ARM: imx: gpc: Initialize all power domains

From: Fabio Estevam <[email protected]>

Since commit 0159ec670763dd ("PM / Domains: Verify the PM domain is present when adding a provider") the following regression is observed on imx6:

imx-gpc: probe of 20dc000.gpc failed with error -22

The imx-gpc driver probe failure causes several issues such as:

- When booting a kernel built with multi_v7_defconfig a kernel crash is seen.

- When booting a kernel built with imx_v6_v7_defconfig the Etnaviv GPU driver is not loaded due to the lack of power domains.

The gpc probe fails because of_genpd_add_provider_onecell() now checks if all the domains are initialized via pm_genpd_present() function and it fails because not all the power domains are initialized.

In order to fix this error, initialize all the power domains from imx_gpc_domains[], not only the imx6q_pu_domain.base one.

---
 arch/arm/mach-imx/gpc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 0df062d..d0463e9 100644
@@ -430,7 +430,8 @@ static int imx_gpc_genpd_init(struct device *dev, struct regulator *pu_reg)
 	if (!IS_ENABLED(CONFIG_PM_GENERIC_DOMAINS))
 		return 0;
 
-	pm_genpd_init(&imx6q_pu_domain.base, NULL, false);
+	for (i = 0; i < ARRAY_SIZE(imx_gpc_domains); i++)
+		pm_genpd_init(imx_gpc_domains[i], NULL, false);
 	return of_genpd_add_provider_onecell(dev->of_node,
 					     &imx_gpc_onecell_data);
 

2.7.4

Thanks for the update. Now we understand that this is a regression caused by power domain core change. But my question on why kernel crashes with multi_v7_defconfig but not with imx_v6_v7_defconfig stays unanswered.

Shawn

Hi Shawn,

imx_v6_v7_defconfig selects GPU and VPU drivers, which are consumers of the GPC driver and change the refcount of the pu_reg regulator.

I think we should not continue to boot in case imx_gpc_probe() fails.

So I will send a patch series with this patch and another one that prevents the boot when imx_gpc_probe() fails.

So in addition to the original patch, I would like to add a second one in the series to treat the imx_gpc_probe() case.

Option 1:

commit eb8aad1447e2227d5706dfa628bdb4c82381202e

Author: Fabio Estevam <[email protected]>
Date:   Fri Oct 21 11:31:06 2016 -0200

ARM: imx: gpc: Fix the imx_gpc_genpd_init() error path

If of_genpd_add_provider_onecell() fails the following kernel crash is observed on a kernel built with multi_v7_defconfig:

[ 1.723358] Unable to handle kernel NULL pointer dereference at
virtual address 00000040

    [    1.731500] pgd = c0204000
    [    1.731863] hctosys: unable to open rtc device (rtc0)
    [    1.739301] [00000040] *pgd=00000000
    [    1.739310] Internal error: Oops: 5 [#1] SMP ARM
    [    1.739319] Modules linked in:
    [    1.739328] CPU: 1 PID: 95 Comm: kworker/1:4 Not tainted
4.8.0-11897-g6b5e09a #1
    [    1.739331] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
    [    1.739352] Workqueue: pm genpd_power_off_work_fn
    [    1.739356] task: ee63d400 task.stack: ee70a000
    [    1.739365] PC is at mutex_lock+0xc/0x4c
    [    1.739374] LR is at regulator_disable+0x2c/0x60
    [    1.739379] pc : [<c0bc0da0>]    lr : [<c06e4b10>]    psr: 60000013
    [    1.739379] sp : ee70beb0  ip : 10624dd3  fp : ee6e6280
    [    1.739382] r10: eefb0900  r9 : 00000000  r8 : c1309918
    [    1.739385] r7 : 00000000  r6 : 00000040  r5 : 00000000  r4 : 00000040
    [    1.739390] r3 : 0000004c  r2 : 7fffd540  r1 : 000001e4  r0 : 00000040

Instead of returning of_genpd_add_provider_onecell() directly, we should check its return value and in the case of error we should unwind the previously actions, which in these case are:
- Call imx6q_pm_pu_power_off()
- Set imx6q_pu_domain.reg back to NULL

Setting imx6q_pu_domain.reg to NULL in the error case is required as it will prevent further operations on the pu_reg regulator.

This kernel crash is not observed with imx_v6_v7_defconfig because it selects GPU and VPU drivers, which are consumers of the GPC block.

index d0463e9..b54db47 100644
@@ -408,7 +408,7 @@ static struct genpd_onecell_data imx_gpc_onecell_data = {
 static int imx_gpc_genpd_init(struct device *dev, struct regulator *pu_reg)
 {
        struct clk *clk;
-       int i;
+       int i, ret;
 
        imx6q_pu_domain.reg = pu_reg;
 

@@ -432,12 +432,20 @@ static int imx_gpc_genpd_init(struct device

*dev, struct regulator *pu_reg)

        for (i = 0; i < ARRAY_SIZE(imx_gpc_domains); i++)
                pm_genpd_init(imx_gpc_domains[i], NULL, false);
-       return of_genpd_add_provider_onecell(dev->of_node,
+
+       ret =  of_genpd_add_provider_onecell(dev->of_node,
                                             &imx_gpc_onecell_data);

+ if (ret)
+ goto power_off;
+
+ return 0;

+power_off:
+ imx6q_pm_pu_power_off(&imx6q_pu_domain.base);
clk_err:

        while (i--)
                clk_put(imx6q_pu_domain.clk[i]);
+       imx6q_pu_domain.reg = NULL;
        return -EINVAL;

}

Or we can not even try to boot the kernel:

Option 2:

@@ -459,7 +459,13 @@ static int imx_gpc_probe(struct platform_device *pdev)
                return ret;
        }
 
-       return imx_gpc_genpd_init(&pdev->dev, pu_reg);
+       ret = imx_gpc_genpd_init(&pdev->dev, pu_reg);
+       /*
+        * If imx_gpc_probe() fails we should better not continue
+        * to boot as the SoC may be in an impredictable state
+        */
+       BUG_ON(ret);
+       return ret;
 }
 

I am more inclined towards option 2, but would like to get some feedback first.