All of lore.kernel.org
 help / color / mirror / Atom feed
From: Raj Vishwanathan <raj.vishwanathan@gmail.com>
To: opensbi@lists.infradead.org
Subject: [PATCH v2] We add a new configuration CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT as an int
Date: Sun, 19 Jan 2025 12:04:47 -0800	[thread overview]
Message-ID: <20250119200448.662074-1-Raj.Vishwanathan@gmail.com> (raw)
In-Reply-To: <20250110003451.2800-1-cfu@mips.com>

If it is set to 0 or not defined, we will continue with the previous
defintion of allocating pointer size chunks. Otherwise we will use the
chacheline size of 64.
We have the option of increasing the scratch allocation alignment to 64 bytes
as the cache line size, to avoid two atomic variables from
the same cache line that may cause livelock on some platforms.
---
 lib/sbi/Kconfig       |  8 ++++++++
 lib/sbi/sbi_scratch.c | 18 ++++++++++++++++--
 2 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/lib/sbi/Kconfig b/lib/sbi/Kconfig
index c6cc04b..1ffa170 100644
--- a/lib/sbi/Kconfig
+++ b/lib/sbi/Kconfig
@@ -69,4 +69,12 @@ config SBI_ECALL_SSE
 config SBI_ECALL_MPXY
 	bool "MPXY extension"
 	default y
+config SBI_SCRATCH_ALLOC_ALIGNMENT
+	int  "Scratch allocation alignment"
+	default 0
+	help
+	  We provide the option for allocation alignment to 64 bytes to avoid 
+	  livelock on certain platforms due to atomic variables from the same 
+	  cache line.  Leave it 0 for default behaviour.
+
 endmenu
diff --git a/lib/sbi/sbi_scratch.c b/lib/sbi/sbi_scratch.c
index ccbbc68..88ea3c7 100644
--- a/lib/sbi/sbi_scratch.c
+++ b/lib/sbi/sbi_scratch.c
@@ -14,6 +14,13 @@
 #include <sbi/sbi_scratch.h>
 #include <sbi/sbi_string.h>
 
+#if !defined(CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT) || (CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT==0)
+#define SCRATCH_ALLOC_ALIGNMENT __SIZEOF_POINTER__
+#else
+#define SCRATCH_ALLOC_ALIGNMENT CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT
+#endif
+
+
 u32 last_hartindex_having_scratch = 0;
 u32 hartindex_to_hartid_table[SBI_HARTMASK_MAX_BITS + 1] = { -1U };
 struct sbi_scratch *hartindex_to_scratch_table[SBI_HARTMASK_MAX_BITS + 1] = { 0 };
@@ -70,8 +77,15 @@ unsigned long sbi_scratch_alloc_offset(unsigned long size)
 	if (!size)
 		return 0;
 
-	size += __SIZEOF_POINTER__ - 1;
-	size &= ~((unsigned long)__SIZEOF_POINTER__ - 1);
+	/*
+	 * We let the allocation align to the cache line size, so
+	 * certain platforms due to atomic variables from the same cache line.
+	 * This ensures that the LR/SC variables are in a separate cache line 
+	 * to avoid live lock.
+	 */
+
+	size += SCRATCH_ALLOC_ALIGNMENT- 1;
+	size &= ~((unsigned long)SCRATCH_ALLOC_ALIGNMENT - 1);
 
 	spin_lock(&extra_lock);
 
-- 
2.43.0



  parent reply	other threads:[~2025-01-19 20:04 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-10  0:34 [PATCH] Align the scratch allocation to 64 bytes Chao-ying Fu
2025-01-10  4:15 ` Xiang W
2025-01-10  2:36   ` Chao-ying Fu
2025-01-19 20:04 ` Raj Vishwanathan [this message]
2025-01-20  3:58   ` [PATCH v2] We add a new configuration CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT as an int Xiang W
2025-01-20 18:56     ` [EXTERNAL]Re: " Raj Vishwanathan
2025-01-21  1:46   ` [PATCH v3] New configuration CONFIG_SBI_SCRATCH_ALLOC_ALIGNMENT Raj Vishwanathan
2025-01-21  3:21     ` Xiang W
2025-01-21 22:55   ` Raj Vishwanathan
2025-01-24 23:26   ` Raj Vishwanathan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250119200448.662074-1-Raj.Vishwanathan@gmail.com \
    --to=raj.vishwanathan@gmail.com \
    --cc=opensbi@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.