Message ID | 20200122125750.9737-1-wambui.karugax@gmail.com (mailing list archive) |
---|---|
Headers | show |
Series | drm/i915/gem: conversion to new drm logging macros | expand |
Quoting Wambui Karuga (2020-01-22 12:57:48) > This series is a part of the conversion to the new struct drm_device > based logging macros in drm/i915. > This series focuses on the drm/i915/gem directory and converts all > straightforward instances of the printk based logging macros to the new > macros. Overall, I'm not keen on this as it perpetuates the mistake of putting client debug message in dmesg and now gives them even more an air of being device driver debug messages. We need a mechanism by which we report the details of what a client did wrong back to that client (tracefs + context/client getparam to return an isolated debug fd is my idea). > Wambui Karuga (2): > drm/i915/gem: initial conversion to new logging macros using > coccinelle. > drm/i915/gem: manual conversion to struct drm_device logging macros. Still this is a necessary evil for the current situation, Acked-by: Chris Wilson <chris@chris-wilson.co.uk> -Chris
On Sat, Jan 25, 2020 at 04:08:39PM +0000, Chris Wilson wrote: > Quoting Wambui Karuga (2020-01-22 12:57:48) > > This series is a part of the conversion to the new struct drm_device > > based logging macros in drm/i915. > > This series focuses on the drm/i915/gem directory and converts all > > straightforward instances of the printk based logging macros to the new > > macros. > > Overall, I'm not keen on this as it perpetuates the mistake of putting > client debug message in dmesg and now gives them even more an air of > being device driver debug messages. We need a mechanism by which we > report the details of what a client did wrong back to that client > (tracefs + context/client getparam to return an isolated debug fd is my > idea). Sean is working on that, but it's a global thing still. Well since it's tracefs should be easy to filter for a given process at least. We've had long discussion about how to expose that, big fear (especially with atomic) is that clients/compositors will start to look at random debug strings and make them uapi. But I think for stuff like igt we should be able to wire it up easily and get it dumped when things go wrong. Maybe similar when mesa gets an unexpected errno. -Daniel > > Wambui Karuga (2): > > drm/i915/gem: initial conversion to new logging macros using > > coccinelle. > > drm/i915/gem: manual conversion to struct drm_device logging macros. > > Still this is a necessary evil for the current situation, > Acked-by: Chris Wilson <chris@chris-wilson.co.uk> > -Chris
Quoting Daniel Vetter (2020-01-27 09:05:57) > On Sat, Jan 25, 2020 at 04:08:39PM +0000, Chris Wilson wrote: > > Quoting Wambui Karuga (2020-01-22 12:57:48) > > > This series is a part of the conversion to the new struct drm_device > > > based logging macros in drm/i915. > > > This series focuses on the drm/i915/gem directory and converts all > > > straightforward instances of the printk based logging macros to the new > > > macros. > > > > Overall, I'm not keen on this as it perpetuates the mistake of putting > > client debug message in dmesg and now gives them even more an air of > > being device driver debug messages. We need a mechanism by which we > > report the details of what a client did wrong back to that client > > (tracefs + context/client getparam to return an isolated debug fd is my > > idea). > > Sean is working on that, but it's a global thing still. Go look at how I suggest we can use tracefs in that thread :) -Chris
On Mon, Jan 27, 2020 at 09:08:01AM +0000, Chris Wilson wrote: > Quoting Daniel Vetter (2020-01-27 09:05:57) > > On Sat, Jan 25, 2020 at 04:08:39PM +0000, Chris Wilson wrote: > > > Quoting Wambui Karuga (2020-01-22 12:57:48) > > > > This series is a part of the conversion to the new struct drm_device > > > > based logging macros in drm/i915. > > > > This series focuses on the drm/i915/gem directory and converts all > > > > straightforward instances of the printk based logging macros to the new > > > > macros. > > > > > > Overall, I'm not keen on this as it perpetuates the mistake of putting > > > client debug message in dmesg and now gives them even more an air of > > > being device driver debug messages. We need a mechanism by which we > > > report the details of what a client did wrong back to that client > > > (tracefs + context/client getparam to return an isolated debug fd is my > > > idea). > > > > Sean is working on that, but it's a global thing still. > > Go look at how I suggest we can use tracefs in that thread :) Hm I think we're a few threads further already? Steven Rostedt has jumped in now too ... -Daniel
On Sat, 25 Jan 2020, Chris Wilson <chris@chris-wilson.co.uk> wrote: > Quoting Wambui Karuga (2020-01-22 12:57:48) >> This series is a part of the conversion to the new struct drm_device >> based logging macros in drm/i915. >> This series focuses on the drm/i915/gem directory and converts all >> straightforward instances of the printk based logging macros to the new >> macros. > > Overall, I'm not keen on this as it perpetuates the mistake of putting > client debug message in dmesg and now gives them even more an air of > being device driver debug messages. We need a mechanism by which we > report the details of what a client did wrong back to that client > (tracefs + context/client getparam to return an isolated debug fd is my > idea). I don't disagree, but I also don't think this makes things (much) worse in that regard. > >> Wambui Karuga (2): >> drm/i915/gem: initial conversion to new logging macros using >> coccinelle. >> drm/i915/gem: manual conversion to struct drm_device logging macros. > > Still this is a necessary evil for the current situation, > Acked-by: Chris Wilson <chris@chris-wilson.co.uk> Thanks, pushed both. BR, Jani.