| Message ID | 20230227214524.914050-2-yury.norov@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/2] lib/test_bitmap: increment failure counter properly | expand |
From: Yury Norov <yury.norov@gmail.com> Date: Mon, 27 Feb 2023 13:45:24 -0800 > bitmap_{from,to}_arr64() optimization is overly optimistic on 32-bit LE > architectures when it's wired to bitmap_copy_clear_tail(). > > bitmap_copy_clear_tail() takes care of unused bits in the bitmap up to > the next word boundary. But on 32-bit machines when copying bits from > bitmap to array of 64-bit words, it's expected that the unused part of > a recipient array must be cleared up to 64-bit boundary, so the last 4 > bytes may stay untouched when nbits % 64 <= 32. > > While the copying part of the optimization works correct, that clear-tail > trick makes corresponding tests reasonably fail: > > test_bitmap: bitmap_to_arr64(nbits == 1): tail is not safely cleared: 0xa5a5a5a500000001 (must be 0x0000000000000001) > > Fix it by removing bitmap_{from,to}_arr64() optimization for 32-bit LE > arches. > > Reported-by: Guenter Roeck <linux@roeck-us.net> > Link: https://lore.kernel.org/lkml/20230225184702.GA3587246@roeck-us.net/ > Fixes: 0a97953fd221 ("lib: add bitmap_{from,to}_arr64") > Signed-off-by: Yury Norov <yury.norov@gmail.com> > Tested-by: Guenter Roeck <linux@roeck-us.net> > --- > include/linux/bitmap.h | 8 +++----- > lib/bitmap.c | 2 +- > 2 files changed, 4 insertions(+), 6 deletions(-) > > diff --git a/include/linux/bitmap.h b/include/linux/bitmap.h > index 40e53a2ecc0d..7d4c90eb3df4 100644 > --- a/include/linux/bitmap.h > +++ b/include/linux/bitmap.h > @@ -302,12 +302,10 @@ void bitmap_to_arr32(u32 *buf, const unsigned long *bitmap, > #endif > > /* > - * On 64-bit systems bitmaps are represented as u64 arrays internally. On LE32 > - * machines the order of hi and lo parts of numbers match the bitmap structure. > - * In both cases conversion is not needed when copying data from/to arrays of > - * u64. > + * On 64-bit systems bitmaps are represented as u64 arrays internally. So, > + * the conversion is not needed when copying data from/to arrays of u64. > */ > -#if (BITS_PER_LONG == 32) && defined(__BIG_ENDIAN) > +#if BITS_PER_LONG == 32 > void bitmap_from_arr64(unsigned long *bitmap, const u64 *buf, unsigned int nbits); > void bitmap_to_arr64(u64 *buf, const unsigned long *bitmap, unsigned int nbits); > #else > diff --git a/lib/bitmap.c b/lib/bitmap.c > index 1c81413c51f8..ddb31015e38a 100644 > --- a/lib/bitmap.c > +++ b/lib/bitmap.c > @@ -1495,7 +1495,7 @@ void bitmap_to_arr32(u32 *buf, const unsigned long *bitmap, unsigned int nbits) > EXPORT_SYMBOL(bitmap_to_arr32); > #endif > > -#if (BITS_PER_LONG == 32) && defined(__BIG_ENDIAN) > +#if BITS_PER_LONG == 32 > /** > * bitmap_from_arr64 - copy the contents of u64 array of bits to bitmap > * @bitmap: array of unsigned longs, the destination bitmap For the series: Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com> Thanks, Olek
diff --git a/include/linux/bitmap.h b/include/linux/bitmap.h index 40e53a2ecc0d..7d4c90eb3df4 100644 --- a/include/linux/bitmap.h +++ b/include/linux/bitmap.h @@ -302,12 +302,10 @@ void bitmap_to_arr32(u32 *buf, const unsigned long *bitmap, #endif /* - * On 64-bit systems bitmaps are represented as u64 arrays internally. On LE32 - * machines the order of hi and lo parts of numbers match the bitmap structure. - * In both cases conversion is not needed when copying data from/to arrays of - * u64. + * On 64-bit systems bitmaps are represented as u64 arrays internally. So, + * the conversion is not needed when copying data from/to arrays of u64. */ -#if (BITS_PER_LONG == 32) && defined(__BIG_ENDIAN) +#if BITS_PER_LONG == 32 void bitmap_from_arr64(unsigned long *bitmap, const u64 *buf, unsigned int nbits); void bitmap_to_arr64(u64 *buf, const unsigned long *bitmap, unsigned int nbits); #else diff --git a/lib/bitmap.c b/lib/bitmap.c index 1c81413c51f8..ddb31015e38a 100644 --- a/lib/bitmap.c +++ b/lib/bitmap.c @@ -1495,7 +1495,7 @@ void bitmap_to_arr32(u32 *buf, const unsigned long *bitmap, unsigned int nbits) EXPORT_SYMBOL(bitmap_to_arr32); #endif -#if (BITS_PER_LONG == 32) && defined(__BIG_ENDIAN) +#if BITS_PER_LONG == 32 /** * bitmap_from_arr64 - copy the contents of u64 array of bits to bitmap * @bitmap: array of unsigned longs, the destination bitmap
bitmap_{from,to}_arr64() optimization is overly optimistic on 32-bit LE architectures when it's wired to bitmap_copy_clear_tail(). bitmap_copy_clear_tail() takes care of unused bits in the bitmap up to the next word boundary. But on 32-bit machines when copying bits from bitmap to array of 64-bit words, it's expected that the unused part of a recipient array must be cleared up to 64-bit boundary, so the last 4 bytes may stay untouched when nbits % 64 <= 32. While the copying part of the optimization works correct, that clear-tail trick makes corresponding tests reasonably fail: test_bitmap: bitmap_to_arr64(nbits == 1): tail is not safely cleared: 0xa5a5a5a500000001 (must be 0x0000000000000001) Fix it by removing bitmap_{from,to}_arr64() optimization for 32-bit LE arches. Reported-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/lkml/20230225184702.GA3587246@roeck-us.net/ Fixes: 0a97953fd221 ("lib: add bitmap_{from,to}_arr64") Signed-off-by: Yury Norov <yury.norov@gmail.com> Tested-by: Guenter Roeck <linux@roeck-us.net> --- include/linux/bitmap.h | 8 +++----- lib/bitmap.c | 2 +- 2 files changed, 4 insertions(+), 6 deletions(-)