summaryrefslogtreecommitdiff
path: root/sys/dev/ipmi.c
AgeCommit message (Collapse)Author
2008-02-16Call sensordev_install() iff we find and sensor_attach() the individual sensors.Constantine A. Murenin
The call got misplaced in ipmi.c#rev1.59 on 2007-08-31. ok marco; tested todd
2007-11-25KNF and remove a bad free right before kthread_exitMarco Peereboom
2007-11-01correct ipmi attach time failure messages (why this one machine fails...Theo de Raadt
still has to be looked at by marco when he has time)
2007-10-08More simple memset(,0,) -> M_ZERO changes. In this batch move toKenneth R Westerback
size(*p) as the first malloc() parameter where p is declared locally and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c even though there is no obvious bzero or memset nearby.
2007-08-31Defer ipmi sensor creation after kernel goes multi user. This removes theMarco Peereboom
annoying waiting period during kernel boots. pushed for by deraadt and beck Lots of testing todd, ckuethe and others ok beck
2007-05-29Correct fix for the last backout use strlcpy() to fill dv_xname. OK beck@Claudio Jeker
2007-05-29back out last commit. It helps to compile before commiting. OK beck@ art@Claudio Jeker
2007-05-29work around really bad softc abuse by marco; proper fixes for this willTheo de Raadt
be part of a later rototilling
2007-04-13move the debugging messages into #ifdef IPMI_DEBUG, we do not need toReyk Floeter
compile them by default. ok marco@
2007-04-08x4100 machines need a hold off in between reads and write to the bmcMarco Peereboom
2007-03-22split userland & kernel struct sensor/sensordev so that the additionTheo de Raadt
of new fields in the future is less disruptive. This is done similar to how struct proc is handled for ps(1). ok jmc (man page changes) tested fkr simon, and more suggestions from millert
2006-12-23adapt to new two-level sensor api; Constantine A. MureninTheo de Raadt
2006-12-23if errors happen while reading the sdr list, give up on handling theTheo de Raadt
device as cleanly as possible. ok marco
2006-12-23wrong printing; ok marcoTheo de Raadt
2006-09-19fix all malloc()s to nowait|canfail and do not ignore errors reading sensor ↵Michael Shalayeff
status; marco@ ok
2006-09-17Move get ipmi id into the match function. This prevents attach from beingMarco Peereboom
called whenever there is no BMC and hence suppresses a ugly message when a bios lied that it had ipmi. tested beck@
2006-08-01Add a wait between send and receive command. Talked through with jordan.Marco Peereboom
2006-08-01Silence intermittent read errors.Marco Peereboom
Fix a bug in bt where the "get id" command fails by giving the bmc some extra time to gather the data.
2006-07-28KNFMarco Peereboom
2006-07-28Only read 1 sensor at a time instead of all of them. This should helpMarco Peereboom
systems that have a burst of activity every 10 seconds. Also lower the polling frequency from 10 to 5 seconds. ok jordan@
2006-07-28Fix BT interface by waiting for BMC_BUSY on writes.Jordan Hargrave
Found and ok marco
2006-05-21Replace lockmgr with rwlock; help and ok marco@Aleksander Piotrowski
2006-05-15Add device name when printing out errors. Pointed out by deraadt@Marco Peereboom
2006-05-15Make _bmc_io_wait 1 second instead of 500ms.Marco Peereboom
Prevent multiple reads and writes to the bmc at the same time which could interrupt a complete bmc transaction. tested by various people.
2006-05-12code we can readTheo de Raadt
2006-05-08Add smbios support for i386 and amd64, fix ipmi to use this new functionallity,Gordon Willem Klok
hook up some sysctls to add system vendor/product/version and UUID reporting. "get it in" deraadt@
2006-05-06if the fan data rate is per second convert it to per minute, ok marcoPeter Stromberg
2006-04-27from PAE work:Michael Shalayeff
since we are only interested in the page offset bits cast paddr_t to (smaller) u_long to shuddup gcc (;
2006-04-20Fix various printf() issues: too many arguments, not enough arguments, argumentMiod Vallat
order reversed, wrong modifiers. ok deraadt@ marco@ mickey@
2006-01-29Disable iic whenever ipmi is enabled. This fixes boxes like Sun x4200 thatMarco Peereboom
freak out when more than one device driver touch the i2c devices. Reported by: Srebrenko Sehic <ssehic at gmail dot com> ok kettenis@
2006-01-19Replace SENSOR_ADD() macro with a pair of functionsAlexander Yurchenko
sensor_add()/sensor_del() so that sensors can be attached and detached dynamicaly. ok kettenis@ deraadt@ dlg@
2006-01-18Fix false positives of ipmi on older machines.Marco Peereboom
ok jordan@ tested by others
2006-01-05Kill all typedef struct goo.Marco Peereboom
2006-01-05Work around flawed IBM 325 eServer IPMI implementation.Marco Peereboom
Help and ok deraadt@ jordan@
2006-01-05Better debug verbosity values.Marco Peereboom
2006-01-05Adjust debug levels.Marco Peereboom
2006-01-05I got the time scale wrong with the previous commit. Now the advertisedMarco Peereboom
50ms timeout is actually 50ms during runtime.
2006-01-04Adjust timeout values. This was determined by using beck's ibm e325.Marco Peereboom
ok jordan
2005-12-16Fix 2 issues reported by beck.Marco Peereboom
1. Sun v20z rebooted dunring autoconf. This was due to reading a byte and discarding it at an inapropriate time. 2. The Sun is so fast that it could exhaust the busy counter. This has been replaced with a timeout/tsleep mechanism that is more predictable. As a bonus we no longer busy wait in ipmi which should improve overall throughput. ok beck@ jordan@ dlg@
2005-12-13Add check for return value in recvcmd.Marco Peereboom
Unmap ipmi registers if attach fails. ok jordan@
2005-12-13no need for device name before failTheo de Raadt
2005-12-13Don't continue to poke ipmi registers if the ID command fails.Marco Peereboom
ok jordan@
2005-11-30spacingTheo de Raadt
2005-11-28Added ipmi watchdog code, based on esm watchdog code by dlg@Jordan Hargrave
Added signextend function to cleanup conversion routine ok marco@
2005-11-22Some BIOS' lie about location of ports. Prevent panic by checking ↵Marco Peereboom
bus_space_map. This works around bob's ipmi panic issue on the amd64 e325 IBM box. ok jordan@
2005-10-24Remove panic if polling thread can not be created.Marco Peereboom
ok jordan@
2005-10-21pretty up dmesg outputTheo de Raadt
2005-10-20KNF; marco okTheo de Raadt
2005-10-19Fixup case statements to use #define'sJordan Hargrave
suggested by deraadt ok marco@
2005-10-19Added readable case names for ipmi device typesJordan Hargrave
ok marco@