Oops on boot for omap3xxx, EHCI fails.

[PATCH] Fix OMAP EHCI suspend/resume failure (i693) '354ab856' causes an oops on boot for all omap3xxx platforms that use usbhs_omap for EHCI. The actual oops seems to come from faulty ehci-omap cleanup, but the failure caused by the change is evidenced here:

[    3.655059] ehci-omap ehci-omap.0: utmi_p1_gfclk failed error:-2
[    3.661376] ehci-omap: probe of ehci-omap.0 failed with error -2

The problem is the call:

omap->utmi_p1_fck = clk_get(dev, "utmi_p1_gfclk");

Where dev here is ehci-omap. As shown below, clk_get will fail with dev_id "ehci-omap" and clk_id "utmi_p1_gfclk" (but succeed on omap44xx).

arch/arm/mach-omap2/clock3xxx_data.c: CLK("usbhs_omap", "utmi_p1_gfclk",
arch/arm/mach-omap2/clock44xx_data.c: CLK(NULL, "utmi_p1_gfclk",

The oops caused by bad cleanup is shown below:

[    3.951934] Unable to handle kernel NULL pointer dereference at
virtual address 00000000
[    3.960418] pgd = c0004000
[    3.963256] [00000000] *pgd=00000000
[    3.967010] Internal error: Oops: 5 [#1] SMP ARM
[    3.971801] Modules linked in:
[    3.975006] CPU: 0    Not tainted  (3.5.0-rc1-ktest-00040-g354ab85 #401)
[    3.981964] PC is at hub_port_init+0x28c/0x968
[    3.986602] LR is at hub_port_init+0xc4/0x968
[    3.991119] pc : [<c03304f8>]    lr : [<c0330330>]    psr: 60000013
[    3.991119] sp : de11deb8  ip : c05eae4c  fp : 00000000
[    4.003082] r10: 00000000  r9 : de38ac00  r8 : 00000032
[    4.008514] r7 : de388800  r6 : 00000002  r5 : 00000000  r4 : de391800
[    4.015289] r3 : 00000000  r2 : c05de444  r1 : c05de468  r0 : de391868
[    4.022094] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM
Segment kernel
[    4.029693] Control: 10c5387d  Table: 80004019  DAC: 00000017
[    4.035675] Process khubd (pid: 242, stack limit = 0xde11c2f8)
[    4.041748] Stack: (0xde11deb8 to 0xde11e000)
[    4.046264] dea0:
    00000002 de38ac00
[    4.054779] dec0: 00000002 00000000 00000004 c00891b0 00000001
00000002 00000003 de378000
[    4.063293] dee0: 00000000 c0479d28 00000001 de391800 de388800
de378000 de378018 de38ac00
[    4.071807] df00: 00000002 00000000 00000002 c0333d68 00000501
00000000 c05df1b0 c069d600
[    4.080322] df20: 10a5bb6a de38ac08 de388420 de38889c de388400
00000004 de38ac70 de38ac01
[    4.088806] df40: de388800 de38ac68 de388800 de388420 de378000
00000009 de11df8c c0064834
[    4.097320] df60: 00000000 de11c000 00000000 de11f440 c005c75c
de11df74 de11df74 00000000
[    4.105834] df80: 00000501 c00891b0 de02ff18 de02fef0 00000000
c03336fc 00000013 00000000
[    4.114349] dfa0: 00000000 00000000 00000000 c005c064 00000000
00000000 00000000 00000000
[    4.122863] dfc0: 00000000 dead4ead ffffffff ffffffff c0747830
00000000 00000000 c0597cd0
[    4.131378] dfe0: de11dfe0 de11dfe0 de02fef0 c005bfe0 c0014154
c0014154 00000200 00000000
[    4.139892] [<c03304f8>] (hub_port_init+0x28c/0x968) from
[<c0333d68>] (hub_thread+0x66c/0x13d8)
[    4.149047] [<c0333d68>] (hub_thread+0x66c/0x13d8) from
[<c005c064>] (kthread+0x84/0x90)
[    4.157501] [<c005c064>] (kthread+0x84/0x90) from [<c0014154>]
(kernel_thread_exit+0x0/0x8)
[    4.166168] Code: e59f2690 e5933080 e2840068 e59f1688 (e593e000)
[    4.172607] ---[ end trace f11541c260e8a205 ]---

A recent commit, [PATCH] Fix OMAP EHCI suspend/resume failure (i693) '354ab856' causes ehci probe to fail on omap3xxx. This exposed bugs in the ehci_hcd_omap_probe error path causing an oops.

On the error path, call usb_remove_hcd if usb_add_hcd has been called, and call usb_put_hcd if usb_alloc_hcd has been called.

Tested on BB-xM.

---
 drivers/usb/host/ehci-omap.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

index 17cfb8a..6e15fc8 100644
@@ -347,7 +347,7 @@ static int ehci_hcd_omap_probe(struct platform_device *pdev)
 	ret = usb_add_hcd(hcd, irq, IRQF_SHARED);
 	if (ret) {
 		dev_err(dev, "failed to add hcd with err %d\n", ret);
-		goto err_add_hcd;
+		goto err_pm_runtime;
 	}
 
 	/* root ports should always stay powered */

@@ -424,8 +424,12 @@ err_utmi_p1_fck:
 	clk_put(utmi_p1_fck);
 
 err_add_hcd:
+	usb_remove_hcd(hcd);
+
+err_pm_runtime:
 	disable_put_regulator(pdata);
 	pm_runtime_put_sync(dev);
+	usb_put_hcd(hcd);
 
 err_io:
 	iounmap(regs);

1.7.10.4

[PATCH] Fix OMAP EHCI suspend/resume failure (i693) '354ab856' causes an oops on boot for all omap3xxx platforms that use usbhs_omap for EHCI. The actual oops comes from faulty ehci-omap cleanup, but the failure caused by the change is evidenced here:

[    3.655059] ehci-omap ehci-omap.0: utmi_p1_gfclk failed error:-2
[    3.661376] ehci-omap: probe of ehci-omap.0 failed with error -2

utmi_p1_gfclk is a clock that exists on OMAP4, but not OMAP3. In the OMAP3 case, it is configured as a dummy clock. However, OMAP4 lists the dev_id as NULL, but OMAP3 lists it as "usbhs_omap".

Attempting to get that clock from ehci-omap then fails. The solution is to just change the clock3xxx_data.c for dummy clocks used in the errata fix to match the dev_id, NULL, used in clock44xx_data.c.

Tested on BB-xM.

---
 arch/arm/mach-omap2/clock3xxx_data.c |   14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

index 4e1a3b0..59d5315 100644
@@ -3391,15 +3391,15 @@ static struct omap_clk omap3xxx_clks[] = {
 	CLK(NULL,	"usbhost_48m_fck", &usbhost_48m_fck, CK_3430ES2PLUS | CK_AM35XX | CK_36XX),
 	CLK(NULL,	"usbhost_ick",	&usbhost_ick,	CK_3430ES2PLUS | CK_AM35XX | CK_36XX),
 	CLK("usbhs_omap",	"usbhost_ick",	&usbhost_ick,	CK_3430ES2PLUS | CK_AM35XX | CK_36XX),
-	CLK("usbhs_omap",	"utmi_p1_gfclk",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"utmi_p2_gfclk",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"xclk60mhsp1_ck",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"xclk60mhsp2_ck",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"usb_host_hs_utmi_p1_clk",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"usb_host_hs_utmi_p2_clk",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"utmi_p1_gfclk",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"utmi_p2_gfclk",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"xclk60mhsp1_ck",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"xclk60mhsp2_ck",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"usb_host_hs_utmi_p1_clk",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"usb_host_hs_utmi_p2_clk",	&dummy_ck,	CK_3XXX),
 	CLK("usbhs_omap",	"usb_tll_hs_usb_ch0_clk",	&dummy_ck,	CK_3XXX),
 	CLK("usbhs_omap",	"usb_tll_hs_usb_ch1_clk",	&dummy_ck,	CK_3XXX),
-	CLK("usbhs_omap",	"init_60m_fclk",	&dummy_ck,	CK_3XXX),
+	CLK(NULL,	"init_60m_fclk",	&dummy_ck,	CK_3XXX),
 	CLK(NULL,	"usim_fck",	&usim_fck,	CK_3430ES2PLUS | CK_36XX),
 	CLK(NULL,	"gpt1_fck",	&gpt1_fck,	CK_3XXX),
 	CLK(NULL,	"wkup_32k_fck",	&wkup_32k_fck,	CK_3XXX),

1.7.10.4

Hello.

Usually commit ID follows the word "commit", and then the commit summary.

It should be:

WBR, Sergei