Skip to contentSkip to conversation

bte_copy of BTE_MAX_XFER trips BUG_ON.

PATCH New Robin Holt 1 participant 2 messages Feb 4, 2009

lore
Share & tools
lore

Robin Holt

author
#0 lore raw

PATCH 1/1 bte_copy of BTE_MAX_XFER trips BUG_ON.

BTE_MAX_XFER is wrong. It is one greater than the number of cache lines the BTE is actually able to transfer. If you request a transfer of exactly BTE_MAX_XFER size, you trip a very cryptic BUG_ON() which should certainly be made more clear.

This patch fixes that constant and also cleans up the BUG_ON()s in arch/ia64/sn/kernel/bte.c to test one condition per line.

---

 arch/ia64/include/asm/sn/bte.h |    2 +-
 arch/ia64/sn/kernel/bte.c      |    7 ++++---
 2 files changed, 5 insertions(+), 4 deletions(-)

Index: bte_max_xfer_fix/arch/ia64/include/asm/sn/bte.h
=================================--- bte_max_xfer_fix.orig/arch/ia64/include/asm/sn/bte.h 2009-02-03 18:07:26.000000000 -0600
+++ bte_max_xfer_fix/arch/ia64/include/asm/sn/bte.h 2009-02-03 18:09:38.749729353 -0600
@@ -39,7 +39,7 @@
/* BTE status register only supports 16 bits for length field */ #define BTE_LEN_BITS (16)
#define BTE_LEN_MASK ((1 << BTE_LEN_BITS) - 1)
-#define BTE_MAX_XFER ((1 << BTE_LEN_BITS) * L1_CACHE_BYTES) +#define BTE_MAX_XFER (BTE_LEN_MASK << L1_CACHE_SHIFT)

/* Define hardware */
Index: bte_max_xfer_fix/arch/ia64/sn/kernel/bte.c
=================================--- bte_max_xfer_fix.orig/arch/ia64/sn/kernel/bte.c 2009-02-03 18:07:26.000000000 -0600
+++ bte_max_xfer_fix/arch/ia64/sn/kernel/bte.c 2009-02-03 18:15:42.854040475 -0600
@@ -97,9 +97,10 @@ bte_result_t bte_copy(u64 src, u64 dest,
return BTE_SUCCESS;
}

- BUG_ON((len & L1_CACHE_MASK) ||
- (src & L1_CACHE_MASK) || (dest & L1_CACHE_MASK));
- BUG_ON(!(len < ((BTE_LEN_MASK + 1) << L1_CACHE_SHIFT)));
+ BUG_ON(len & L1_CACHE_MASK);
+ BUG_ON(src & L1_CACHE_MASK);
+ BUG_ON(dest & L1_CACHE_MASK);
+ BUG_ON(len > BTE_MAX_XFER);

/*
* Start with interface corresponding to cpu number

author
#1 lore raw

PATCH 1/1 bte_copy of BTE_MAX_XFER trips BUG_ON -V2

BTE_MAX_XFER is wrong. It is one greater than the number of cache lines the BTE is actually able to transfer. If you request a transfer of exactly BTE_MAX_XFER size, you trip a very cryptic BUG_ON() which should certainly be made more clear.

This patch fixes that constant and also cleans up the BUG_ON()s in arch/ia64/sn/kernel/bte.c to test one condition per line.

---

Changes since -V1:
Base the BTE_LEN_MASK upon a 1UL instead of 1 for correctness.

 arch/ia64/include/asm/sn/bte.h |    4 ++--
 arch/ia64/sn/kernel/bte.c      |    7 ++++---
 2 files changed, 6 insertions(+), 5 deletions(-)

Index: bte_fixup/arch/ia64/include/asm/sn/bte.h
=================================--- bte_fixup.orig/arch/ia64/include/asm/sn/bte.h 2009-02-04 10:37:29.594750841 -0600
+++ bte_fixup/arch/ia64/include/asm/sn/bte.h 2009-02-04 10:40:32.495791608 -0600
@@ -38,8 +38,8 @@

/* BTE status register only supports 16 bits for length field */ #define BTE_LEN_BITS (16)
-#define BTE_LEN_MASK ((1 << BTE_LEN_BITS) - 1) -#define BTE_MAX_XFER ((1 << BTE_LEN_BITS) * L1_CACHE_BYTES) +#define BTE_LEN_MASK ((1UL << BTE_LEN_BITS) - 1) +#define BTE_MAX_XFER (BTE_LEN_MASK << L1_CACHE_SHIFT)

/* Define hardware */
Index: bte_fixup/arch/ia64/sn/kernel/bte.c
=================================--- bte_fixup.orig/arch/ia64/sn/kernel/bte.c 2009-02-04 10:37:30.026753839 -0600
+++ bte_fixup/arch/ia64/sn/kernel/bte.c 2009-02-04 10:40:11.711701203 -0600
@@ -97,9 +97,10 @@ bte_result_t bte_copy(u64 src, u64 dest,
return BTE_SUCCESS;
}

- BUG_ON((len & L1_CACHE_MASK) ||
- (src & L1_CACHE_MASK) || (dest & L1_CACHE_MASK));
- BUG_ON(!(len < ((BTE_LEN_MASK + 1) << L1_CACHE_SHIFT)));
+ BUG_ON(len & L1_CACHE_MASK);
+ BUG_ON(src & L1_CACHE_MASK);
+ BUG_ON(dest & L1_CACHE_MASK);
+ BUG_ON(len > BTE_MAX_XFER);

/*
* Start with interface corresponding to cpu number