sna: CustomEDID fix
A message that is not in the archive (sent off-list or not yet received): 16051285.77697168.1517513962156.JavaMail.root@zimbra33-e6
PATCH 1/1 sna: CustomEDID fix
Hello,
For my HTPC setup, I'm using the option "CustomEDID".
With this option, output attaching and destroying events leads to crashes.
The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
- Starting Xorg
- Connect HDMI2
- Disconnect HDMI2
- Reconnect HDMI2
-> Crash
The crash happens in xf86OutputSetEDID (xorg/xserver/hw/xfree86/modes/xf86Crtc.c) at "free(output->MonInfo)". MonInfo is assigned with sna_output->fake_edid_mon which is allocated by intel driver in sna_output_load_fake_edid (src/sna/sna_display.c).
Sequence details:
- Starting Xorg
-> fake_edid_mon is initialized
- Connect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is NULL
- MonInfo is assigned with fake_edid_mon pointer
- MonInfo is read by Xorg
- Disconnect HDMI2
- Reconnect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is freed thus also fake_edid_mon
- MonInfo is assigned with fake_edid_mon
- MonInfo is read but it was freed -> CRASH
The fix consists of a new instance of xf86MonPtr for each calls of xf86OutputSetEDID. This instance is initialized with fake_edid_raw which render fake_edid_mon useless. With this proposal, the behaviour of an EDID override is similar to a "real" EDID.
Regards,
To: Chris Wilson <chris at chris-wilson.co.uk>
src/sna/sna_display.c +0 -0
@@ -263,7 +263,6 @@ struct sna_output {
uint32_t edid_blob_id;
uint32_t edid_len;
void *edid_raw;
- xf86MonPtr fake_edid_mon;
void *fake_edid_raw;
bool has_panel_limits;
@@ -4102,13 +4101,21 @@ static DisplayModePtr sna_output_override_edid(xf86OutputPtr output)
{
struct sna_output *sna_output = output->driver_private;
+ xf86MonPtr mon = NULL;
+
+ if (sna_output->fake_edid_raw == NULL)
+ return NULL;
- if (sna_output->fake_edid_mon == NULL)
+ mon = xf86InterpretEDID(output->scrn->scrnIndex, sna_output->fake_edid_raw); + if (mon == NULL) {
return NULL;
+ }
+
+ mon->flags |= MONITOR_EDID_COMPLETE_RAWDATA;
+
+ xf86OutputSetEDID(output, mon);
- xf86OutputSetEDID(output, sna_output->fake_edid_mon);
- return xf86DDCGetModes(output->scrn->scrnIndex,
- sna_output->fake_edid_mon);
+ return xf86DDCGetModes(output->scrn->scrnIndex, mon);
}
static DisplayModePtr
@@ -4896,7 +4903,6 @@ sna_output_load_fake_edid(xf86OutputPtr output) FILE *file;
void *raw;
int size;
- xf86MonPtr mon;
filename = fake_edid_name(output);
if (filename == NULL)
@@ -4928,16 +4934,6 @@ sna_output_load_fake_edid(xf86OutputPtr output) }
fclose(file);
- mon = xf86InterpretEDID(output->scrn->scrnIndex, raw);
- if (mon == NULL) {
- free(raw);
- goto err;
- }
-
- if (mon && size > 128)
- mon->flags |= MONITOR_EDID_COMPLETE_RAWDATA;
-
- sna_output->fake_edid_mon = mon;
sna_output->fake_edid_raw = raw;
xf86DrvMsg(output->scrn->scrnIndex, X_CONFIG,
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 5 lines · On Fri, 02 Feb 2018, [email protected] wrote:
For my HTPC setup, I'm using the option "CustomEDID".
With this option, output attaching and destroying events leads to crashes.The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
I know nothing about xorg customedid stuff.
But there's drm.edid_firmware=HDMI-2:/etc/my_edid.bin (or drm_kms_helper.edid_firmware on older kernels) module parameter for the kernel that does this transparently across the stack.
BR,
Jani.
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting Jani Nikula (2018-02-02 19:13:53)
Quoted 12 lines · On Fri, 02 Feb 2018, [email protected] wrote:
Quoted 5 lines
For my HTPC setup, I'm using the option "CustomEDID".
With this option, output attaching and destroying events leads to crashes.The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"I know nothing about xorg customedid stuff.
But there's drm.edid_firmware=HDMI-2:/etc/my_edid.bin (or drm_kms_helper.edid_firmware on older kernels) module parameter for the kernel that does this transparently across the stack.
There's also the difference in that Xorg only uses the custom EDID to conveniently load a set of modelines, but the kernel EDID override describes the sink completely (e.g. audio modes, link protocols, pixel formats etc).
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting [email protected] (2018-02-02 18:37:12)
Quoted 14 lines
Hello,
For my HTPC setup, I'm using the option "CustomEDID".
With this option, output attaching and destroying events leads to crashes.The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
- Starting Xorg
- Connect HDMI2
- Disconnect HDMI2
- Reconnect HDMI2
-> Crash
Quoted 3 lines
The crash happens in xf86OutputSetEDID (
xorg/xserver/hw/xfree86/modes/xf86Crtc.c) at "free(output->MonInfo)". MonInfo is assigned withsna_output->fake_edid_monwhich is allocated by intel driver insna_output_load_fake_edid(src/sna/sna_display.c).
Ho hum, looks like I never tested it more than a quick start of Xorg. Sorry about that :(
Quoted 23 lines
Sequence details:
- Starting Xorg
->fake_edid_monis initialized- Connect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is NULL
- MonInfo is assigned withfake_edid_monpointer
- MonInfo is read by Xorg- Disconnect HDMI2
- Reconnect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is freed thus alsofake_edid_mon
- MonInfo is assigned withfake_edid_mon
- MonInfo is read but it was freed -> CRASHThe fix consists of a new instance of xf86MonPtr for each calls of xf86OutputSetEDID. This instance is initialized with
fake_edid_rawwhich renderfake_edid_monuseless. With this proposal, the behaviour of an EDID override is similar to a "real" EDID.
Quoted 5 lines
Regards,
Signed-off-by Dominique Constant <[email protected]>To: Chris Wilson <chris at chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk> -Chris
_______________________________________________
Intel-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting [email protected] (2018-02-02 18:37:12)
Quoted 42 lines
Hello,
For my HTPC setup, I'm using the option "CustomEDID".
With this option, output attaching and destroying events leads to crashes.The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
- Starting Xorg
- Connect HDMI2
- Disconnect HDMI2
- Reconnect HDMI2
-> CrashThe crash happens in xf86OutputSetEDID (
xorg/xserver/hw/xfree86/modes/xf86Crtc.c) at "free(output->MonInfo)". MonInfo is assigned withsna_output->fake_edid_monwhich is allocated by intel driver insna_output_load_fake_edid(src/sna/sna_display.c).Sequence details:
- Starting Xorg
->fake_edid_monis initialized- Connect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is NULL
- MonInfo is assigned withfake_edid_monpointer
- MonInfo is read by Xorg- Disconnect HDMI2
- Reconnect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is freed thus alsofake_edid_mon
- MonInfo is assigned withfake_edid_mon
- MonInfo is read but it was freed -> CRASHThe fix consists of a new instance of xf86MonPtr for each calls of xf86OutputSetEDID. This instance is initialized with
fake_edid_rawwhich renderfake_edid_monuseless. With this proposal, the behaviour of an EDID override is similar to a "real" EDID.
Quoted 5 lines
Regards,
Signed-off-by Dominique Constant <[email protected]>To: Chris Wilson <chris at chris-wilson.co.uk>
Apologies if this is a resend, but could you send me this patch using "git send-email" (or attach the output of "git format-patch"). git am is not happy with it.
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 55 lines
Quoting [email protected] (2018-02-02 18:37:12)
Quoted 52 lines
Hello,
For my HTPC setup, I'm using the option "CustomEDID". With this option, output attaching and destroying events leads to crashes.
The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
- Starting Xorg
- Connect HDMI2
- Disconnect HDMI2
- Reconnect HDMI2
-> CrashThe crash happens in xf86OutputSetEDID (
xorg/xserver/hw/xfree86/modes/xf86Crtc.c)
at "free(output->MonInfo)". MonInfo is assigned withsna_output->fake_edid_mon
which is allocated by intel driver insna_output_load_fake_edid(src/sna/sna_display.c).Sequence details:
- Starting Xorg
->fake_edid_monis initialized- Connect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is NULL
- MonInfo is assigned withfake_edid_monpointer
- MonInfo is read by Xorg- Disconnect HDMI2
- Reconnect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is freed thus alsofake_edid_mon
- MonInfo is assigned withfake_edid_mon
- MonInfo is read but it was freed -> CRASHThe fix consists of a new instance of xf86MonPtr for each calls of xf86OutputSetEDID.
This instance is initialized withfake_edid_rawwhich renderfake_edid_monuseless.
With this proposal, the behaviour of an EDID override is similar to a "real" EDID.Regards,
Quoted 8 lines
Quoted 2 lines
Signed-off-by Dominique Constant <[email protected]>To: Chris Wilson <chris at chris-wilson.co.uk>
Apologies if this is a resend, but could you send me this patch using "git send-email" (or attach the output of "git format-patch"). git am is not happy with it.
-Chris
Hello,
Thank you for taking into consideration this patch.
Regards,
Dominique CONSTANT
--- src/sna/sna_display.c | 28 ++++++++++++---------------- 1 file changed, 12 insertions(+), 16 deletions(-)
src/sna/sna_display.c +12 -16
@@ -263,7 +263,6 @@ struct sna_output {uint32_t edid_blob_id;uint32_t edid_len;void *edid_raw; -xf86MonPtr fake_edid_mon;void *fake_edid_raw;bool has_panel_limits;
@@ -4102,13 +4101,21 @@ static DisplayModePtr sna_output_override_edid(xf86OutputPtr output) {struct sna_output *sna_output = output->driver_private; +xf86MonPtr mon =NULL ; + +if (sna_output->fake_edid_raw ==NULL ) +return NULL ; -if (sna_output->fake_edid_mon ==NULL ) + mon =xf86InterpretEDID (output->scrn->scrnIndex, sna_output->fake_edid_raw); +if (mon ==NULL ) {return NULL ; + } + + mon->flags |=MONITOR_EDID_COMPLETE_RAWDATA ; + +xf86OutputSetEDID (output, mon); -xf86OutputSetEDID (output, sna_output->fake_edid_mon); -return xf86DDCGetModes (output->scrn->scrnIndex, - sna_output->fake_edid_mon); +return xf86DDCGetModes (output->scrn->scrnIndex, mon); } static DisplayModePtr
@@ -4896,7 +4903,6 @@ sna_output_load_fake_edid(xf86OutputPtr output)FILE *file;void *raw;int size; -xf86MonPtr mon; filename =fake_edid_name (output);if (filename ==NULL )
@@ -4928,16 +4934,6 @@ sna_output_load_fake_edid(xf86OutputPtr output) }fclose (file); - mon =xf86InterpretEDID (output->scrn->scrnIndex, raw); -if (mon ==NULL ) { -free (raw); - goto err; - } - -if (mon && size >128 ) - mon->flags |=MONITOR_EDID_COMPLETE_RAWDATA ; - - sna_output->fake_edid_mon = mon; sna_output->fake_edid_raw = raw; xf86DrvMsg(output->scrn->scrnIndex, X_CONFIG,
Signature · 7 lines
2.15.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting [email protected] (2018-02-06 18:57:05)
Quoted 65 lines
Quoted 55 lines
Quoting [email protected] (2018-02-02 18:37:12)
Quoted 52 lines
Hello,
For my HTPC setup, I'm using the option "CustomEDID". With this option, output attaching and destroying events leads to crashes.
The following sequence leads to a crash:
- In xorg.conf: Option "CustomEDID" "HDMI2:/etc/my_edid.bin"
- Starting Xorg
- Connect HDMI2
- Disconnect HDMI2
- Reconnect HDMI2
-> CrashThe crash happens in xf86OutputSetEDID (
xorg/xserver/hw/xfree86/modes/xf86Crtc.c)
at "free(output->MonInfo)". MonInfo is assigned withsna_output->fake_edid_mon
which is allocated by intel driver insna_output_load_fake_edid(src/sna/sna_display.c).Sequence details:
- Starting Xorg
->fake_edid_monis initialized- Connect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is NULL
- MonInfo is assigned withfake_edid_monpointer
- MonInfo is read by Xorg- Disconnect HDMI2
- Reconnect HDMI2
-> xf86OutputSetEDID is called:
- MonInfo is freed thus alsofake_edid_mon
- MonInfo is assigned withfake_edid_mon
- MonInfo is read but it was freed -> CRASHThe fix consists of a new instance of xf86MonPtr for each calls of xf86OutputSetEDID.
This instance is initialized withfake_edid_rawwhich renderfake_edid_monuseless.
With this proposal, the behaviour of an EDID override is similar to a "real" EDID.Regards,
Quoted 8 lines
Quoted 2 lines
Signed-off-by Dominique Constant <[email protected]>To: Chris Wilson <chris at chris-wilson.co.uk>
Apologies if this is a resend, but could you send me this patch using "git send-email" (or attach the output of "git format-patch"). git am is not happy with it.
-Chris
Quoted 2 lines
Hello,
Thank you for taking into consideration this patch.
Finally, got around to picking up patches for -intel. Thank you very much for the fix,
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx