Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
To: "Summers, Stuart" <stuart.summers@intel.com>
Cc: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
	"dri-devel@lists.freedesktop.org"
	<dri-devel@lists.freedesktop.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Nilawar, Badal" <badal.nilawar@intel.com>,
	"gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
	"Usyskin, Alexander" <alexander.usyskin@intel.com>,
	"Ceraolo Spurio, Daniele" <daniele.ceraolospurio@intel.com>,
	"Gupta, Anshuman" <anshuman.gupta@intel.com>
Subject: Re: [PATCH 8/9] drm/xe/xe_late_bind_fw: Introduce debug fs node to disable late binding
Date: Fri, 1 Aug 2025 11:17:16 -0400	[thread overview]
Message-ID: <aIzafHWskNOEcXuC@intel.com> (raw)
In-Reply-To: <1f323cf45af3ba781d46d2e889f89bc26aa9b85a.camel@intel.com>

On Thu, Jul 31, 2025 at 08:03:36PM +0000, Summers, Stuart wrote:
> On Thu, 2025-07-10 at 11:08 -0400, Rodrigo Vivi wrote:
> > From: Badal Nilawar <badal.nilawar@intel.com>
> > 
> > Introduce a debug filesystem node to disable late binding fw reload
> > during the system or runtime resume. This is intended for situations
> > where the late binding fw needs to be loaded from user mode,
> > perticularly for validation purpose.
> > Note that xe kmd doesn't participate in late binding flow from user
> > space. Binary loaded from the userspace will be lost upon entering to
> > D3 cold hence user space app need to handle this situation.
> > 
> > v2:
> >   - s/(uval == 1) ? true : false/!!uval/ (Daniele)
> > v3:
> >   - Refine the commit message (Daniele)
> > 
> > Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
> > Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> 
> Bit of a flyby comment here, but would it be better if this was
> configfs rather than debugfs in case we wanted to skip this on the
> first probe?

Indeed. Configfs seems the best alternative for this. Otherwise
by the time we get to have the debugfs, the fw was already
attempted to get loaded at least once.

> 
> Thanks,
> Stuart
> 
> > ---
> >  drivers/gpu/drm/xe/xe_debugfs.c            | 41
> > ++++++++++++++++++++++
> >  drivers/gpu/drm/xe/xe_late_bind_fw.c       |  3 ++
> >  drivers/gpu/drm/xe/xe_late_bind_fw_types.h |  2 ++
> >  3 files changed, 46 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/xe/xe_debugfs.c
> > b/drivers/gpu/drm/xe/xe_debugfs.c
> > index d83cd6ed3fa8..d1f6f556efa2 100644
> > --- a/drivers/gpu/drm/xe/xe_debugfs.c
> > +++ b/drivers/gpu/drm/xe/xe_debugfs.c
> > @@ -226,6 +226,44 @@ static const struct file_operations
> > atomic_svm_timeslice_ms_fops = {
> >         .write = atomic_svm_timeslice_ms_set,
> >  };
> >  
> > +static ssize_t disable_late_binding_show(struct file *f, char __user
> > *ubuf,
> > +                                        size_t size, loff_t *pos)
> > +{
> > +       struct xe_device *xe = file_inode(f)->i_private;
> > +       struct xe_late_bind *late_bind = &xe->late_bind;
> > +       char buf[32];
> > +       int len;
> > +
> > +       len = scnprintf(buf, sizeof(buf), "%d\n", late_bind-
> > >disable);
> > +
> > +       return simple_read_from_buffer(ubuf, size, pos, buf, len);
> > +}
> > +
> > +static ssize_t disable_late_binding_set(struct file *f, const char
> > __user *ubuf,
> > +                                       size_t size, loff_t *pos)
> > +{
> > +       struct xe_device *xe = file_inode(f)->i_private;
> > +       struct xe_late_bind *late_bind = &xe->late_bind;
> > +       u32 uval;
> > +       ssize_t ret;
> > +
> > +       ret = kstrtouint_from_user(ubuf, size, sizeof(uval), &uval);
> > +       if (ret)
> > +               return ret;
> > +
> > +       if (uval > 1)
> > +               return -EINVAL;
> > +
> > +       late_bind->disable = !!uval;
> > +       return size;
> > +}
> > +
> > +static const struct file_operations disable_late_binding_fops = {
> > +       .owner = THIS_MODULE,
> > +       .read = disable_late_binding_show,
> > +       .write = disable_late_binding_set,
> > +};
> > +
> >  void xe_debugfs_register(struct xe_device *xe)
> >  {
> >         struct ttm_device *bdev = &xe->ttm;
> > @@ -249,6 +287,9 @@ void xe_debugfs_register(struct xe_device *xe)
> >         debugfs_create_file("atomic_svm_timeslice_ms", 0600, root,
> > xe,
> >                             &atomic_svm_timeslice_ms_fops);
> >  
> > +       debugfs_create_file("disable_late_binding", 0600, root, xe,
> > +                           &disable_late_binding_fops);
> > +
> >         for (mem_type = XE_PL_VRAM0; mem_type <= XE_PL_VRAM1;
> > ++mem_type) {
> >                 man = ttm_manager_type(bdev, mem_type);
> >  
> > diff --git a/drivers/gpu/drm/xe/xe_late_bind_fw.c
> > b/drivers/gpu/drm/xe/xe_late_bind_fw.c
> > index df43523e9043..88355adce1d0 100644
> > --- a/drivers/gpu/drm/xe/xe_late_bind_fw.c
> > +++ b/drivers/gpu/drm/xe/xe_late_bind_fw.c
> > @@ -167,6 +167,9 @@ int xe_late_bind_fw_load(struct xe_late_bind
> > *late_bind)
> >         if (!late_bind->component_added)
> >                 return -ENODEV;
> >  
> > +       if (late_bind->disable)
> > +               return 0;
> > +
> >         for (fw_id = 0; fw_id < XE_LB_FW_MAX_ID; fw_id++) {
> >                 lbfw = &late_bind->late_bind_fw[fw_id];
> >                 if (lbfw->payload) {
> > diff --git a/drivers/gpu/drm/xe/xe_late_bind_fw_types.h
> > b/drivers/gpu/drm/xe/xe_late_bind_fw_types.h
> > index 5c0574aff7b9..158dc1abe072 100644
> > --- a/drivers/gpu/drm/xe/xe_late_bind_fw_types.h
> > +++ b/drivers/gpu/drm/xe/xe_late_bind_fw_types.h
> > @@ -65,6 +65,8 @@ struct xe_late_bind {
> >         struct workqueue_struct *wq;
> >         /** @component_added: whether the component has been added */
> >         bool component_added;
> > +       /** @disable: to block late binding reload during pm resume
> > flow*/
> > +       bool disable;
> >  };
> >  
> >  #endif
> 

  reply	other threads:[~2025-08-01 15:17 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-10 15:08 [PATCH 0/9] Introducing firmware late binding Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 1/9] mei: bus: add mei_cldev_mtu interface Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 2/9] mei: late_bind: add late binding component driver Rodrigo Vivi
2025-07-16 11:48   ` Greg KH
2025-07-16 11:58     ` Usyskin, Alexander
2025-07-16 12:07       ` Greg KH
2025-07-16 14:26         ` Usyskin, Alexander
2025-07-16 14:45           ` Greg KH
2025-07-27 14:46             ` Usyskin, Alexander
2025-08-13  9:51               ` Nilawar, Badal
2025-08-13 12:13                 ` Greg KH
2025-07-10 15:08 ` [PATCH 3/9] drm/xe/xe_late_bind_fw: Introducing xe_late_bind_fw Rodrigo Vivi
2025-07-25 21:38   ` Lucas De Marchi
2025-07-30 11:47     ` Nilawar, Badal
2025-07-10 15:08 ` [PATCH 4/9] drm/xe/xe_late_bind_fw: Initialize late binding firmware Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 5/9] drm/xe/xe_late_bind_fw: Load " Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 6/9] drm/xe/xe_late_bind_fw: Reload late binding fw in rpm resume Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 7/9] drm/xe/xe_late_bind_fw: Reload late binding fw during system resume Rodrigo Vivi
2025-07-10 15:08 ` [PATCH 8/9] drm/xe/xe_late_bind_fw: Introduce debug fs node to disable late binding Rodrigo Vivi
2025-07-31 20:03   ` Summers, Stuart
2025-08-01 15:17     ` Rodrigo Vivi [this message]
2025-09-05 15:40       ` Nilawar, Badal
2025-07-10 15:08 ` [PATCH 9/9] drm/xe/xe_late_bind_fw: Extract and print version info Rodrigo Vivi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=aIzafHWskNOEcXuC@intel.com \
    --to=rodrigo.vivi@intel.com \
    --cc=alexander.usyskin@intel.com \
    --cc=anshuman.gupta@intel.com \
    --cc=badal.nilawar@intel.com \
    --cc=daniele.ceraolospurio@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stuart.summers@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox