[<prev] [next>] [day] [month] [year] [list]
Message-Id: <201105022254.p42MsQlr030357@farm-0023.internal.tilera.com>
Date: Mon, 2 May 2011 15:54:32 -0400
From: Chris Metcalf <cmetcalf@...era.com>
To: linux-kernel@...r.kernel.org
Subject: [PATCH] arch/tile: disable GX prefetcher during cache flush
Otherwise, it's possible to end up with the prefetcher pulling
data into cache that the code believes has been flushed.
Signed-off-by: Chris Metcalf <cmetcalf@...era.com>
---
arch/tile/lib/cacheflush.c | 18 ++++++++++++++++++
1 files changed, 18 insertions(+), 0 deletions(-)
diff --git a/arch/tile/lib/cacheflush.c b/arch/tile/lib/cacheflush.c
index 35c1d8c..8928aac 100644
--- a/arch/tile/lib/cacheflush.c
+++ b/arch/tile/lib/cacheflush.c
@@ -15,6 +15,7 @@
#include <asm/page.h>
#include <asm/cacheflush.h>
#include <arch/icache.h>
+#include <arch/spr_def.h>
void __flush_icache_range(unsigned long start, unsigned long end)
@@ -39,6 +40,18 @@ void finv_buffer_remote(void *buffer, size_t size, int hfh)
char *p, *base;
size_t step_size, load_count;
const unsigned long STRIPE_WIDTH = 8192;
+#ifdef __tilegx__
+ /*
+ * On TILE-Gx, we must disable the dstream prefetcher before doing
+ * a cache flush; otherwise, we could end up with data in the cache
+ * that we don't want there. Note that normally we'd do an mf
+ * after the SPR write to disabling the prefetcher, but we do one
+ * below, before any further loads, so there's no need to do it
+ * here.
+ */
+ uint_reg_t old_dstream_pf = __insn_mfspr(SPR_DSTREAM_PF);
+ __insn_mtspr(SPR_DSTREAM_PF, 0);
+#endif
/*
* Flush and invalidate the buffer out of the local L1/L2
@@ -122,4 +135,9 @@ void finv_buffer_remote(void *buffer, size_t size, int hfh)
/* Wait for the load+inv's (and thus finvs) to have completed. */
__insn_mf();
+
+#ifdef __tilegx__
+ /* Reenable the prefetcher. */
+ __insn_mtspr(SPR_DSTREAM_PF, old_dstream_pf);
+#endif
}
--
1.6.5.2
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists