Description of problem:
RHEL5 beta2 will set 1280x1024@75hz to ATI ES1000. This mode is beyond what
the ES1000 can handle when using 32bpp at 200Mhz.
RHEL2 beta2 use xorg-x11-drv-ati-6.6.3-1 version driver. This driver has
enlarged the pixel clock range for ES1000.Thus some large modes that exceed
ES1000 capacity are added to the valid mode list. However, these big modes
will lead to uncertain status of ES1000 at 200Mhz memory clock.
According to ES1000 chip spec. At 200Mhz memory clock ES1000 cannot support
modes bigger than:
please refer to bug #182511 for detailed information:
Version-Release number of selected component (if applicable):
RHEL2 beta2 with xorg-x11-drv-ati-6.6.3-1
Steps to Reproduce:
1.install RHEL5 beta2 on ATI ES1000 system.
2.choose a monitor that can support high modes to the system
3.at 24bpp choose resolution 1280x1024
5.excute xrandr to check the modes
mode 1280x1024@75hz at 24bpp is added to the valid mode list
at 24bpp,200Mhz memory clock, ES1000 cannot handle mode larger than
Created attachment 142459 [details]
limit the rn50 pixel clock range within its capacity
Trivially correct, thanks. Built as xorg-x11-drv-ati 6.6.3-3.el5, should go out
in the next weekly snapshot.
This patch caused a regression on several machines here in RDU. See bug 211504.
Hi, Jay, I am not authorized to see bug 211504, would you please do me a favor
to grant it to me?
as a matter of fact, in our patch, we only change the empirical value from 24
to 18, the code is in function RADEONGetClockInfo of radeon_driver.c :
if (info->ChipFamily == CHIP_FAMILY_RV100 &&!info->HasCRTC2)
nothing else has been changed in our patch. Would it possible you just
mannually change the empirical value and then build the driver again?
no right to access bug 211504
Dell would like to know if the MODIFIED state of this issue is still accurate ?
We do see the rn50-pixel-clock-limit.patch in the RCS6 which contains
xorg-x11-drv-ati-6.6.3-3.2.el5. Question is, does this resolve the issue ??
Since this issue was found and fixed in RHEL3 and RHEL4 (caused a lot of
heart-ache), we would hate to see this resurrect right back into RHEL5. Hence
raising severity to get the right visibility.
Can the involved parties: Alan J, Jay T, please comment on this ASAP since
time's running out.
xorg-x11-drv-ati-6.6.3-3.2.el5 resolves the issue. thanks for the cooperation
from you all.
Setting to VERIFIED based on Comment #7.
xorg-x11-drv-ati-6.6.3-3.2.el5 included in 20070124.1.