Re: sparc64 panics - ctx_free: context xxxx still active
Claudio Jeker <[email protected]>
| Newsgroups | gmane.os.openbsd.bugs |
|---|---|
| Message-ID | <[email protected]> |
On Tue, Apr 14, 2026 at 11:56:30AM +0200, Mark Kettenis wrote: > > Date: Sun, 12 Apr 2026 22:39:14 +0200 > > From: Claudio Jeker <[email protected]> > > > > Running with this for some time now doing the perl FPU torture. > > > > ctx_free: context 7952 still active in DMMU TSB > > ctx_free: tag 0x1f10000000008845 data 0x8000000ed2b42490 > > ctx_free: context 994 still active in DMMU TSB > > ctx_free: tag 0x03e200000001fd88 data 0x8000000ed2b42490 > > So interesting that the TSB data is the same for both. These are > mappings for the same physical page. And it is a PROT_EXEC mapping > (sun4v uses a combined D/I-TLB, so it shows up in the DMMU). Do you > have more of these? Is the data always the same? Until you reboot at > least? > > I wonder if this is the sigcode page that doesn't get cleaned up... ctx_free: tag 0x0ae100000003529c data 0x8000000ed2b42490 ctx_free: context 3200 still active in DMMU TSB ctx_free: tag 0x0c8000000003a8e6 data 0x8000000ed2b42490 ctx_free: context 5918 still active in DMMU TSB ctx_free: tag 0x171e0000000157f0 data 0x8000000ed2b42490 ctx_free: context 5558 still active in DMMU TSB ctx_free: tag 0x15b600000003639f data 0x8000000ed2b42490 ctx_free: context 5168 still active in DMMU TSB ctx_free: tag 0x143000000001ee2b data 0x8000000ed2b42490 ctx_free: context 2672 still active in DMMU TSB ctx_free: tag 0x0a70000000005e0b data 0x8000000ed2b42490 ctx_free: context 6315 still active in DMMU TSB ctx_free: tag 0x18ab000000038b2b data 0x8000000ed2b42490 ctx_free: context 1328 still active in DMMU TSB ctx_free: tag 0x053000000001e302 data 0x8000000ed2b42490 ctx_free: context 6830 still active in DMMU TSB ctx_free: tag 0x1aae00000002c805 data 0x8000000ed2b42490 ctx_free: context 3546 still active in DMMU TSB ctx_free: tag 0x0dda000000037313 data 0x8000000ed2b42490 ctx_free: context 2574 still active in DMMU TSB ctx_free: tag 0x0a0e0000000075f8 data 0x8000000ed2b42490 ctx_free: context 6470 still active in DMMU TSB ctx_free: tag 0x1946000000009e06 data 0x8000000ed2b42490 ctx_free: context 7904 still active in DMMU TSB ctx_free: tag 0x1ee000000002068d data 0x8000000ed2b42490 ctx_free: context 89 still active in DMMU TSB ctx_free: tag 0x0059000000019d6d data 0x8000000ed2b42490 ctx_free: context 4783 still active in DMMU TSB ctx_free: tag 0x12af00000003366f data 0x8000000ed2b42490 ctx_free: context 1995 still active in DMMU TSB ctx_free: tag 0x07cb000000033e48 data 0x8000000ed2b42490 Here a few more. It is indeed always the same entry. The sigcode page is an interesting idea. I was wondering if this was some CoW mapping that triggers an update of a vmpage at the same time. -- :wq Claudio