Re: [PATCH v2 1/4] dt-bindings: arm: ras: Introduce bindings for ARM RAS error sources
Umang Chheda <[email protected]>
| Newsgroups | org.kernel.vger.linux-edac,org.infradead.lists.linux-arm-kernel,org.kernel.vger.linux-acpi,org.kernel.vger.linux-arm-msm,org.kernel.vger.linux-devicetree,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <[email protected]> |
Hello Krzysztof, Trilok, On 7/28/2026 2:17 PM, Krzysztof Kozlowski wrote: > On Mon, Jul 27, 2026 at 01:51:08PM -0700, Trilok Soni wrote: >> On 7/20/2026 1:19 AM, Umang Chheda wrote: >>> new file mode 100644 >>> index 000000000000..c2f4e1f8243e >>> --- /dev/null >>> +++ b/include/dt-bindings/arm/arm-ras.h >>> @@ -0,0 +1,11 @@ >>> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ >>> + >>> +#ifndef _DT_BINDINGS_ARM_RAS_H >>> +#define _DT_BINDINGS_ARM_RAS_H >>> + >>> +/* arm,group-format - error record group register window page size */ >>> +#define ARM_RAS_GROUP_4K 0 /* 4 KiB, 1 ERRGSR */ >>> +#define ARM_RAS_GROUP_16K 1 /* 16 KiB, 4 ERRGSRs */ >>> +#define ARM_RAS_GROUP_64K 2 /* 64 KiB, 14 ERRGSRs */ >> >> Do you want me to have different DT for 4k and 16k and 64k? This is not a right approach. >> > > Is this page size or register window page size? If the latter, why would > device registers be a binding? Thanks for the review. To clarify both points: On "page size" vs "register window page size": The comment in arm-ras.h is misleading — I will fix it. It is neither a memory page size nor a kernel page granularity. The 4K/16K/64K values refer to the three fixed error record group register window layouts defined by the ARM RAS architecture specification. The naming comes from the ARM spec itself which uses these sizes to describe the three possible layouts of the MMIO window. I will update the comment to make this clear. On "why would device registers be a binding": This is a static hardware property — it describes which of the three fixed register layouts the hardware block implements, and it is not self-discoverable at runtime. There is no register within the window that encodes the layout. The driver needs it upfront to: - Determine the correct ioremap size - Know how many ERRGSR registers to scan - Know the array length for arm,record-impl and arm,status-reporting I will also fix the misleading comment in arm-ras.h from "register window page size" to "error record group register window layout" in v3. > > Best regards, > Krzysztof > Thanks, Umang