drivers/net/ethernet/marvell/mvpp2/mvpp2.h    |  2 +-
 .../staging/media/atomisp/pci/sh_css_frac.h   | 26 ++++++++++++++-----
 include/linux/skbuff.h                        |  6 ++++-
 3 files changed, 25 insertions(+), 9 deletions(-)

--
2.46.0

[PATCH hotfix 6.11 v2 1/3] minmax: reduce min/max macro expansion in mvpp2 driver

[Lorenzo Stoakes]

Avoid unnecessary nested min()/max() which reults in egregious macro
expansion. Use clamp_t() as this introduces the least possible expansion.

Not doing so results in an impact on build times.

Cc: stable@vger.kernel.org
Fixes: 867046cc7027 ("minmax: relax check to allow comparison between unsigned arguments and signed constants")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 drivers/net/ethernet/marvell/mvpp2/mvpp2.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2.h b/drivers/net/ethernet/marvell/mvpp2/mvpp2.h
index e809f91c08fb..8b431f90efc3 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2.h
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2.h
@@ -23,7 +23,7 @@
 /* The PacketOffset field is measured in units of 32 bytes and is 3 bits wide,
  * so the maximum offset is 7 * 32 = 224
  */
-#define MVPP2_SKB_HEADROOM	min(max(XDP_PACKET_HEADROOM, NET_SKB_PAD), 224)
+#define MVPP2_SKB_HEADROOM	clamp_t(int, XDP_PACKET_HEADROOM, NET_SKB_PAD, 224)

 #define MVPP2_XDP_PASS		0
 #define MVPP2_XDP_DROPPED	BIT(0)
--
2.46.0

[PATCH hotfix 6.11 v2 2/3] minmax: reduce min/max macro expansion in skbuff

[Lorenzo Stoakes]

Avoid unnecessary nested min()/max() which reults in egregious macro
expansion.

The nesting occurs when NET_SKB_PAD is used in a min() or max() invocation,
for instance, various ethernet drivers wrap it in a max().

Not doing so results in an impact on build times.

Cc: stable@vger.kernel.org
Fixes: 867046cc7027 ("minmax: relax check to allow comparison between unsigned arguments and signed constants")
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 include/linux/skbuff.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 29c3ea5b6e93..d53b296df504 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -3164,7 +3164,11 @@ static inline int pskb_network_may_pull(struct sk_buff *skb, unsigned int len)
  * NET_IP_ALIGN(2) + ethernet_header(14) + IP_header(20/40) + ports(8)
  */
 #ifndef NET_SKB_PAD
-#define NET_SKB_PAD	max(32, L1_CACHE_BYTES)
+#if L1_CACHE_BYTES < 32
+#define NET_SKB_PAD	32
+#else
+#define NET_SKB_PAD	L1_CACHE_BYTES
+#endif
 #endif

 int ___pskb_trim(struct sk_buff *skb, unsigned int len);
--
2.46.0

[PATCH hotfix 6.11 v2 3/3] minmax: reduce min/max macro expansion in atomisp driver

[Lorenzo Stoakes]

Avoid unnecessary nested min()/max() which results in egregious macro
expansion. Use clamp_t() as this introduces the least possible expansion.

Not doing so results in an impact on build times.

This resolves an issue with slackware 15.0 32-bit compilation as reported
by Richard Narron.

Presumably the min/max fixups would be difficult to backport, this patch
should be easier and fix's Richard's problem in 5.15.

Reported-by: Richard Narron <richard@aaazen.com>
Closes: https://lore.kernel.org/all/4a5321bd-b1f-1832-f0c-cea8694dc5aa@aaazen.com/
Fixes: 867046cc7027 ("minmax: relax check to allow comparison between unsigned arguments and signed constants")
Cc: stable@vger.kernel.org
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 .../staging/media/atomisp/pci/sh_css_frac.h   | 26 ++++++++++++++-----
 1 file changed, 19 insertions(+), 7 deletions(-)

diff --git a/drivers/staging/media/atomisp/pci/sh_css_frac.h b/drivers/staging/media/atomisp/pci/sh_css_frac.h
index b90b5b330dfa..8ba65161f7a9 100644
--- a/drivers/staging/media/atomisp/pci/sh_css_frac.h
+++ b/drivers/staging/media/atomisp/pci/sh_css_frac.h
@@ -32,12 +32,24 @@
 #define uISP_VAL_MAX		      ((unsigned int)((1 << uISP_REG_BIT) - 1))

 /* a:fraction bits for 16bit precision, b:fraction bits for ISP precision */
-#define sDIGIT_FITTING(v, a, b) \
-	min_t(int, max_t(int, (((v) >> sSHIFT) >> max(sFRACTION_BITS_FITTING(a) - (b), 0)), \
-	  sISP_VAL_MIN), sISP_VAL_MAX)
-#define uDIGIT_FITTING(v, a, b) \
-	min((unsigned int)max((unsigned)(((v) >> uSHIFT) \
-	>> max((int)(uFRACTION_BITS_FITTING(a) - (b)), 0)), \
-	  uISP_VAL_MIN), uISP_VAL_MAX)
+static inline int sDIGIT_FITTING(int v, int a, int b)
+{
+	int fit_shift = sFRACTION_BITS_FITTING(a) - b;
+
+	v >>= sSHIFT;
+	v >>= fit_shift > 0 ? fit_shift : 0;
+
+	return clamp_t(int, v, sISP_VAL_MIN, sISP_VAL_MAX);
+}
+
+static inline unsigned int uDIGIT_FITTING(unsigned int v, int a, int b)
+{
+	int fit_shift = uFRACTION_BITS_FITTING(a) - b;
+
+	v >>= uSHIFT;
+	v >>= fit_shift > 0 ? fit_shift : 0;
+
+	return clamp_t(unsigned int, v, uISP_VAL_MIN, uISP_VAL_MAX);
+}

 #endif /* __SH_CSS_FRAC_H */
--
2.46.0

Re: [PATCH hotfix 6.11 v2 3/3] minmax: reduce min/max macro expansion in atomisp driver

[Hans de Goede]

Hi,

On 9/11/24 7:51 PM, Lorenzo Stoakes wrote:
> Avoid unnecessary nested min()/max() which results in egregious macro
> expansion. Use clamp_t() as this introduces the least possible expansion.
> 
> Not doing so results in an impact on build times.
> 
> This resolves an issue with slackware 15.0 32-bit compilation as reported
> by Richard Narron.
> 
> Presumably the min/max fixups would be difficult to backport, this patch
> should be easier and fix's Richard's problem in 5.15.
> 
> Reported-by: Richard Narron <richard@aaazen.com>
> Closes: https://lore.kernel.org/all/4a5321bd-b1f-1832-f0c-cea8694dc5aa@aaazen.com/
> Fixes: 867046cc7027 ("minmax: relax check to allow comparison between unsigned arguments and signed constants")
> Cc: stable@vger.kernel.org
> Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

Thanks, patch looks good to me:

Reviewed-by: Hans de Goede <hdegoede@redhat.com>

Regards,

Hans




> ---
>  .../staging/media/atomisp/pci/sh_css_frac.h   | 26 ++++++++++++++-----
>  1 file changed, 19 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/staging/media/atomisp/pci/sh_css_frac.h b/drivers/staging/media/atomisp/pci/sh_css_frac.h
> index b90b5b330dfa..8ba65161f7a9 100644
> --- a/drivers/staging/media/atomisp/pci/sh_css_frac.h
> +++ b/drivers/staging/media/atomisp/pci/sh_css_frac.h
> @@ -32,12 +32,24 @@
>  #define uISP_VAL_MAX		      ((unsigned int)((1 << uISP_REG_BIT) - 1))
> 
>  /* a:fraction bits for 16bit precision, b:fraction bits for ISP precision */
> -#define sDIGIT_FITTING(v, a, b) \
> -	min_t(int, max_t(int, (((v) >> sSHIFT) >> max(sFRACTION_BITS_FITTING(a) - (b), 0)), \
> -	  sISP_VAL_MIN), sISP_VAL_MAX)
> -#define uDIGIT_FITTING(v, a, b) \
> -	min((unsigned int)max((unsigned)(((v) >> uSHIFT) \
> -	>> max((int)(uFRACTION_BITS_FITTING(a) - (b)), 0)), \
> -	  uISP_VAL_MIN), uISP_VAL_MAX)
> +static inline int sDIGIT_FITTING(int v, int a, int b)
> +{
> +	int fit_shift = sFRACTION_BITS_FITTING(a) - b;
> +
> +	v >>= sSHIFT;
> +	v >>= fit_shift > 0 ? fit_shift : 0;
> +
> +	return clamp_t(int, v, sISP_VAL_MIN, sISP_VAL_MAX);
> +}
> +
> +static inline unsigned int uDIGIT_FITTING(unsigned int v, int a, int b)
> +{
> +	int fit_shift = uFRACTION_BITS_FITTING(a) - b;
> +
> +	v >>= uSHIFT;
> +	v >>= fit_shift > 0 ? fit_shift : 0;
> +
> +	return clamp_t(unsigned int, v, uISP_VAL_MIN, uISP_VAL_MAX);
> +}
> 
>  #endif /* __SH_CSS_FRAC_H */
> --
> 2.46.0
> 

Re: [PATCH hotfix 6.11 v2 3/3] minmax: reduce min/max macro expansion in atomisp driver

[Linus Torvalds]

On Wed, 11 Sept 2024 at 10:51, Lorenzo Stoakes
<lorenzo.stoakes@oracle.com> wrote:
>
> Avoid unnecessary nested min()/max() which results in egregious macro
> expansion. Use clamp_t() as this introduces the least possible expansion.

I took this (single) patch directly, since that's the one that
actually causes build problems in limited environments (admittedly not
in current git with the more invasive min/max cleanups, but in order
to be back-ported).

Plus it cleans up the code with more legible inline functions, rather
than just doing some minimal syntactic changes. I expanded on the
commit message to say that.

The two others I'll leave for now and see what maintainers of their
respective areas think.

            Linus