Paths

Table of Contentst

-
sys/
-
arm64/arm64/
-
arm64/
24
intr_machdep.c
24
nexus.c
-
dev/fdt/
-
fdt/
24
simplebus.c
-
kern/
24
bus_if.m
-
subr_bus.c
-
sys/
-
bus.h

Add BUS_UNMASK_INTR method
AbandonedPublic
Actions

Authored by wma_semihalf.com on Jul 9 2015, 11:21 AM.

Details

Reviewers

zbb
emaste
andrew
imp
• ian

Summary

This commit allows other subsystem to explicitly unmask

previously configured interrupt on the current CPU.
Before it was not possible. For example, ARM timer should
unmask per-processor interrupts on each secondary CPU the
system boots. It cannot call bus_setup_intr then, because
was already invoked on primary CPU, thus the new method was
added.

Mostly will help to cleanup GIC init_secondary routines and move TMR PPI unmasking to timer drivers rather than using hardcoded values in the interrupt controller (future commits).

The armv8 function arm_unmask_irq is undefined yet and is waiting for https://reviews.freebsd.org/D3029 to process.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

wma_semihalf.com updated this revision to Diff 6811.Jul 9 2015, 11:21 AM

wma_semihalf.com retitled this revision from to Add BUS_UNMASK_INTR method.

wma_semihalf.com updated this object.

wma_semihalf.com edited the test plan for this revision. (Show Details)

wma_semihalf.com added reviewers: emaste, imp, • ian, andrew, zbb.

wma_semihalf.com set the repository for this revision to rS FreeBSD src repository - subversion.

Herald added subscribers: emaste, andrew, imp. · View Herald TranscriptJul 9 2015, 11:21 AM

mmel added a subscriber: mmel.Jul 11 2015, 7:42 PM

Any comments?

andrew mentioned this in D3029: Fix intr_machdep.c for armv8.Jul 13 2015, 9:59 AM

Please send your comments asap. We want to commit this soon.

zbb edited edge metadata.Jul 14 2015, 9:18 AM

zbb added a subscriber: freebsd-arm-list.

Moved arm_unmask_irq here.

With full respect, I don’t think that this new bus method is right solution for our problem(s).
The proposed bus_unmask_intr() cannot be used from outside of secondary CPUs startup context. It cannot be used for SPI, it cannot be used of PPI after scheduler is started on given core.
Why you cannot simply call arm_unmask_irq() directly from driver?

Well, I do not like to change bus interface so fast. What about to think of PPI on each core like separate interrupt with its irq number, counter, handler, and handler argument. This way bus_setup_intr will be called only once for PPI on each core. Of course, some coding is needed in nexus and/or interrupt framework.

Sadly, there is a new interrupt framework waiting for review for long time. Thus, this thing could be already discussed without such hurry.

I guess treating PPIs are separate IRQs number will cause a huge mess. I'm starting to port fbsd on 96 core armv8 platform and the idea of 1536 vectors wasted for PPIs is outrageous.

My point in adding separate interface is to support multi-PIC systems. Calling BUS_UNMASK_INTR with an appropriate resource will allow to pick up the correct PIC controller in a generic way (as I looked into intrng last time it was capable of associating device with specific PIC). Adding the same code to timer driver might be really inelegant, thus require a static map either in the driver or in intr_machdep file.

The proposed bus_unmask_intr() cannot be used from outside of secondary CPUs startup context. It cannot be used for SPI, it cannot be used of PPI after scheduler is started on given core.

Actually, I'm not quite sure I understand your point. Why cannot one use it for SPI (except it might not have a sense) or what the scheduler has to do with PPI? The only risk I see (in current implementation) is that someone calls unmask when irq is not configured.
The freebsd also lacks the fine-grained interrupt blocking mechanism, like, for example, mask/unmask pair exported as bus interface (that would require adding mask call here also). I bet if you ever tried to write a driver for some queer and exotic piece of hardware you wished these functions existed, like in other OSes. Or at least I have...

In D3034#61087, @wma_semihalf.com wrote:

I guess treating PPIs are separate IRQs number will cause a huge mess. I'm starting to port fbsd on 96 core armv8 platform and the idea of 1536 vectors wasted for PPIs is outrageous.

In fact, PPIs are separate interrupts. So if you have 96 core armv8 platform with 16 PPIs on each core, then you have 1536 PPIs. If you have 96 core, then you have probably enough memory to run them all. And I certainly will not call using of 1536 IRQ numbers as wasting in this case. Moreover, 1536 IRQ numbers are allocated only if all of them are used.

Further, with such many cores, I can imagine more than before that same PPI could be served differently on various subsets of cores. However, I do not say that it's necessary to have an IRQ number for each PPI on each core. I just say that it should be so in principle. It could be optimalized according to usage.

Think about PPI statistics for example. Will be PPI interrupts counted on a single counter? (Just do not forget to use atomic increment in that case. I have confirmed that even two cores count badly without atomics.)

Anyhow, I still think that problem with PPIs you desribed can be dealt with either in nexus or in interrupt framework without a need to change bus interface. (You can change the way how bus_setup_intr() and bus_teardown_intr() are implemented for PPI kind of interrupts.)

My point in adding separate interface is to support multi-PIC systems. Calling BUS_UNMASK_INTR with an appropriate resource will allow to pick up the correct PIC controller in a generic way (as I looked into intrng last time it was capable of associating device with specific PIC). Adding the same code to timer driver might be really inelegant, thus require a static map either in the driver or in intr_machdep file.

Note that only root PIC is capable of PPIs and SGIs (IPIs).

However, if you really want to add some new bus interface, I would like to know more about it. For example, what about to add man pages for it? What about to add bus_mask_intr() too? What is difference between it and bus_activate_resource()? What about to add cpuset_t argument to it? (Which, I believe, was Michal's point for using with SPI.) Must be ensured that no cpu migration happens during the call? I have a really bad feeling with the change how it's introduced now.

The proposed bus_unmask_intr() cannot be used from outside of secondary CPUs startup context. It cannot be used for SPI, it cannot be used of PPI after scheduler is started on given core.

Actually, I'm not quite sure I understand your point. Why cannot one use it for SPI (except it might not have a sense) or what the scheduler has to do with PPI? The only risk I see (in current implementation) is that someone calls unmask when irq is not configured.
The freebsd also lacks the fine-grained interrupt blocking mechanism, like, for example, mask/unmask pair exported as bus interface (that would require adding mask call here also). I bet if you ever tried to write a driver for some queer and exotic piece of hardware you wished these functions existed, like in other OSes. Or at least I have...

Thanks! That was really helpful. However I'm still confused about the system I have. I didn't powered up the full configuration yet, but I'm afraid it has two root pics then (one per each 48-cpu partition). Nevertheless, I'm abandoning this review for a while and stick to hacking generic_timer. When I test it on actual hw I will get back to this patch. However I'm still thinking the generic mask/unmask pair might be a useful feature.

I would like to see something like how Linux handles this. When they create interrupts they can mark them as per-core, i.e. they will be unmasked on all cores. I think we could do this with a new flag in bus_setup_intr.

Care would need to be taken to handle devices created both before and after we have enabled the secondary cpus as I would expect the code paths to be slightly different, e.g. before enabling we would need to record the need to unmask when new cpus are brought up, after enabling we will also need to signal to the other cpus to unmask the interrupt.

Revision Contents
Changeset List

Path

Size

sys/

arm64/

intr_machdep.c

7 lines

nexus.c

14 lines

dev/

fdt/

simplebus.c

1 line

kern/

bus_if.m

15 lines

subr_bus.c

38 lines

sys/

bus.h

3 lines

Diff 6910

View Options

sys/arm64/arm64/intr_machdep.c

	.\"	.\"
	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>
		.\" Copyright (c) 2014 The FreeBSD Foundation
	.\" All rights reserved.	.\" All rights reserved.
	.\"	.\"
	.\" Redistribution and use in source and binary forms, with or without	.\" Redistribution and use in source and binary forms, with or without
Context not available.
	.\"	.\"
	.\" $FreeBSD$	.\" $FreeBSD$
	.\"	.\"
	.Dd July 21, 2003	.Dd August 13, 2014
	.Dt PMAP_ENTER 9	.Dt PMAP_ENTER 9
	.Os	.Os
	.Sh NAME	.Sh NAME
Context not available.
	.In sys/param.h	.In sys/param.h
	.In vm/vm.h	.In vm/vm.h
	.In vm/pmap.h	.In vm/pmap.h
	.Ft void	.Ft int
	.Fo pmap_enter	.Fo pmap_enter
	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t p" "vm_prot_t prot"	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t m" "vm_prot_t prot"
	.Fa "boolean_t wired"	.Fa "u_int flags" "int8_t psind"
	.Fc	.Fc
	.Sh DESCRIPTION	.Sh DESCRIPTION
	The	The
	.Fn pmap_enter	.Fn pmap_enter
	function inserts the given physical page	function creates a mapping in the physical map
		alcUnsubmitted Not Done Inline Actions This is really a comment about the existing first sentence. I find "... inserts ... into the physical map ..." to be utterly meaningless. How does the first sentence not include the word "mapping". I suggest changing this sentence to "... function creates a mapping in the physical map ... from the virtual address ... to the physical page ... with the protection ..." Maybe add the following sentence: "Any previous mapping at the virtual address ... is destroyed." alc: This is really a comment about the existing first sentence. I find "... inserts ... into the…
	.Fa p ,	.Fa pmap
	into the physical map	from the virtual address
	.Fa pmap ,	.Fa va
	at the virtual address	to the physical page
	.Fa va ,	.Fa m
	with the protection	with the protection
	.Fa prot .	.Fa prot .
	If	Any previous mapping at the virtual address
	.Fa wired	.Fa va
	is	is destroyed.
	.Dv TRUE ,	.Pp
	then increment the wired count for the page as soon as the mapping	The
	is inserted into	.Fa flags
	.Fa pmap .	argument takes the following values:
		alcUnsubmitted Not Done Inline Actions "... takes ..." -> "... may have ..." alc: "... takes ..." -> "... may have ..."
	.Sh IMPLEMENTATION NOTES	.Bl -tag -width ".Dv PMAP_ENTER_NOSLEEP"
	This routine MAY NOT lazy-evaluate the entry; it is required by	.It Dv VM_PROT_READ
	specification to make the requested entry at the time it is called.	The page access, which initiated the call, was for read.
		alcUnsubmitted Not Done Inline Actions As a rule, try to avoid passive voice, so... "A read access to the given virtual address triggered the call." alc: As a rule, try to avoid passive voice, so... "A read access to the given virtual address…
		.It Dv VM_PROT_WRITE
		The page access, which initiated the call, was for write.
		alcUnsubmitted Not Done Inline Actions "A write access to the given virtual address triggered the call." alc: "A write access to the given virtual address triggered the call."
		.It Dv VM_PROT_EXECUTE
		The page access, which initiated the call, was for execution.
		alcUnsubmitted Not Done Inline Actions I would simply say, "The mapping should be marked as wired." alc: I would simply say, "The mapping should be marked as wired."
		alcUnsubmitted Not Done Inline Actions "An execute access to the given virtual address triggered the call." alc: "An execute access to the given virtual address triggered the call."
		.It Dv PMAP_ENTER_WIRED
		The mapping should be marked as wired.
		alcUnsubmitted Not Done Inline Actions The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying, "This function may not ..." alc: The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying…
		.It Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions "If the mapping cannot be created without sleeping, ..." alc: "If the mapping cannot be created without sleeping, ..."
		This function may not sleep during creation of the mapping.
		If the mapping cannot be created without sleeping, an appropriate
		Mach VM error must be returned.
		alcUnsubmitted Not Done Inline Actions "... error code is returned." alc: "... error code is returned."
		.El
		If the
		.Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would say, "... this function must create the requested mapping before returning. It may not fail in the creation of the mapping." alc: I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would…
		flag is not specified, this function must create the requested mapping
		before returning.
		It may not fail in the creation of the mapping.
		alcUnsubmitted Not Done Inline Actions I think you can simply say "It may not fail. alc: I think you can simply say "It may not fail.
		In order to create the requested mapping, this function may destroy
		alcUnsubmitted Not Done Inline Actions "In order to create the requested mapping, this function may destroy any non-wired mapping in any pmap." alc: "In order to create the requested mapping, this function may destroy any non-wired mapping in…
		any non-wired mapping in any pmap.
		.Pp
		alcUnsubmitted Not Done Inline Actions Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies the page size that should be used by the mapping. The supported page sizes are described by the global array pagesizes[]. The desired page size is specified by passing the index of the array element that equals the desired page size. Maybe add a bugs or implementation notes section saying that the implementation is incomplete on amd64, arm, and x86. Only 0 is supported. alc: Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies…
		When destroying a managed mapping during creation of the requested entry,
		including the pre-existing mapping for the address
		.Fa va ,
		the
		.Va PGA_REFERENCED
		aflag for the corresponding page should be updated by the information
		alcUnsubmitted Not Done Inline Actions The word "managed" should really be in the first clause, not the second. I need to stop here for now. alc: The word "managed" should really be in the first clause, not the second. I need to stop here…
		from hardware.
		alcUnsubmitted Not Done Inline Actions "When the pmap_enter function destroys or updates a managed mapping, including an existing mapping at virtual address va, it updates the vm_page structure corresponding to the previously mapped physical page. If the physical page was accessed through the managed mapping, then the vm_page structure's PGA_REFERENCED aflag is set. alc: "When the pmap_enter function destroys or updates a managed mapping, including an existing…
		If the hardware indicated that the destroyed mapping was used to write
		to the mapped page, the page
		.Va dirty
		field must be set with
		.Fn vm_page_dirty
		call.
		alcUnsubmitted Not Done Inline Actions "If the physical page was modified through the managed mapping, then the vm_page_dirty function is called on the vm_page structure. alc: "If the physical page was modified through the managed mapping, then the vm_page_dirty function…
		.Pp
		The
		.Va PGA_WRITEABLE
		aflag must be set for the page
		.Fa m
		if the new mapping is writeable.
		alcUnsubmitted Not Done Inline Actions We only set PGA_WRITEABLE for managed mappings. alc: We only set PGA_WRITEABLE for managed mappings.
		It is advised to clear
		.Va PGA_WRITEABLE
		for destroyed mappings if the implementation can ensure
		that no other writeable managed mappings for the previously
		mapped pages exist.
		.Pp
		If the passed page
		alcUnsubmitted Not Done Inline Actions Drop "passed". The use of "m" here makes it clear that we're talking about the given physical page. alc: Drop "passed". The use of "m" here makes it clear that we're talking about the given physical…
		.Fa m
		is managed, the page must be busied by the caller,
		or the owning object must be locked, in which case
		alcUnsubmitted Not Done Inline Actions "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the PMAP_ENTER_NOSLEEP flag must be specified by the caller. alc: "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the…
		.Dv PMAP_ENTER_NOSLEEP
		have to be specified in
		.Fa flags .
		.Pp
		The
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this function", "the pmap_enter function", or simply "pmap_enter". alc: This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this…
		must handle the multiprocessor TLB consistency for the given address.
		.Sh RETURN VALUES
		If successful, the
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions See previous comment. alc: See previous comment.
		returns
		.Er KERN_SUCCESS.
		If the
		.Dv PMAP_ENTER_NOSLEEP
		flag was specified, and required resources cannot be acquired
		in non-sleeping manner, typical implementation returns
		alcUnsubmitted Not Done Inline Actions "... specified (no comma here) and the resources required for the mapping cannot be acquired without sleeping, KERN_RESOURCE_SHORTAGE is returned." I don't see the point of encouraging people to invent different return values. Didn't we use the same return value in all the implementations so far? alc: "... specified (no comma here) and the resources required for the mapping cannot be acquired…
		.Dv KERN_RESOURCE_SHORTAGE
		error.
	.Sh SEE ALSO	.Sh SEE ALSO
	.Xr pmap 9	.Xr pmap 9
	.Sh AUTHORS	.Sh AUTHORS
	This manual page was written by	This manual page was written by
		alcUnsubmitted Not Done Inline Actions "... was first written ... alc: "... was first written ...
	.An Bruce M Simpson Aq Mt bms@spc.org .	.An Bruce M Simpson Aq Mt bms@spc.org
		and
		.An Konstantin Belousov Aq Mt kib@freebsd.org .
Context not available.
		alcUnsubmitted Not Done Inline Actions I think that your name should go first. There is nothing left of the original text. alc: I think that your name should go first. There is nothing left of the original text.

View Options

sys/arm64/arm64/nexus.c

	.\"	.\"
	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>
		.\" Copyright (c) 2014 The FreeBSD Foundation
	.\" All rights reserved.	.\" All rights reserved.
	.\"	.\"
	.\" Redistribution and use in source and binary forms, with or without	.\" Redistribution and use in source and binary forms, with or without
Context not available.
	.\"	.\"
	.\" $FreeBSD$	.\" $FreeBSD$
	.\"	.\"
	.Dd July 21, 2003	.Dd August 13, 2014
	.Dt PMAP_ENTER 9	.Dt PMAP_ENTER 9
	.Os	.Os
	.Sh NAME	.Sh NAME
Context not available.
	.In sys/param.h	.In sys/param.h
	.In vm/vm.h	.In vm/vm.h
	.In vm/pmap.h	.In vm/pmap.h
	.Ft void	.Ft int
	.Fo pmap_enter	.Fo pmap_enter
	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t p" "vm_prot_t prot"	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t m" "vm_prot_t prot"
	.Fa "boolean_t wired"	.Fa "u_int flags" "int8_t psind"
	.Fc	.Fc
	.Sh DESCRIPTION	.Sh DESCRIPTION
	The	The
	.Fn pmap_enter	.Fn pmap_enter
	function inserts the given physical page	function creates a mapping in the physical map
		alcUnsubmitted Not Done Inline Actions This is really a comment about the existing first sentence. I find "... inserts ... into the physical map ..." to be utterly meaningless. How does the first sentence not include the word "mapping". I suggest changing this sentence to "... function creates a mapping in the physical map ... from the virtual address ... to the physical page ... with the protection ..." Maybe add the following sentence: "Any previous mapping at the virtual address ... is destroyed." alc: This is really a comment about the existing first sentence. I find "... inserts ... into the…
	.Fa p ,	.Fa pmap
	into the physical map	from the virtual address
	.Fa pmap ,	.Fa va
	at the virtual address	to the physical page
	.Fa va ,	.Fa m
	with the protection	with the protection
	.Fa prot .	.Fa prot .
	If	Any previous mapping at the virtual address
	.Fa wired	.Fa va
	is	is destroyed.
	.Dv TRUE ,	.Pp
	then increment the wired count for the page as soon as the mapping	The
	is inserted into	.Fa flags
	.Fa pmap .	argument takes the following values:
		alcUnsubmitted Not Done Inline Actions "... takes ..." -> "... may have ..." alc: "... takes ..." -> "... may have ..."
	.Sh IMPLEMENTATION NOTES	.Bl -tag -width ".Dv PMAP_ENTER_NOSLEEP"
	This routine MAY NOT lazy-evaluate the entry; it is required by	.It Dv VM_PROT_READ
	specification to make the requested entry at the time it is called.	The page access, which initiated the call, was for read.
		alcUnsubmitted Not Done Inline Actions As a rule, try to avoid passive voice, so... "A read access to the given virtual address triggered the call." alc: As a rule, try to avoid passive voice, so... "A read access to the given virtual address…
		.It Dv VM_PROT_WRITE
		The page access, which initiated the call, was for write.
		alcUnsubmitted Not Done Inline Actions "A write access to the given virtual address triggered the call." alc: "A write access to the given virtual address triggered the call."
		.It Dv VM_PROT_EXECUTE
		The page access, which initiated the call, was for execution.
		alcUnsubmitted Not Done Inline Actions I would simply say, "The mapping should be marked as wired." alc: I would simply say, "The mapping should be marked as wired."
		alcUnsubmitted Not Done Inline Actions "An execute access to the given virtual address triggered the call." alc: "An execute access to the given virtual address triggered the call."
		.It Dv PMAP_ENTER_WIRED
		The mapping should be marked as wired.
		alcUnsubmitted Not Done Inline Actions The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying, "This function may not ..." alc: The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying…
		.It Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions "If the mapping cannot be created without sleeping, ..." alc: "If the mapping cannot be created without sleeping, ..."
		This function may not sleep during creation of the mapping.
		If the mapping cannot be created without sleeping, an appropriate
		Mach VM error must be returned.
		alcUnsubmitted Not Done Inline Actions "... error code is returned." alc: "... error code is returned."
		.El
		If the
		.Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would say, "... this function must create the requested mapping before returning. It may not fail in the creation of the mapping." alc: I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would…
		flag is not specified, this function must create the requested mapping
		before returning.
		It may not fail in the creation of the mapping.
		alcUnsubmitted Not Done Inline Actions I think you can simply say "It may not fail. alc: I think you can simply say "It may not fail.
		In order to create the requested mapping, this function may destroy
		alcUnsubmitted Not Done Inline Actions "In order to create the requested mapping, this function may destroy any non-wired mapping in any pmap." alc: "In order to create the requested mapping, this function may destroy any non-wired mapping in…
		any non-wired mapping in any pmap.
		.Pp
		alcUnsubmitted Not Done Inline Actions Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies the page size that should be used by the mapping. The supported page sizes are described by the global array pagesizes[]. The desired page size is specified by passing the index of the array element that equals the desired page size. Maybe add a bugs or implementation notes section saying that the implementation is incomplete on amd64, arm, and x86. Only 0 is supported. alc: Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies…
		When destroying a managed mapping during creation of the requested entry,
		including the pre-existing mapping for the address
		.Fa va ,
		the
		.Va PGA_REFERENCED
		aflag for the corresponding page should be updated by the information
		alcUnsubmitted Not Done Inline Actions The word "managed" should really be in the first clause, not the second. I need to stop here for now. alc: The word "managed" should really be in the first clause, not the second. I need to stop here…
		from hardware.
		alcUnsubmitted Not Done Inline Actions "When the pmap_enter function destroys or updates a managed mapping, including an existing mapping at virtual address va, it updates the vm_page structure corresponding to the previously mapped physical page. If the physical page was accessed through the managed mapping, then the vm_page structure's PGA_REFERENCED aflag is set. alc: "When the pmap_enter function destroys or updates a managed mapping, including an existing…
		If the hardware indicated that the destroyed mapping was used to write
		to the mapped page, the page
		.Va dirty
		field must be set with
		.Fn vm_page_dirty
		call.
		alcUnsubmitted Not Done Inline Actions "If the physical page was modified through the managed mapping, then the vm_page_dirty function is called on the vm_page structure. alc: "If the physical page was modified through the managed mapping, then the vm_page_dirty function…
		.Pp
		The
		.Va PGA_WRITEABLE
		aflag must be set for the page
		.Fa m
		if the new mapping is writeable.
		alcUnsubmitted Not Done Inline Actions We only set PGA_WRITEABLE for managed mappings. alc: We only set PGA_WRITEABLE for managed mappings.
		It is advised to clear
		.Va PGA_WRITEABLE
		for destroyed mappings if the implementation can ensure
		that no other writeable managed mappings for the previously
		mapped pages exist.
		.Pp
		If the passed page
		alcUnsubmitted Not Done Inline Actions Drop "passed". The use of "m" here makes it clear that we're talking about the given physical page. alc: Drop "passed". The use of "m" here makes it clear that we're talking about the given physical…
		.Fa m
		is managed, the page must be busied by the caller,
		or the owning object must be locked, in which case
		alcUnsubmitted Not Done Inline Actions "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the PMAP_ENTER_NOSLEEP flag must be specified by the caller. alc: "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the…
		.Dv PMAP_ENTER_NOSLEEP
		have to be specified in
		.Fa flags .
		.Pp
		The
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this function", "the pmap_enter function", or simply "pmap_enter". alc: This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this…
		must handle the multiprocessor TLB consistency for the given address.
		.Sh RETURN VALUES
		If successful, the
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions See previous comment. alc: See previous comment.
		returns
		.Er KERN_SUCCESS.
		If the
		.Dv PMAP_ENTER_NOSLEEP
		flag was specified, and required resources cannot be acquired
		in non-sleeping manner, typical implementation returns
		alcUnsubmitted Not Done Inline Actions "... specified (no comma here) and the resources required for the mapping cannot be acquired without sleeping, KERN_RESOURCE_SHORTAGE is returned." I don't see the point of encouraging people to invent different return values. Didn't we use the same return value in all the implementations so far? alc: "... specified (no comma here) and the resources required for the mapping cannot be acquired…
		.Dv KERN_RESOURCE_SHORTAGE
		error.
	.Sh SEE ALSO	.Sh SEE ALSO
	.Xr pmap 9	.Xr pmap 9
	.Sh AUTHORS	.Sh AUTHORS
	This manual page was written by	This manual page was written by
		alcUnsubmitted Not Done Inline Actions "... was first written ... alc: "... was first written ...
	.An Bruce M Simpson Aq Mt bms@spc.org .	.An Bruce M Simpson Aq Mt bms@spc.org
		and
		.An Konstantin Belousov Aq Mt kib@freebsd.org .
Context not available.
		alcUnsubmitted Not Done Inline Actions I think that your name should go first. There is nothing left of the original text. alc: I think that your name should go first. There is nothing left of the original text.

View Options

sys/dev/fdt/simplebus.c

	.\"	.\"
	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>
		.\" Copyright (c) 2014 The FreeBSD Foundation
	.\" All rights reserved.	.\" All rights reserved.
	.\"	.\"
	.\" Redistribution and use in source and binary forms, with or without	.\" Redistribution and use in source and binary forms, with or without
Context not available.
	.\"	.\"
	.\" $FreeBSD$	.\" $FreeBSD$
	.\"	.\"
	.Dd July 21, 2003	.Dd August 13, 2014
	.Dt PMAP_ENTER 9	.Dt PMAP_ENTER 9
	.Os	.Os
	.Sh NAME	.Sh NAME
Context not available.
	.In sys/param.h	.In sys/param.h
	.In vm/vm.h	.In vm/vm.h
	.In vm/pmap.h	.In vm/pmap.h
	.Ft void	.Ft int
	.Fo pmap_enter	.Fo pmap_enter
	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t p" "vm_prot_t prot"	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t m" "vm_prot_t prot"
	.Fa "boolean_t wired"	.Fa "u_int flags" "int8_t psind"
	.Fc	.Fc
	.Sh DESCRIPTION	.Sh DESCRIPTION
	The	The
	.Fn pmap_enter	.Fn pmap_enter
	function inserts the given physical page	function creates a mapping in the physical map
		alcUnsubmitted Not Done Inline Actions This is really a comment about the existing first sentence. I find "... inserts ... into the physical map ..." to be utterly meaningless. How does the first sentence not include the word "mapping". I suggest changing this sentence to "... function creates a mapping in the physical map ... from the virtual address ... to the physical page ... with the protection ..." Maybe add the following sentence: "Any previous mapping at the virtual address ... is destroyed." alc: This is really a comment about the existing first sentence. I find "... inserts ... into the…
	.Fa p ,	.Fa pmap
	into the physical map	from the virtual address
	.Fa pmap ,	.Fa va
	at the virtual address	to the physical page
	.Fa va ,	.Fa m
	with the protection	with the protection
	.Fa prot .	.Fa prot .
	If	Any previous mapping at the virtual address
	.Fa wired	.Fa va
	is	is destroyed.
	.Dv TRUE ,	.Pp
	then increment the wired count for the page as soon as the mapping	The
	is inserted into	.Fa flags
	.Fa pmap .	argument takes the following values:
		alcUnsubmitted Not Done Inline Actions "... takes ..." -> "... may have ..." alc: "... takes ..." -> "... may have ..."
	.Sh IMPLEMENTATION NOTES	.Bl -tag -width ".Dv PMAP_ENTER_NOSLEEP"
	This routine MAY NOT lazy-evaluate the entry; it is required by	.It Dv VM_PROT_READ
	specification to make the requested entry at the time it is called.	The page access, which initiated the call, was for read.
		alcUnsubmitted Not Done Inline Actions As a rule, try to avoid passive voice, so... "A read access to the given virtual address triggered the call." alc: As a rule, try to avoid passive voice, so... "A read access to the given virtual address…
		.It Dv VM_PROT_WRITE
		The page access, which initiated the call, was for write.
		alcUnsubmitted Not Done Inline Actions "A write access to the given virtual address triggered the call." alc: "A write access to the given virtual address triggered the call."
		.It Dv VM_PROT_EXECUTE
		The page access, which initiated the call, was for execution.
		alcUnsubmitted Not Done Inline Actions I would simply say, "The mapping should be marked as wired." alc: I would simply say, "The mapping should be marked as wired."
		alcUnsubmitted Not Done Inline Actions "An execute access to the given virtual address triggered the call." alc: "An execute access to the given virtual address triggered the call."
		.It Dv PMAP_ENTER_WIRED
		The mapping should be marked as wired.
		alcUnsubmitted Not Done Inline Actions The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying, "This function may not ..." alc: The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying…
		.It Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions "If the mapping cannot be created without sleeping, ..." alc: "If the mapping cannot be created without sleeping, ..."
		This function may not sleep during creation of the mapping.
		If the mapping cannot be created without sleeping, an appropriate
		Mach VM error must be returned.
		alcUnsubmitted Not Done Inline Actions "... error code is returned." alc: "... error code is returned."
		.El
		If the
		.Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would say, "... this function must create the requested mapping before returning. It may not fail in the creation of the mapping." alc: I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would…
		flag is not specified, this function must create the requested mapping
		before returning.
		It may not fail in the creation of the mapping.
		alcUnsubmitted Not Done Inline Actions I think you can simply say "It may not fail. alc: I think you can simply say "It may not fail.
		In order to create the requested mapping, this function may destroy
		alcUnsubmitted Not Done Inline Actions "In order to create the requested mapping, this function may destroy any non-wired mapping in any pmap." alc: "In order to create the requested mapping, this function may destroy any non-wired mapping in…
		any non-wired mapping in any pmap.
		.Pp
		alcUnsubmitted Not Done Inline Actions Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies the page size that should be used by the mapping. The supported page sizes are described by the global array pagesizes[]. The desired page size is specified by passing the index of the array element that equals the desired page size. Maybe add a bugs or implementation notes section saying that the implementation is incomplete on amd64, arm, and x86. Only 0 is supported. alc: Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies…
		When destroying a managed mapping during creation of the requested entry,
		including the pre-existing mapping for the address
		.Fa va ,
		the
		.Va PGA_REFERENCED
		aflag for the corresponding page should be updated by the information
		alcUnsubmitted Not Done Inline Actions The word "managed" should really be in the first clause, not the second. I need to stop here for now. alc: The word "managed" should really be in the first clause, not the second. I need to stop here…
		from hardware.
		alcUnsubmitted Not Done Inline Actions "When the pmap_enter function destroys or updates a managed mapping, including an existing mapping at virtual address va, it updates the vm_page structure corresponding to the previously mapped physical page. If the physical page was accessed through the managed mapping, then the vm_page structure's PGA_REFERENCED aflag is set. alc: "When the pmap_enter function destroys or updates a managed mapping, including an existing…
		If the hardware indicated that the destroyed mapping was used to write
		to the mapped page, the page
		.Va dirty
		field must be set with
		.Fn vm_page_dirty
		call.
		alcUnsubmitted Not Done Inline Actions "If the physical page was modified through the managed mapping, then the vm_page_dirty function is called on the vm_page structure. alc: "If the physical page was modified through the managed mapping, then the vm_page_dirty function…
		.Pp
		The
		.Va PGA_WRITEABLE
		aflag must be set for the page
		.Fa m
		if the new mapping is writeable.
		alcUnsubmitted Not Done Inline Actions We only set PGA_WRITEABLE for managed mappings. alc: We only set PGA_WRITEABLE for managed mappings.
		It is advised to clear
		.Va PGA_WRITEABLE
		for destroyed mappings if the implementation can ensure
		that no other writeable managed mappings for the previously
		mapped pages exist.
		.Pp
		If the passed page
		alcUnsubmitted Not Done Inline Actions Drop "passed". The use of "m" here makes it clear that we're talking about the given physical page. alc: Drop "passed". The use of "m" here makes it clear that we're talking about the given physical…
		.Fa m
		is managed, the page must be busied by the caller,
		or the owning object must be locked, in which case
		alcUnsubmitted Not Done Inline Actions "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the PMAP_ENTER_NOSLEEP flag must be specified by the caller. alc: "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the…
		.Dv PMAP_ENTER_NOSLEEP
		have to be specified in
		.Fa flags .
		.Pp
		The
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this function", "the pmap_enter function", or simply "pmap_enter". alc: This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this…
		must handle the multiprocessor TLB consistency for the given address.
		.Sh RETURN VALUES
		If successful, the
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions See previous comment. alc: See previous comment.
		returns
		.Er KERN_SUCCESS.
		If the
		.Dv PMAP_ENTER_NOSLEEP
		flag was specified, and required resources cannot be acquired
		in non-sleeping manner, typical implementation returns
		alcUnsubmitted Not Done Inline Actions "... specified (no comma here) and the resources required for the mapping cannot be acquired without sleeping, KERN_RESOURCE_SHORTAGE is returned." I don't see the point of encouraging people to invent different return values. Didn't we use the same return value in all the implementations so far? alc: "... specified (no comma here) and the resources required for the mapping cannot be acquired…
		.Dv KERN_RESOURCE_SHORTAGE
		error.
	.Sh SEE ALSO	.Sh SEE ALSO
	.Xr pmap 9	.Xr pmap 9
	.Sh AUTHORS	.Sh AUTHORS
	This manual page was written by	This manual page was written by
		alcUnsubmitted Not Done Inline Actions "... was first written ... alc: "... was first written ...
	.An Bruce M Simpson Aq Mt bms@spc.org .	.An Bruce M Simpson Aq Mt bms@spc.org
		and
		.An Konstantin Belousov Aq Mt kib@freebsd.org .
Context not available.
		alcUnsubmitted Not Done Inline Actions I think that your name should go first. There is nothing left of the original text. alc: I think that your name should go first. There is nothing left of the original text.

View Options

sys/kern/bus_if.m

	.\"	.\"
	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>	.\" Copyright (c) 2003 Bruce M Simpson <bms@spc.org>
		.\" Copyright (c) 2014 The FreeBSD Foundation
	.\" All rights reserved.	.\" All rights reserved.
	.\"	.\"
	.\" Redistribution and use in source and binary forms, with or without	.\" Redistribution and use in source and binary forms, with or without
Context not available.
	.\"	.\"
	.\" $FreeBSD$	.\" $FreeBSD$
	.\"	.\"
	.Dd July 21, 2003	.Dd August 13, 2014
	.Dt PMAP_ENTER 9	.Dt PMAP_ENTER 9
	.Os	.Os
	.Sh NAME	.Sh NAME
Context not available.
	.In sys/param.h	.In sys/param.h
	.In vm/vm.h	.In vm/vm.h
	.In vm/pmap.h	.In vm/pmap.h
	.Ft void	.Ft int
	.Fo pmap_enter	.Fo pmap_enter
	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t p" "vm_prot_t prot"	.Fa "pmap_t pmap" "vm_offset_t va" "vm_page_t m" "vm_prot_t prot"
	.Fa "boolean_t wired"	.Fa "u_int flags" "int8_t psind"
	.Fc	.Fc
	.Sh DESCRIPTION	.Sh DESCRIPTION
	The	The
	.Fn pmap_enter	.Fn pmap_enter
	function inserts the given physical page	function creates a mapping in the physical map
		alcUnsubmitted Not Done Inline Actions This is really a comment about the existing first sentence. I find "... inserts ... into the physical map ..." to be utterly meaningless. How does the first sentence not include the word "mapping". I suggest changing this sentence to "... function creates a mapping in the physical map ... from the virtual address ... to the physical page ... with the protection ..." Maybe add the following sentence: "Any previous mapping at the virtual address ... is destroyed." alc: This is really a comment about the existing first sentence. I find "... inserts ... into the…
	.Fa p ,	.Fa pmap
	into the physical map	from the virtual address
	.Fa pmap ,	.Fa va
	at the virtual address	to the physical page
	.Fa va ,	.Fa m
	with the protection	with the protection
	.Fa prot .	.Fa prot .
	If	Any previous mapping at the virtual address
	.Fa wired	.Fa va
	is	is destroyed.
	.Dv TRUE ,	.Pp
	then increment the wired count for the page as soon as the mapping	The
	is inserted into	.Fa flags
	.Fa pmap .	argument takes the following values:
		alcUnsubmitted Not Done Inline Actions "... takes ..." -> "... may have ..." alc: "... takes ..." -> "... may have ..."
	.Sh IMPLEMENTATION NOTES	.Bl -tag -width ".Dv PMAP_ENTER_NOSLEEP"
	This routine MAY NOT lazy-evaluate the entry; it is required by	.It Dv VM_PROT_READ
	specification to make the requested entry at the time it is called.	The page access, which initiated the call, was for read.
		alcUnsubmitted Not Done Inline Actions As a rule, try to avoid passive voice, so... "A read access to the given virtual address triggered the call." alc: As a rule, try to avoid passive voice, so... "A read access to the given virtual address…
		.It Dv VM_PROT_WRITE
		The page access, which initiated the call, was for write.
		alcUnsubmitted Not Done Inline Actions "A write access to the given virtual address triggered the call." alc: "A write access to the given virtual address triggered the call."
		.It Dv VM_PROT_EXECUTE
		The page access, which initiated the call, was for execution.
		alcUnsubmitted Not Done Inline Actions I would simply say, "The mapping should be marked as wired." alc: I would simply say, "The mapping should be marked as wired."
		alcUnsubmitted Not Done Inline Actions "An execute access to the given virtual address triggered the call." alc: "An execute access to the given virtual address triggered the call."
		.It Dv PMAP_ENTER_WIRED
		The mapping should be marked as wired.
		alcUnsubmitted Not Done Inline Actions The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying, "This function may not ..." alc: The phrasing "The pmap implementation ..." sounds like the entire pmap. Restrict it by saying…
		.It Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions "If the mapping cannot be created without sleeping, ..." alc: "If the mapping cannot be created without sleeping, ..."
		This function may not sleep during creation of the mapping.
		If the mapping cannot be created without sleeping, an appropriate
		Mach VM error must be returned.
		alcUnsubmitted Not Done Inline Actions "... error code is returned." alc: "... error code is returned."
		.El
		If the
		.Dv PMAP_ENTER_NOSLEEP
		alcUnsubmitted Not Done Inline Actions I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would say, "... this function must create the requested mapping before returning. It may not fail in the creation of the mapping." alc: I don't think that "lazy-evaluate" adds anything here to the clarity of the statement. I would…
		flag is not specified, this function must create the requested mapping
		before returning.
		It may not fail in the creation of the mapping.
		alcUnsubmitted Not Done Inline Actions I think you can simply say "It may not fail. alc: I think you can simply say "It may not fail.
		In order to create the requested mapping, this function may destroy
		alcUnsubmitted Not Done Inline Actions "In order to create the requested mapping, this function may destroy any non-wired mapping in any pmap." alc: "In order to create the requested mapping, this function may destroy any non-wired mapping in…
		any non-wired mapping in any pmap.
		.Pp
		alcUnsubmitted Not Done Inline Actions Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies the page size that should be used by the mapping. The supported page sizes are described by the global array pagesizes[]. The desired page size is specified by passing the index of the array element that equals the desired page size. Maybe add a bugs or implementation notes section saying that the implementation is incomplete on amd64, arm, and x86. Only 0 is supported. alc: Insert a paragraph with a couple sentences about psind here? "The psind parameter specifies…
		When destroying a managed mapping during creation of the requested entry,
		including the pre-existing mapping for the address
		.Fa va ,
		the
		.Va PGA_REFERENCED
		aflag for the corresponding page should be updated by the information
		alcUnsubmitted Not Done Inline Actions The word "managed" should really be in the first clause, not the second. I need to stop here for now. alc: The word "managed" should really be in the first clause, not the second. I need to stop here…
		from hardware.
		alcUnsubmitted Not Done Inline Actions "When the pmap_enter function destroys or updates a managed mapping, including an existing mapping at virtual address va, it updates the vm_page structure corresponding to the previously mapped physical page. If the physical page was accessed through the managed mapping, then the vm_page structure's PGA_REFERENCED aflag is set. alc: "When the pmap_enter function destroys or updates a managed mapping, including an existing…
		If the hardware indicated that the destroyed mapping was used to write
		to the mapped page, the page
		.Va dirty
		field must be set with
		.Fn vm_page_dirty
		call.
		alcUnsubmitted Not Done Inline Actions "If the physical page was modified through the managed mapping, then the vm_page_dirty function is called on the vm_page structure. alc: "If the physical page was modified through the managed mapping, then the vm_page_dirty function…
		.Pp
		The
		.Va PGA_WRITEABLE
		aflag must be set for the page
		.Fa m
		if the new mapping is writeable.
		alcUnsubmitted Not Done Inline Actions We only set PGA_WRITEABLE for managed mappings. alc: We only set PGA_WRITEABLE for managed mappings.
		It is advised to clear
		.Va PGA_WRITEABLE
		for destroyed mappings if the implementation can ensure
		that no other writeable managed mappings for the previously
		mapped pages exist.
		.Pp
		If the passed page
		alcUnsubmitted Not Done Inline Actions Drop "passed". The use of "m" here makes it clear that we're talking about the given physical page. alc: Drop "passed". The use of "m" here makes it clear that we're talking about the given physical…
		.Fa m
		is managed, the page must be busied by the caller,
		or the owning object must be locked, in which case
		alcUnsubmitted Not Done Inline Actions "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the PMAP_ENTER_NOSLEEP flag must be specified by the caller. alc: "..., then it must be ... caller (no comma here) or ... locked. In the latter case, the…
		.Dv PMAP_ENTER_NOSLEEP
		have to be specified in
		.Fa flags .
		.Pp
		The
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this function", "the pmap_enter function", or simply "pmap_enter". alc: This is equivalent to saying, "the Alan", which I don't think that you would do. :-) Use "this…
		must handle the multiprocessor TLB consistency for the given address.
		.Sh RETURN VALUES
		If successful, the
		.Fn pmap_enter
		alcUnsubmitted Not Done Inline Actions See previous comment. alc: See previous comment.
		returns
		.Er KERN_SUCCESS.
		If the
		.Dv PMAP_ENTER_NOSLEEP
		flag was specified, and required resources cannot be acquired
		in non-sleeping manner, typical implementation returns
		alcUnsubmitted Not Done Inline Actions "... specified (no comma here) and the resources required for the mapping cannot be acquired without sleeping, KERN_RESOURCE_SHORTAGE is returned." I don't see the point of encouraging people to invent different return values. Didn't we use the same return value in all the implementations so far? alc: "... specified (no comma here) and the resources required for the mapping cannot be acquired…
		.Dv KERN_RESOURCE_SHORTAGE
		error.
	.Sh SEE ALSO	.Sh SEE ALSO
	.Xr pmap 9	.Xr pmap 9
	.Sh AUTHORS	.Sh AUTHORS
	This manual page was written by	This manual page was written by
		alcUnsubmitted Not Done Inline Actions "... was first written ... alc: "... was first written ...
	.An Bruce M Simpson Aq Mt bms@spc.org .	.An Bruce M Simpson Aq Mt bms@spc.org
		and
		.An Konstantin Belousov Aq Mt kib@freebsd.org .
Context not available.
		alcUnsubmitted Not Done Inline Actions I think that your name should go first. There is nothing left of the original text. alc: I think that your name should go first. There is nothing left of the original text.

View Options

sys/kern/subr_bus.c

View Options

Add BUS_UNMASK_INTR methodAbandonedPublicActions

Details

Diff Detail

Event Timeline

Revision ContentsChangeset List

Diff 6910

sys/arm64/arm64/intr_machdep.c

sys/arm64/arm64/nexus.c

sys/dev/fdt/simplebus.c

sys/kern/bus_if.m

sys/kern/subr_bus.c

sys/sys/bus.h

Add BUS_UNMASK_INTR method
AbandonedPublic
Actions

Revision Contents
Changeset List