openwrt/target/linux/starfive/patches-6.6/0072-drivers-nvme-Add-precheck-and-delay-for-CQE-pending-.patch

41 lines
1.1 KiB
Diff
Raw Normal View History

From f0b4cffe4d1813305f783d208f260747ecc56c50 Mon Sep 17 00:00:00 2001
From: "Kevin.xie" <kevin.xie@starfivetech.com>
Date: Thu, 24 Nov 2022 16:59:12 +0800
Subject: [PATCH 072/116] drivers: nvme: Add precheck and delay for CQE pending
status.
To workaroud the NVMe I/O timeout problem in bootup S10udev case
which caused by the CQE update lantancy.
Signed-off-by: Kevin.xie <kevin.xie@starfivetech.com>
---
drivers/nvme/host/pci.c | 10 ++++++++++
1 file changed, 10 insertions(+)
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -28,6 +28,7 @@
#include <linux/io-64-nonatomic-hi-lo.h>
#include <linux/sed-opal.h>
#include <linux/pci-p2pdma.h>
+#include <linux/delay.h>
#include "trace.h"
#include "nvme.h"
kernel: bump 6.6 to 6.6.64 Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.64 Manually rebased: generic/hack-6.6/780-usb-net-MeigLink_modem_support.patch bcm27xx/patches-6.6/950-0585-drm-vc4-Introduce-generation-number-enum.patch bcm27xx/patches-6.6/950-0610-drm-vc4-hvs-Support-BCM2712-HVS.patch bcm27xx/patches-6.6/950-0829-vc4-hvs-Add-support-for-D0-register-changes.patch Removed upstreamed: bcm27xx/patches-6.6/950-0597-drm-vc4-hdmi-Avoid-hang-with-debug-registers-when-su.patch[1] bcm27xx/patches-6.6/950-0599-drm-vc4-Fix-dlist-debug-not-resetting-the-next-entry.patch[2] bcm27xx/patches-6.6/950-0600-drm-vc4-Remove-incorrect-limit-from-hvs_dlist-debugf.patch[3] bcm27xx/patches-6.6/950-0708-drm-vc4-Correct-logic-on-stopping-an-HVS-channel.patch[4] ramips/patches-6.6/002-01-v6.13-clk-ralink-mtmips-fix-clock-plan-for-Ralink-SoC-RT38.patch[5] ramips/patches-6.6/002-02-v6.13-clk-ralink-mtmips-fix-clocks-probe-order-in-oldest-r.patch[6] All other patches automatically rebased. 1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=223ee2567a55e4f80315c768d2969e6a3b9fb23d 2. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=8182b5ca19c6f173b6498d1c6d3e4b034b76bbde 3. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=52c1716f65a558174e381360bd88f18dae4be85c 4. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=9728b508b01a5eeeac79ceb676364c674dd951ac 5. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=f85a1d06afbcc57ac44176db8f9d7a934979952c 6. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.64&id=fbb13732c6ffa9d58cedafabcd5ce8fd7ef8ae5a Build system: x86/64 Build-tested: x86/64/AMD Cezanne, flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3 Run-tested: x86/64/AMD Cezanne, flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3 Co-authored-by: Mieczyslaw Nalewaj <namiltd@yahoo.com> Signed-off-by: John Audia <therealgraysky@proton.me> Link: https://github.com/openwrt/openwrt/pull/17217 Signed-off-by: Robert Marko <robimarko@gmail.com> Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> (cherry picked from commit 5158e28769cb5e5c6a59bd9a176f55cf2d568742)
2024-12-09 23:19:18 +00:00
@@ -1058,6 +1059,15 @@ static inline int nvme_poll_cq(struct nv
{
int found = 0;
+ /*
+ * In some cases, such as udev trigger, cqe status may update
+ * a little bit later than MSI, which cause an irq handle missing.
+ * To workaound, here we will prefetch the status first, and wait
+ * 1us if we get nothing.
+ */
+ if (!nvme_cqe_pending(nvmeq))
+ udelay(1);
+
while (nvme_cqe_pending(nvmeq)) {
found++;
/*