linux-kernel - Re: [PATCH] phy: qcom-qmp: Correct READY

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Mon, 5 Aug 2019 17:43:42 -0700
From:   Bjorn Andersson <bjorn.andersson@...aro.org>
To:     Marc Gonzalez <marc.w.gonzalez@...e.fr>
Cc:     Kishon Vijay Abraham I <kishon@...com>,
        MSM <linux-arm-msm@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Evan Green <evgreen@...omium.org>,
        Vivek Gautam <vivek.gautam@...eaurora.org>,
        Niklas Cassel <niklas.cassel@...aro.org>,
        Stanimir Varbanov <svarbanov@...sol.com>
Subject: Re: [PATCH] phy: qcom-qmp: Correct READY_STATUS poll break condition

On Fri 19 Jul 08:50 PDT 2019, Marc Gonzalez wrote:

> On 13/06/2019 11:10, Marc Gonzalez wrote:
> 
> > Here are my observations for a 8998 board:
> > 
> > 1) If I apply only the readl_poll_timeout() fix (not the mask_pcs_ready fixup)
> > qcom_pcie_probe() fails with a timeout in phy_init.
> > => this is in line with your regression analysis.
> > 
> > 2) Your patch also fixes a long-standing bug in UFS init whereby sending
> > lots of information to the console during phy init would lead to an
> > incorrectly diagnosed time-out.
> > 
> > Good stuff!
> > 
> > Reviewed-by: Marc Gonzalez <marc.w.gonzalez@...e.fr>
> > Tested-by: Marc Gonzalez <marc.w.gonzalez@...e.fr>
> 
> It looks like this patch fixed UFS, but broke PCIe and USB3 ^_^
> 
> qcom-qmp-phy 1c06000.phy: Registered Qcom-QMP phy
> qcom-qmp-phy c010000.phy: Registered Qcom-QMP phy
> qcom-qmp-phy 1da7000.phy: Registered Qcom-QMP phy
> 
> qcom-qmp-phy 1c06000.phy: BEFORE=000000a6 AFTER=000000a6
> qcom-qmp-phy 1c06000.phy: phy initialization timed-out
> phy phy-1c06000.phy.0: phy init failed --> -110
> qcom-pcie: probe of 1c00000.pci failed with error -110
> 
> qcom-qmp-phy 1da7000.phy: BEFORE=00000040 AFTER=0000000d
> 
> qcom-qmp-phy c010000.phy: BEFORE=69696969 AFTER=b7b7b7b7
> qcom-qmp-phy c010000.phy: phy initialization timed-out
> phy phy-c010000.phy.1: phy init failed --> -110
> dwc3 a800000.dwc3: failed to initialize core: -110
> dwc3: probe of a800000.dwc3 failed with error -110
> 
> 
> Downstream code for PCIe is:
> 
> static bool pcie_phy_is_ready(struct msm_pcie_dev_t *dev)
> {
> 	if (dev->phy_ver >= 0x20) {
> 		if (readl_relaxed(dev->phy + PCIE_N_PCS_STATUS(dev->rc_idx, dev->common_phy)) &	BIT(6))
> 			return false;
> 		else
> 			return true;
> 	}
> 
> 	if (!(readl_relaxed(dev->phy + PCIE_COM_PCS_READY_STATUS) & 0x1))
> 		return false;
> 	else
> 		return true;
> }
> 
> AFAICT:
> PCIe and USB3 QMP PHYs are ready when PHYSTATUS=BIT(6) goes to 0.
> But UFS is ready when PCS_READY=BIT(0) goes to 1.
> 
> 
> Can someone verify that USB3 is broken on 845 with 885bd765963b?
> 

Both USB3 and PCIe still works for me on 845, but you are correct;
PHYSTATUS is ready low, while PCS_READY is ready high.

I posted a patch that attempts to clean up the differences, please give
it a spin (and review).

Regards,
Bjorn