All of lore.kernel.org
 help / color / mirror / Atom feed
From: Caleb Connolly <caleb@connolly.tech>
To: caleb@connolly.tech, Andy Gross <agross@kernel.org>,
	Bjorn Andersson <bjorn.andersson@linaro.org>,
	Rob Herring <robh+dt@kernel.org>
Cc: ~postmarketos/upstreaming@lists.sr.ht,
	phone-devel@vger.kernel.org, linux-arm-msm@vger.kernel.org,
	devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH 3/4] dts: qcom: sdm845-oneplus-common: guard rmtfs-mem
Date: Sun, 02 May 2021 01:42:57 +0000	[thread overview]
Message-ID: <20210502014146.85642-4-caleb@connolly.tech> (raw)

The rmtfs_mem region is a weird one, downstream allocates it
dynamically, and supports a "qcom,guard-memory" property which when set
will reserve 4k above and below the rmtfs memory.

A common from qcom 4.9 kernel msm_sharedmem driver:

/*
 * If guard_memory is set, then the shared memory region
 * will be guarded by SZ_4K at the start and at the end.
 * This is needed to overcome the XPU limitation on few
 * MSM HW, so as to make this memory not contiguous with
 * other allocations that may possibly happen from other
 * clients in the system.
*/

When the kernel tries to touch memory that is too close the
rmtfs region it may cause an XPU violation. Such is the case on the
OnePlus 6 where random crashes would occur usually after boot.

Reserve 4k above and below the rmtfs_mem to avoid hitting these XPU
Violations.

This doesn't entirely solve the random crashes on the OnePlus 6/6T but
it does seem to prevent the ones which happen shortly after modem
bringup.

Signed-off-by: Caleb Connolly <caleb@connolly.tech>
---
 arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi b/arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi
index 8f3f5c687b4a..96c370b90550 100644
--- a/arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi
+++ b/arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi
@@ -46,6 +46,14 @@ vol-up {
 	};
 
 	reserved-memory {
+		/* The rmtfs_mem needs to be guarded due to "XPU limitations"
+		 * it is otherwise possible for an allocation adjacent to the
+		 * rmtfs_mem region to trigger an XPU violation, causing a crash.
+		 */
+		rmtfs_lower_guard: memory@f5b00000 {
+			no-map;
+			reg = <0 0xf5b00000 0 0x1000>;
+		};
 		/*
 		 * The rmtfs memory region in downstream is 'dynamically allocated'
 		 * but given the same address every time. Hard code it as this address is
@@ -59,6 +67,10 @@ rmtfs_mem: memory@f5b01000 {
 			qcom,client-id = <1>;
 			qcom,vmid = <15>;
 		};
+		rmtfs_upper_guard: memory@f5d01000 {
+			no-map;
+			reg = <0 0xf5d01000 0 0x2000>;
+		};
 
 		/*
 		 * It seems like reserving the old rmtfs_mem region is also needed to prevent
-- 
2.30.2



                 reply	other threads:[~2021-05-02  1:43 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20210502014146.85642-4-caleb@connolly.tech \
    --to=caleb@connolly.tech \
    --cc=agross@kernel.org \
    --cc=bjorn.andersson@linaro.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=phone-devel@vger.kernel.org \
    --cc=robh+dt@kernel.org \
    --cc=~postmarketos/upstreaming@lists.sr.ht \
    /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.