Re: [PATCH] PR target/111778 - Fix undefined shifts in PowerPC compiler

[Jiufu Guo <guojiufu@linux.ibm.com>]

Hi,

Thanks for your quick fix!

Michael Meissner <meissner@linux.ibm.com> writes:

> I was building a cross compiler to PowerPC on my x86_86 workstation with the
> latest version of GCC on October 11th.  I could not build the compiler on the
> x86_64 system as it died in building libgcc.  I looked into it, and I
> discovered the compiler was recursing until it ran out of stack space.  If I
> build a native compiler with the same sources on a PowerPC system, it builds
> fine.
>
> I traced this down to a change made around October 10th:
>
> | commit 8f1a70a4fbcc6441c70da60d4ef6db1e5635e18a (HEAD)
> | Author: Jiufu Guo <guojiufu@linux.ibm.com>
> | Date:   Tue Jan 10 20:52:33 2023 +0800
> |
> |   rs6000: build constant via li/lis;rldicl/rldicr
> |
> |   If a constant is possible left/right cleaned on a rotated value from
> |   a negative value of "li/lis".  Then, using "li/lis ; rldicl/rldicr"
> |   to build the constant.
>
> The code was doing a -1 << 64 which is undefined behavior because different
> machines produce different results.  On the x86_64 system, (-1 << 64) produces
> -1 while on a PowerPC 64-bit system, (-1 << 64) produces 0.  The x86_64 then
> recurses until the stack runs out of space.
>
> If I apply this patch, the compiler builds fine on both x86_64 as a PowerPC
> crosss compiler and on a native PowerPC system.
>
> Can I check this into the master branch to fix the problem?
>
> 2023-10-12  Michael Meissner  <meissner@linux.ibm.com>
>
> gcc/
>
> 	PR target/111778
> 	* config/rs6000/rs6000.cc (can_be_built_by_li_lis_and_rldicl): Protect
> 	code from shifts that are undefined.
> 	(can_be_built_by_li_lis_and_rldicr): Likewise.
> 	(can_be_built_by_li_and_rldic): Protect code from shifts that
> 	undefined.  Also replace uses of 1ULL with HOST_WIDE_INT_1U.
>
> ---
>  gcc/config/rs6000/rs6000.cc | 29 ++++++++++++++++++++++++++---
>  1 file changed, 26 insertions(+), 3 deletions(-)
>
> diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc
> index 2828f01413c..cc24dd5301e 100644
> --- a/gcc/config/rs6000/rs6000.cc
> +++ b/gcc/config/rs6000/rs6000.cc
> @@ -10370,6 +10370,11 @@ can_be_built_by_li_lis_and_rldicl (HOST_WIDE_INT c, int *shift,
>    /* Leading zeros may be cleaned by rldicl with a mask.  Change leading zeros
>       to ones and then recheck it.  */
>    int lz = clz_hwi (c);
> +
> +  /* If lz == 0, the left shift is undefined.  */
> +  if (!lz)
> +    return false;
> +
Thanks! This should be checked.
If "lz" is zero, it means for input "C", there is no leading
zeros which are cleanded by "rldicl". And then no future analyzing is
needed.

>    HOST_WIDE_INT unmask_c
>      = c | (HOST_WIDE_INT_M1U << (HOST_BITS_PER_WIDE_INT - lz));
>    int n;
> @@ -10398,6 +10403,11 @@ can_be_built_by_li_lis_and_rldicr (HOST_WIDE_INT c, int *shift,
>    /* Tailing zeros may be cleaned by rldicr with a mask.  Change tailing zeros
>       to ones and then recheck it.  */
>    int tz = ctz_hwi (c);
> +
> +  /* If tz == HOST_BITS_PER_WIDE_INT, the left shift is undefined.  */
> +  if (tz >= HOST_BITS_PER_WIDE_INT)
> +    return false;
> +
This is correct in theory and could make sure "tz" is ok.
Just one minor thing:
"ctz_hwi" would not return value greater than HOST_BITS_PER_WIDE_INT
other than 0, right? 

>    HOST_WIDE_INT unmask_c = c | ((HOST_WIDE_INT_1U << tz) - 1);
>    int n;
>    if (can_be_rotated_to_lowbits (~unmask_c, 15, &n)
> @@ -10428,8 +10438,15 @@ can_be_built_by_li_and_rldic (HOST_WIDE_INT c, int *shift, HOST_WIDE_INT *mask)
>       right bits are shifted as 0's, and left 1's(and x's) are cleaned.  */
>    int tz = ctz_hwi (c);
>    int lz = clz_hwi (c);
> +
> +  /* If lz == HOST_BITS_PER_WIDE_INT, the left shift is undefined.  */
> +  if (lz >= HOST_BITS_PER_WIDE_INT)
> +    return false;
> +
This maybe similar.

>    int middle_ones = clz_hwi (~(c << lz));
> -  if (tz + lz + middle_ones >= ones)
> +  if (tz + lz + middle_ones >= ones
> +      && (tz - lz) < HOST_BITS_PER_WIDE_INT
> +      && tz < HOST_BITS_PER_WIDE_INT)
>      {
>        *mask = ((1LL << (HOST_BITS_PER_WIDE_INT - tz - lz)) - 1LL) << tz;
>        *shift = tz;
> @@ -10440,7 +10457,8 @@ can_be_built_by_li_and_rldic (HOST_WIDE_INT c, int *shift, HOST_WIDE_INT *mask)
>    int leading_ones = clz_hwi (~c);
>    int tailing_ones = ctz_hwi (~c);
>    int middle_zeros = ctz_hwi (c >> tailing_ones);
> -  if (leading_ones + tailing_ones + middle_zeros >= ones)
> +  if (leading_ones + tailing_ones + middle_zeros >= ones
> +      && middle_zeros < HOST_BITS_PER_WIDE_INT)
Thanks.
>      {
>        *mask = ~(((1ULL << middle_zeros) - 1ULL) << tailing_ones);
>        *shift = tailing_ones + middle_zeros;
> @@ -10450,10 +10468,15 @@ can_be_built_by_li_and_rldic (HOST_WIDE_INT c, int *shift, HOST_WIDE_INT *mask)
>    /* xx1..1xx: --> xx0..01..1xx: some 1's(following x's) are cleaned. */
>    /* Get the position for the first bit of successive 1.
>       The 24th bit would be in successive 0 or 1.  */
> -  HOST_WIDE_INT low_mask = (1LL << 24) - 1LL;
> +  HOST_WIDE_INT low_mask = (HOST_WIDE_INT_1U << 24) - HOST_WIDE_INT_1U;
Yes.
>    int pos_first_1 = ((c & (low_mask + 1)) == 0)
>  		      ? clz_hwi (c & low_mask)
>  		      : HOST_BITS_PER_WIDE_INT - ctz_hwi (~(c | low_mask));
> +
> +  /* Make sure the left and right shifts are defined.  */
> +  if (!IN_RANGE (pos_first_1, 1, HOST_BITS_PER_WIDE_INT-1))
> +    return false;
> +
Yes, this change would be safer.

Thanks again for the enhancement!

BR,
Jeff (Jiufu Guo)
>    middle_ones = clz_hwi (~c << pos_first_1);
>    middle_zeros = ctz_hwi (c >> (HOST_BITS_PER_WIDE_INT - pos_first_1));
>    if (pos_first_1 < HOST_BITS_PER_WIDE_INT
> -- 
> 2.41.0