linux-kernel - Re: [PATCH V2 0/4] soc: imx: add i.MX BLK-CTL support

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

Message-ID: <2f0300ba-2465-b379-6824-6b66cfad1c59@kontron.de>
Date:   Tue, 4 May 2021 10:55:03 +0200
From:   Frieder Schrempf <frieder.schrempf@...tron.de>
To:     "Peng Fan (OSS)" <peng.fan@....nxp.com>, robh+dt@...nel.org,
        shawnguo@...nel.org, s.hauer@...gutronix.de
Cc:     kernel@...gutronix.de, festevam@...il.com, linux-imx@....com,
        p.zabel@...gutronix.de, l.stach@...gutronix.de, krzk@...nel.org,
        agx@...xcpu.org, marex@...x.de, andrew.smirnov@...il.com,
        devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, ping.bai@....com, aford173@...il.com,
        abel.vesa@....com, Peng Fan <peng.fan@....com>
Subject: Re: [PATCH V2 0/4] soc: imx: add i.MX BLK-CTL support

On 03.05.21 16:57, Frieder Schrempf wrote:
> Hi Peng,
> 
> On 30.04.21 07:27, Peng Fan (OSS) wrote:
>> From: Peng Fan <peng.fan@....com>
>>
>> V2:
>>   Fix yaml check failure
>>
>> Previously there is an effort from Abel that take BLK-CTL as clock
>> provider, but it turns out that there is A/B lock issue and we are
>> not able resolve that.
>>
>> Per discuss with Lucas and Jacky, we made an agreement that take BLK-CTL
>> as a power domain provider and use GPC's domain as parent, the consumer
>> node take BLK-CTL as power domain input.
>>
>> This patchset has been tested on i.MX8MM EVK board, but one hack
>> is not included in the patchset is that the DISPMIX BLK-CTL 
>> MIPI_M/S_RESET
>> not implemented. Per Lucas, we will finally have a MIPI DPHY driver,
>> so fine to leave it.
> 
> Thanks for your work. I would like to test this together with the DSIM 
> and PHY driver by Michael and Marek. So far the boot hangs when probing 
> the DSIM, but I'm not even sure if my DT is correct.
> 
> With the DSIM PHY driver (see [1]) in place, the GPR_MIPI_M_RESETN 
> should be set correctly, right?

So I found out, that with the hack below applied (taken from Marek's WIP 
patches for DSIM support) it seems to work properly.

Therefore I guess this is an issue with the DSIM driver, that somehow 
needs to make sure that the proper clocks are enabled at probe time.

diff --git a/drivers/soc/imx/gpcv2.c b/drivers/soc/imx/gpcv2.c
index 04564017bfe9..5a9bca805b0c 100644
--- a/drivers/soc/imx/gpcv2.c
+++ b/drivers/soc/imx/gpcv2.c
@@ -9,6 +9,7 @@
   */

  #include <linux/clk.h>
+#include <linux/clk-provider.h>
  #include <linux/of_device.h>
  #include <linux/platform_device.h>
  #include <linux/pm_domain.h>
@@ -193,6 +194,7 @@ to_imx_pgc_domain(struct generic_pm_domain *genpd)
  static int imx_pgc_power_up(struct generic_pm_domain *genpd)
  {
         struct imx_pgc_domain *domain = to_imx_pgc_domain(genpd);
+       unsigned int i;
         u32 reg_val;
         int ret;

@@ -254,7 +256,14 @@ static int imx_pgc_power_up(struct 
generic_pm_domain *genpd)
         }

         /* Disable reset clocks for all devices in the domain */
-       clk_bulk_disable_unprepare(domain->num_clks, domain->clks);
+       //clk_bulk_disable_unprepare(domain->num_clks, domain->clks);
+       for (i = 0; i < domain->num_clks; i++) {
+               /* Keep the DISPMIX active, it is needed by both LCDIF 
and MIPI */
+               if (strcmp(__clk_get_name(domain->clks[i].clk), 
"disp_root_clk") &&
+                   strcmp(__clk_get_name(domain->clks[i].clk), 
"disp_axi_root_clk") &&
+                   strcmp(__clk_get_name(domain->clks[i].clk), 
"disp_apb_root_clk"))
+                       clk_disable_unprepare(domain->clks[i].clk);
+       }

         return 0;