src - OpenBSD base system

Age	Commit message (Collapse)	Author
5 days	Use imsgbuf_clear() where appropriate instead of msgbuf_clear().	Claudio Jeker
	OK tb@
5 days	Rename imsg_init, imsg_clear, imsg_read, imsg_write and imsg_flush to	Claudio Jeker
	imsgbuf_init, imsgbuf_clear, imsgbuf_read, imsgbuf_write and imsgbuf_flush. This separates the imsgbuf API from the per-imsg API. OK tb@
5 days	Convert imsg_write() callers to the new simplified return logic.	Claudio Jeker
	OK tb@
5 days	Convert imsg_write() callers to the new simplified return logic.	Claudio Jeker
	OK tb@
5 days	Convert ibuf_write() callers to new simplified return logic.	Claudio Jeker
	OK tb@
5 days	Convert msgbuf_write() callers to the new return logic.	Claudio Jeker
	OK tb@
5 days	Use ibuf_write() to send BGP messages to peers instead of msgbuf_write().	Claudio Jeker
	This code does not need to pass fds so there is no need to call msgbuf_write. OK tb@
5 days	Introduce imsg_write() and use it instead of msgbuf_write().	Claudio Jeker
	imsg_write() is just a thin wrapper around msgbuf_write(). So this is mostly search and replace. OK tb@
5 days	Use imsg_flush() instead of implementing the same by hand.	Claudio Jeker
	OK tb@
5 days	imsg_flush no longer returns EAGAIN. Simplify proc_flush_imsg() accordingly.	Claudio Jeker
	Issue spotted by op@
5 days	ibuf_get_string() was added to the ibuf API, remove local version	Claudio Jeker
	OK tb@
5 days	Update ibuf and imsg manpage to include all new functions.	Claudio Jeker
	OK tb@
5 days	Add ibuf_read and msgbuf_read to read from a socket and msgbuf_get	Claudio Jeker
	to return the ibufs generated by the previous two functions. Error out if the hdrsz argument in msgbuf_new_reader is 0 or too big. Also check that the rbuf is allocated in ibuf_read and msgbuf_read. If not return EINVAL. Implement the imsg API using these functions and introduce imsgbuf_set_maxsize() to alter the maximum message size and imsgbuf_allow_fdpass() to allow fd passing (which is now off by default). Also cleanup the internals a bit and make imsgbuf_init() return int. OK tb@
5 days	Make struct msgbuf opaque. Introduce msgbuf_new() and msgbuf_free() for that.	Claudio Jeker
	This does not yet fix the imsgbuf_init() function which can now error. OK tb@
5 days	Remove fd from struct msgbuf, instead pass the fd to imsg_write and	Claudio Jeker
	msgbuf_write OK tb@
5 days	Reshuffle code and rename some internal functions.	Claudio Jeker
	OK tb@
5 days	Retire ibuf_add_buf, replaced by better named ibuf_add_ibuf	Claudio Jeker
	OK tb@
5 days	Only ldapd tried to handle fd exhaustion during imsg fd passing.	Claudio Jeker
	Remove the getdtablecount code from imsgbuf_read() and instead move the getdtablecount code into ldapd. Handle EMSGSIZE (the error returned when there are not enough free file descriptor slots) inside imsgbuf_read() by retrying the read to receive the data but without fd. A caller expecting an fd should check the return value of imsg_get_fd. OK tb@
5 days	Adjust imsgbuf_read to return an int with either 1, 0 or -1.	Claudio Jeker
	Handle EAGAIN by a simple return 1 (same for the fd check). This way the caller will poll again and then retry later. OK tb@
5 days	Introduce imsgbuf_queuelen() which returns the number	Claudio Jeker
	of messages ready for transmission. Returns 0 if nothing is pending. OK tb@
5 days	Rename imsg_init, imsg_clear, imsg_read, imsg_write and imsg_flush	Claudio Jeker
	to imsgbuf_init, imsgbuf_clear, imsgbuf_read, imsgbuf_write and imsgbuf_flush. This separates the imsgbuf API from the per-imsg API. OK tb@
5 days	Simplify imsg_write, msgbuf_write and ibuf_write return codes.	Claudio Jeker
	Return 0 on success or when a temporary error happened (EAGAIN, ENOBUFS). Return -1 on error and set errno otherwise. Kill the old 0 return for EOF. This is not how write operations work. OK tb@
5 days	Introduce imsg_write() and use it instead of msgbuf_write().	Claudio Jeker
	This is just a thin wrapper around msgbuf_write() but it makes the API more consistent. OK tb@
5 days	imsg_flush() should retry the write on EAGAIN errors.	Claudio Jeker
	OK tb@
5 days	Refine imsg_forward(), do not close fd, just don't forward it.	Claudio Jeker
	imsg_free() will close the unclaimed fds at the end. OK tb@
5 days	Use ibuf_fd_set() and ibuf_fd_get() to handle the fd passing in imsgs.	Claudio Jeker
	ibufs already close forgotten fds on free so now imsg_free behaves the same way. OK tb@
5 days	Allow 0 sized ibuf_open() and mark stack buffers by setting the fd to -2.	Claudio Jeker
	For imsgs we want to be able to use ibufs even for empty messages and stash the fd into those ibufs. For that adjust the ibuf code to allow that. This adds an internal IBUF_FD_MARK_ON_STACK define that is now used for on stack ibufs instead of setting max to 0. OK tb@
5 days	nline ibuf_realloc() in ibuf_reserve() it is the only	Claudio Jeker
	user and the function is simple enough. OK tb@
5 days	Add ibuf_get_string() to the ibuf API (used by bgpd and xlockmore)	Claudio Jeker
	OK tb@
5 days	Let the user provide an alternative perfpolicy when on battery	Jeremie Courreges-Anglas
	The current behavior of "auto", which implies running at full speed when on AC power, does not fit all the hardware and use cases. For some people it results in more power consumption, more heat, more noise, etc. Extend the semantics of hw.perfpolicy and provide two buttons to specify the desired behavior: sysctl hw.perfpolicy=<policy while on ac>[,<policy while on battery>] Keep the default behavior of "high,auto". People can opt for "auto,auto" or simply "auto" instead. No objection from deraadt@, input and ok sobrado@ sthen@
5 days	The block size of HMAC-SHA2-{384,512} must be 128 bytes.	YASUOKA Masahiko
	ok tb
5 days	Fix save-buffer documentation - it writes to stdout. From Ilya Grigoriev.	Nicholas Marriott

5 days	When the mouse clicked on the border between two horizontal panes, the	Nicholas Marriott
	location was not being set properly. Pulled the checking of this into separate function for clarity. Fixes dragging on horizontal pane border. From Michael Grant.
5 days	Update lang/ruby module documentation for the removal of Ruby 3.1	Jeremy Evans

5 days	The buffer flipper (incorrectly?) uses pmap_copy_page() from interrupt	Mark Kettenis
	context (when it calls uvm_pagerealloc_multi()). But the current implementation of pmap_copy_page() assumes it only runs in process context. Use splbio() to block the interrupts while we're doing the copy. Same diff as the one committed to arm64 a bit over a week ago. ok mpi@, jca@
5 days	Fix word navigation on lines with tabs, from Alexander Arch.	Nicholas Marriott

6 days	pfctl: clear statistic for specified addresses	kirill
	OK: bluhm@
6 days	Improve error handling of psp(4) driver.	Alexander Bluhm
	The driver for the AMD platform security processor did use a bunch of EINVAL error codes. It is better to propagate the error up from the subsystem or use more specific values. Rename the variable "ret" to "error" to make clear where we deal with an actual errno(2). Document where and why the error code is ignored. OK hshoexer@
6 days	No need to call unsleep() if p_wchan is NULL.	Claudio Jeker
	OK mpi@
6 days	sync	Stuart Henderson

6 days	use ifq_deq_set_oactive if we're in an if_deq_begin "transaction"	David Gwynne
	reported by and fix tested by afresh@
6 days	provide ifq_deq_set_oactive.	David Gwynne
	ifq_deq_set_oactive is a variation on ifq_set_oactive that can be called inside an if_deq_begin "transaction". afresh@ found de(4) was calling ifq_set_oactive while holding the ifq mutex via ifq_deq_begin, which led to a panic because ifq_set_oactive also tries to take the ifq mutex. ifq_deq_set_oactive assumes the caller is already holding the mutex. de(4) is confusing, so it seemed simpler to add a small tweak to ifqs than try and do major surgery on such a hairy driver. tested by afresh@
6 days	use a tailq for the global list of bpf_if structs.	David Gwynne
	this replaces a hand rolled list that's been here since 1.1. ok claudio@ kn@ tb@
7 days	fix SEE ALSO;	Jason McIntyre

7 days	Add qccpucp(4) manual page	Tobias Heider
	ok mlarkin@
7 days	gather ice(4) hardware counters for debugging	Stefan Sperling

7 days	implement ice_add_rx_lldp_filter()	Stefan Sperling

7 days	remove forward declaration of struct mount	Jonathan Gray
	unneeded after uvm_vnp_sync() prototype moved to vnode.h in 2014
7 days	hook up ddb regress	Anton Lindqvist

7 days	Add amd64 ddb regress suite allowing the disassembler to be tested from	Anton Lindqvist
	userspace.