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: <CY5PR11MB6366EF9CA6552ADF6E01A557ED4B2@CY5PR11MB6366.namprd11.prod.outlook.com>
Date: Tue, 29 Oct 2024 11:24:36 +0000
From: "Usyskin, Alexander" <alexander.usyskin@...el.com>
To: "Gupta, Anshuman" <anshuman.gupta@...el.com>, "Vivi, Rodrigo"
	<rodrigo.vivi@...el.com>, "Deak, Imre" <imre.deak@...el.com>
CC: Miquel Raynal <miquel.raynal@...tlin.com>, Richard Weinberger
	<richard@....at>, Vignesh Raghavendra <vigneshr@...com>, "De Marchi, Lucas"
	<lucas.demarchi@...el.com>, Thomas Hellström
	<thomas.hellstrom@...ux.intel.com>, Maarten Lankhorst
	<maarten.lankhorst@...ux.intel.com>, Maxime Ripard <mripard@...nel.org>,
	Thomas Zimmermann <tzimmermann@...e.de>, David Airlie <airlied@...il.com>,
	Simona Vetter <simona@...ll.ch>, Jani Nikula <jani.nikula@...ux.intel.com>,
	Joonas Lahtinen <joonas.lahtinen@...ux.intel.com>, Tvrtko Ursulin
	<tursulin@...ulin.net>, "Weil, Oren jer" <oren.jer.weil@...el.com>,
	"linux-mtd@...ts.infradead.org" <linux-mtd@...ts.infradead.org>,
	"dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>,
	"intel-gfx@...ts.freedesktop.org" <intel-gfx@...ts.freedesktop.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 06/10] mtd: intel-dg: wake card on operations

> -----Original Message-----
> From: Gupta, Anshuman <anshuman.gupta@...el.com>
> Sent: Monday, October 28, 2024 5:02 PM
> To: Vivi, Rodrigo <rodrigo.vivi@...el.com>; Usyskin, Alexander
> <alexander.usyskin@...el.com>; Deak, Imre <imre.deak@...el.com>
> Cc: Miquel Raynal <miquel.raynal@...tlin.com>; Richard Weinberger
> <richard@....at>; Vignesh Raghavendra <vigneshr@...com>; De Marchi,
> Lucas <lucas.demarchi@...el.com>; Thomas Hellström
> <thomas.hellstrom@...ux.intel.com>; Maarten Lankhorst
> <maarten.lankhorst@...ux.intel.com>; Maxime Ripard <mripard@...nel.org>;
> Thomas Zimmermann <tzimmermann@...e.de>; David Airlie
> <airlied@...il.com>; Simona Vetter <simona@...ll.ch>; Jani Nikula
> <jani.nikula@...ux.intel.com>; Joonas Lahtinen
> <joonas.lahtinen@...ux.intel.com>; Tvrtko Ursulin <tursulin@...ulin.net>;
> Weil, Oren jer <oren.jer.weil@...el.com>; linux-mtd@...ts.infradead.org; dri-
> devel@...ts.freedesktop.org; intel-gfx@...ts.freedesktop.org; linux-
> kernel@...r.kernel.org
> Subject: RE: [PATCH 06/10] mtd: intel-dg: wake card on operations
> 
> 
> 
> > -----Original Message-----
> > From: Vivi, Rodrigo <rodrigo.vivi@...el.com>
> > Sent: Monday, October 28, 2024 8:27 PM
> > To: Usyskin, Alexander <alexander.usyskin@...el.com>; Gupta, Anshuman
> > <anshuman.gupta@...el.com>; Deak, Imre <imre.deak@...el.com>
> > Cc: Miquel Raynal <miquel.raynal@...tlin.com>; Richard Weinberger
> > <richard@....at>; Vignesh Raghavendra <vigneshr@...com>; De Marchi,
> > Lucas <lucas.demarchi@...el.com>; Thomas Hellström
> > <thomas.hellstrom@...ux.intel.com>; Maarten Lankhorst
> > <maarten.lankhorst@...ux.intel.com>; Maxime Ripard
> <mripard@...nel.org>;
> > Thomas Zimmermann <tzimmermann@...e.de>; David Airlie
> > <airlied@...il.com>; Simona Vetter <simona@...ll.ch>; Jani Nikula
> > <jani.nikula@...ux.intel.com>; Joonas Lahtinen
> > <joonas.lahtinen@...ux.intel.com>; Tvrtko Ursulin <tursulin@...ulin.net>;
> > Weil, Oren jer <oren.jer.weil@...el.com>; linux-mtd@...ts.infradead.org; dri-
> > devel@...ts.freedesktop.org; intel-gfx@...ts.freedesktop.org; linux-
> > kernel@...r.kernel.org
> > Subject: Re: [PATCH 06/10] mtd: intel-dg: wake card on operations
> >
> > On Tue, Oct 22, 2024 at 01:41:15PM +0300, Alexander Usyskin wrote:
> > > Enable runtime PM in mtd driver to notify graphics driver that whole
> > > card should be kept awake while nvm operations are performed through
> > > this driver.
> > >
> > > CC: Lucas De Marchi <lucas.demarchi@...el.com>
> > > Signed-off-by: Alexander Usyskin <alexander.usyskin@...el.com>
> > > ---
> > >  drivers/mtd/devices/mtd-intel-dg.c | 73
> > > +++++++++++++++++++++++++-----
> > >  1 file changed, 61 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/drivers/mtd/devices/mtd-intel-dg.c
> > > b/drivers/mtd/devices/mtd-intel-dg.c
> > > index 4951438e8faf..3d53f35478e8 100644
> > > --- a/drivers/mtd/devices/mtd-intel-dg.c
> > > +++ b/drivers/mtd/devices/mtd-intel-dg.c
> > > @@ -15,11 +15,14 @@
> > >  #include <linux/module.h>
> > >  #include <linux/mtd/mtd.h>
> > >  #include <linux/mtd/partitions.h>
> > > +#include <linux/pm_runtime.h>
> > >  #include <linux/string.h>
> > >  #include <linux/slab.h>
> > >  #include <linux/sizes.h>
> > >  #include <linux/types.h>
> > >
> > > +#define INTEL_DG_NVM_RPM_TIMEOUT 500
> > > +
> > >  struct intel_dg_nvm {
> > >  	struct kref refcnt;
> > >  	struct mtd_info mtd;
> > > @@ -462,6 +465,7 @@ static int intel_dg_mtd_erase(struct mtd_info
> *mtd,
> > struct erase_info *info)
> > >  	loff_t from;
> > >  	size_t len;
> > >  	size_t total_len;
> > > +	int ret = 0;
> > >
> > >  	if (WARN_ON(!nvm))
> > >  		return -EINVAL;
> > > @@ -476,20 +480,28 @@ static int intel_dg_mtd_erase(struct mtd_info
> > *mtd, struct erase_info *info)
> > >  	total_len = info->len;
> > >  	addr = info->addr;
> > >
> > > +	ret = pm_runtime_resume_and_get(mtd->dev.parent);
> > > +	if (ret < 0) {
> > > +		dev_err(&mtd->dev, "rpm: get failed %d\n", ret);
> > > +		return ret;
> > > +	}
> > > +
> > >  	guard(mutex)(&nvm->lock);
> > >
> > >  	while (total_len > 0) {
> > >  		if (!IS_ALIGNED(addr, SZ_4K) || !IS_ALIGNED(total_len,
> > SZ_4K)) {
> > >  			dev_err(&mtd->dev, "unaligned erase %llx %zx\n",
> > addr, total_len);
> > >  			info->fail_addr = addr;
> > > -			return -ERANGE;
> > > +			ret = -ERANGE;
> > > +			goto out;
> > >  		}
> > >
> > >  		idx = idg_nvm_get_region(nvm, addr);
> > >  		if (idx >= nvm->nregions) {
> > >  			dev_err(&mtd->dev, "out of range");
> > >  			info->fail_addr = MTD_FAIL_ADDR_UNKNOWN;
> > > -			return -ERANGE;
> > > +			ret = -ERANGE;
> > > +			goto out;
> > >  		}
> > >
> > >  		from = addr - nvm->regions[idx].offset; @@ -505,14 +517,18
> > @@
> > > static int intel_dg_mtd_erase(struct mtd_info *mtd, struct erase_info
> *info)
> > >  		if (bytes < 0) {
> > >  			dev_dbg(&mtd->dev, "erase failed with %zd\n",
> > bytes);
> > >  			info->fail_addr += nvm->regions[idx].offset;
> > > -			return bytes;
> > > +			ret = bytes;
> > > +			goto out;
> > >  		}
> > >
> > >  		addr += len;
> > >  		total_len -= len;
> > >  	}
> > >
> > > -	return 0;
> > > +out:
> > > +	pm_runtime_mark_last_busy(mtd->dev.parent);
> > > +	pm_runtime_put_autosuspend(mtd->dev.parent);
> > > +	return ret;
> > >  }
> > >
> > >  static int intel_dg_mtd_read(struct mtd_info *mtd, loff_t from,
> > > size_t len, @@ -541,17 +557,25 @@ static int intel_dg_mtd_read(struct
> > mtd_info *mtd, loff_t from, size_t len,
> > >  	if (len > nvm->regions[idx].size - from)
> > >  		len = nvm->regions[idx].size - from;
> > >
> > > +	ret = pm_runtime_resume_and_get(mtd->dev.parent);
> > > +	if (ret < 0) {
> > > +		dev_err(&mtd->dev, "rpm: get failed %zd\n", ret);
> > > +		return ret;
> > > +	}
> > > +
> > >  	guard(mutex)(&nvm->lock);
> > >
> > >  	ret = idg_read(nvm, region, from, len, buf);
> > >  	if (ret < 0) {
> > >  		dev_dbg(&mtd->dev, "read failed with %zd\n", ret);
> > > -		return ret;
> > > +	} else {
> > > +		*retlen = ret;
> > > +		ret = 0;
> > >  	}
> > >
> > > -	*retlen = ret;
> > > -
> > > -	return 0;
> > > +	pm_runtime_mark_last_busy(mtd->dev.parent);
> > > +	pm_runtime_put_autosuspend(mtd->dev.parent);
> > > +	return ret;
> > >  }
> > >
> > >  static int intel_dg_mtd_write(struct mtd_info *mtd, loff_t to, size_t
> > > len, @@ -580,17 +604,25 @@ static int intel_dg_mtd_write(struct
> mtd_info
> > *mtd, loff_t to, size_t len,
> > >  	if (len > nvm->regions[idx].size - to)
> > >  		len = nvm->regions[idx].size - to;
> > >
> > > +	ret = pm_runtime_resume_and_get(mtd->dev.parent);
> > > +	if (ret < 0) {
> > > +		dev_err(&mtd->dev, "rpm: get failed %zd\n", ret);
> > > +		return ret;
> > > +	}
> > > +
> > >  	guard(mutex)(&nvm->lock);
> > >
> > >  	ret = idg_write(nvm, region, to, len, buf);
> > >  	if (ret < 0) {
> > >  		dev_dbg(&mtd->dev, "write failed with %zd\n", ret);
> > > -		return ret;
> > > +	} else {
> > > +		*retlen = ret;
> > > +		ret = 0;
> > >  	}
> > >
> > > -	*retlen = ret;
> > > -
> > > -	return 0;
> > > +	pm_runtime_mark_last_busy(mtd->dev.parent);
> > > +	pm_runtime_put_autosuspend(mtd->dev.parent);
> > > +	return ret;
> > >  }
> > >
> > >  static void intel_dg_nvm_release(struct kref *kref) @@ -722,6 +754,17
> > > @@ static int intel_dg_mtd_probe(struct auxiliary_device *aux_dev,
> > >  		n++;
> > >  	}
> > >
> > > +	pm_runtime_enable(device);
> > > +
> > > +	pm_runtime_set_autosuspend_delay(device,
> > INTEL_DG_NVM_RPM_TIMEOUT);
> > > +	pm_runtime_use_autosuspend(device);
> >
> > something looks strange here...
> > on the functions above you get and put references for the parent device
> > directly.
> > But here you enable the rpm for this device.
> >
> > If I remember correctly from some old experiments, the rpm is smart enough
> > and wake up the parent before waking up the child. So I'm wondering if we
> > should only do the child without the parent.
> Agree parent can't runtime suspend ind if it has active child.
> Let this driver have it's own get/put routine instead of parent.
> Thanks,
> Anshuman Gupta.

I need to wake DG card before the MTD device is established to scan the partition table on flash.
Thus, if I move rpm down to MTD device I should enable and take parent (auxiliary device) rpm anyway.
Considering above, is this move is justified?
Also, MTD drivers tend to enable parent rpm, not its own one.

- - 
Thanks,
Sasha



> >
> > Cc: Imre Deak <imre.deak@...el.com>
> > Cc: Anshuman Gupta <anshuman.gupta@...el.com>
> >
> > > +
> > > +	ret = pm_runtime_resume_and_get(device);
> > > +	if (ret < 0) {
> > > +		dev_err(device, "rpm: get failed %d\n", ret);
> > > +		goto err_norpm;
> > > +	}
> > > +
> > >  	nvm->base = devm_ioremap_resource(device, &invm->bar);
> > >  	if (IS_ERR(nvm->base)) {
> > >  		dev_err(device, "mmio not mapped\n"); @@ -744,9 +787,13
> > @@ static
> > > int intel_dg_mtd_probe(struct auxiliary_device *aux_dev,
> > >
> > >  	dev_set_drvdata(&aux_dev->dev, nvm);
> > >
> > > +	pm_runtime_put(device);
> > >  	return 0;
> > >
> > >  err:
> > > +	pm_runtime_put(device);
> > > +err_norpm:
> > > +	pm_runtime_disable(device);
> > >  	kref_put(&nvm->refcnt, intel_dg_nvm_release);
> > >  	return ret;
> > >  }
> > > @@ -758,6 +805,8 @@ static void intel_dg_mtd_remove(struct
> > auxiliary_device *aux_dev)
> > >  	if (!nvm)
> > >  		return;
> > >
> > > +	pm_runtime_disable(&aux_dev->dev);
> > > +
> > >  	mtd_device_unregister(&nvm->mtd);
> > >
> > >  	dev_set_drvdata(&aux_dev->dev, NULL);
> > > --
> > > 2.43.0
> > >

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ