Quoting r. Michael S. Tsirkin <mst at mellanox.co.il>: > What I plan to do is using eea0e11c1f0d6ef89e64182b2f1223a4ca2b74a2 > for a couple of days and see how this works out. Ugh. Unfortunately in that kernel version, the e1000 driver says the eeprom checksum is bad (works fine with 2.6.19-rc3). So, I tried some suspends/resumes and things seem to work, but I won't be able to test it under real use conditions. But maybe its another red herring? Andi, could you maybe look at that commit and tell me whether it could cause troubles with ACPI after suspend/resume even theoretically? -- MST
