public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] EDAC/i10nm: don't fail probing if ADXL is missing
@ 2026-04-13 23:13 Vasily Khoruzhick
  2026-04-14 14:29 ` Zhuo, Qiuxu
  0 siblings, 1 reply; 2+ messages in thread
From: Vasily Khoruzhick @ 2026-04-13 23:13 UTC (permalink / raw)
  To: Tony Luck, Borislav Petkov, linux-edac, linux-kernel
  Cc: Vasily Khoruzhick, stable

From: Vasily Khoruzhick <vasilykh@arista.com>

ADXL is not present in Coreboot- or Slimbootloader-based BIOSes and as
result, the driver fails to probe there.

i10nm does not require ADXL for decoding errors since commit
2738c69a8813 ("EDAC/i10nm: Add driver decoder for Ice Lake and Tremont CPUs"),
so we can just switch to driver decoding when it's not present.

Cc: stable@vger.kernel.org # v6.1+
Signed-off-by: Vasily Khoruzhick <vasilykh@arista.com>
---
 drivers/edac/i10nm_base.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/drivers/edac/i10nm_base.c b/drivers/edac/i10nm_base.c
index 89b3e8cc38b1..69a4a255e4c8 100644
--- a/drivers/edac/i10nm_base.c
+++ b/drivers/edac/i10nm_base.c
@@ -79,6 +79,7 @@ static struct res_config *res_cfg;
 static int retry_rd_err_log;
 static int decoding_via_mca;
 static bool mem_cfg_2lm;
+static bool no_adxl;
 
 static struct reg_rrl icx_reg_rrl_ddr = {
 	.set_num = 2,
@@ -1208,8 +1209,14 @@ static int __init i10nm_init(void)
 	}
 
 	rc = skx_adxl_get();
-	if (rc)
-		goto fail;
+	if (rc) {
+		/* Decoding errors via MCA banks for 2LM isn't supported yet */
+		if (rc != -ENODEV || mem_cfg_2lm)
+			goto fail;
+		i10nm_printk(KERN_INFO, "ADXL not found, falling back to MCA-based decoding.\n");
+		no_adxl = true;
+		decoding_via_mca = true;
+	}
 
 	opstate_init();
 	mce_register_decode_chain(&i10nm_mce_dec);
@@ -1243,7 +1250,8 @@ static void __exit i10nm_exit(void)
 
 	skx_teardown_debug();
 	mce_unregister_decode_chain(&i10nm_mce_dec);
-	skx_adxl_put();
+	if (!no_adxl)
+		skx_adxl_put();
 	skx_remove();
 }
 
-- 
2.53.0


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* RE: [PATCH] EDAC/i10nm: don't fail probing if ADXL is missing
  2026-04-13 23:13 [PATCH] EDAC/i10nm: don't fail probing if ADXL is missing Vasily Khoruzhick
@ 2026-04-14 14:29 ` Zhuo, Qiuxu
  0 siblings, 0 replies; 2+ messages in thread
From: Zhuo, Qiuxu @ 2026-04-14 14:29 UTC (permalink / raw)
  To: Vasily Khoruzhick, Luck, Tony, Borislav Petkov,
	linux-edac@vger.kernel.org, linux-kernel@vger.kernel.org
  Cc: Vasiliy Khoruzhick, stable@vger.kernel.org

> From: Vasily Khoruzhick <anarsoul@gmail.com>
> Sent: Tuesday, April 14, 2026 7:14 AM
> To: Luck, Tony <tony.luck@intel.com>; Borislav Petkov <bp@alien8.de>; linux-
> edac@vger.kernel.org; linux-kernel@vger.kernel.org
> Cc: Vasiliy Khoruzhick <vasilykh@arista.com>; stable@vger.kernel.org
> Subject: [PATCH] EDAC/i10nm: don't fail probing if ADXL is missing

s/don't/Don't/
 
> From: Vasily Khoruzhick <vasilykh@arista.com>
> 
> ADXL is not present in Coreboot- or Slimbootloader-based BIOSes and as
> result, the driver fails to probe there.
> 
> i10nm does not require ADXL for decoding errors since commit

ADXL provides more detailed decoded results than the driver decoder. 
It's the preferred method when decoding performance is not a concern. 

> 2738c69a8813 ("EDAC/i10nm: Add driver decoder for Ice Lake and Tremont
> CPUs"), so we can just switch to driver decoding when it's not present.
>

You could phrase this commit message like:

    Since commit 2738c69a8813 ("EDAC/i10nm: Add driver decoder for Ice Lake and Tremont CPUs"),
    i10nm_edac supports driver decoder. Switch to driver decoding when ADXL is not present. 

Btw, please drop the personal pronoun 'we' and use the imperative tone.

> Cc: stable@vger.kernel.org # v6.1+
> Signed-off-by: Vasily Khoruzhick <vasilykh@arista.com>

Other than the comments above:

   Reviewed-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>

[...]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-04-14 14:29 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-04-13 23:13 [PATCH] EDAC/i10nm: don't fail probing if ADXL is missing Vasily Khoruzhick
2026-04-14 14:29 ` Zhuo, Qiuxu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox