OpenCloudOS-Kernel/arch/x86/xen/setup.c

628 lines
17 KiB
C
Raw Normal View History

xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
/*
* Machine specific setup for xen
*
* Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
*/
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/pm.h>
#include <linux/memblock.h>
#include <linux/cpuidle.h>
#include <linux/cpufreq.h>
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#include <asm/elf.h>
#include <asm/vdso.h>
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#include <asm/e820.h>
#include <asm/setup.h>
#include <asm/acpi.h>
xen/boot: Disable NUMA for PV guests. The hypervisor is in charge of allocating the proper "NUMA" memory and dealing with the CPU scheduler to keep them bound to the proper NUMA node. The PV guests (and PVHVM) have no inkling of where they run and do not need to know that right now. In the future we will need to inject NUMA configuration data (if a guest spans two or more NUMA nodes) so that the kernel can make the right choices. But those patches are not yet present. In the meantime, disable the NUMA capability in the PV guest, which also fixes a bootup issue. Andre says: "we see Dom0 crashes due to the kernel detecting the NUMA topology not by ACPI, but directly from the northbridge (CONFIG_AMD_NUMA). This will detect the actual NUMA config of the physical machine, but will crash about the mismatch with Dom0's virtual memory. Variation of the theme: Dom0 sees what it's not supposed to see. This happens with the said config option enabled and on a machine where this scanning is still enabled (K8 and Fam10h, not Bulldozer class) We have this dump then: NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10 Scanning NUMA topology in Northbridge 24 Number of physical nodes 4 Node 0 MemBase 0000000000000000 Limit 0000000040000000 Node 1 MemBase 0000000040000000 Limit 0000000138000000 Node 2 MemBase 0000000138000000 Limit 00000001f8000000 Node 3 MemBase 00000001f8000000 Limit 0000000238000000 Initmem setup node 0 0000000000000000-0000000040000000 NODE_DATA [000000003ffd9000 - 000000003fffffff] Initmem setup node 1 0000000040000000-0000000138000000 NODE_DATA [0000000137fd9000 - 0000000137ffffff] Initmem setup node 2 0000000138000000-00000001f8000000 NODE_DATA [00000001f095e000 - 00000001f0984fff] Initmem setup node 3 00000001f8000000-0000000238000000 Cannot find 159744 bytes in node 3 BUG: unable to handle kernel NULL pointer dereference at (null) IP: [<ffffffff81d220e6>] __alloc_bootmem_node+0x43/0x96 Pid: 0, comm: swapper Not tainted 3.3.6 #1 AMD Dinar/Dinar RIP: e030:[<ffffffff81d220e6>] [<ffffffff81d220e6>] __alloc_bootmem_node+0x43/0x96 .. snip.. [<ffffffff81d23024>] sparse_early_usemaps_alloc_node+0x64/0x178 [<ffffffff81d23348>] sparse_init+0xe4/0x25a [<ffffffff81d16840>] paging_init+0x13/0x22 [<ffffffff81d07fbb>] setup_arch+0x9c6/0xa9b [<ffffffff81683954>] ? printk+0x3c/0x3e [<ffffffff81d01a38>] start_kernel+0xe5/0x468 [<ffffffff81d012cf>] x86_64_start_reservations+0xba/0xc1 [<ffffffff81007153>] ? xen_setup_runstate_info+0x2c/0x36 [<ffffffff81d050ee>] xen_start_kernel+0x565/0x56c " so we just disable NUMA scanning by setting numa_off=1. CC: stable@vger.kernel.org Reported-and-Tested-by: Andre Przywara <andre.przywara@amd.com> Acked-by: Andre Przywara <andre.przywara@amd.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-08-17 22:22:37 +08:00
#include <asm/numa.h>
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#include <asm/xen/hypervisor.h>
#include <asm/xen/hypercall.h>
#include <xen/xen.h>
#include <xen/page.h>
#include <xen/interface/callback.h>
#include <xen/interface/memory.h>
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#include <xen/interface/physdev.h>
#include <xen/features.h>
#include "mmu.h"
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#include "xen-ops.h"
#include "vdso.h"
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
/* These are code, but not functions. Defined in entry.S */
extern const char xen_hypervisor_callback[];
extern const char xen_failsafe_callback[];
xen: Support 64-bit PV guest receiving NMIs This is based on a patch that Zhenzhong Duan had sent - which was missing some of the remaining pieces. The kernel has the logic to handle Xen-type-exceptions using the paravirt interface in the assembler code (see PARAVIRT_ADJUST_EXCEPTION_FRAME - pv_irq_ops.adjust_exception_frame and and INTERRUPT_RETURN - pv_cpu_ops.iret). That means the nmi handler (and other exception handlers) use the hypervisor iret. The other changes that would be neccessary for this would be to translate the NMI_VECTOR to one of the entries on the ipi_vector and make xen_send_IPI_mask_allbutself use different events. Fortunately for us commit 1db01b4903639fcfaec213701a494fe3fb2c490b (xen: Clean up apic ipi interface) implemented this and we piggyback on the cleanup such that the apic IPI interface will pass the right vector value for NMI. With this patch we can trigger NMIs within a PV guest (only tested x86_64). For this to work with normal PV guests (not initial domain) we need the domain to be able to use the APIC ops - they are already implemented to use the Xen event channels. For that to be turned on in a PV domU we need to remove the masking of X86_FEATURE_APIC. Incidentally that means kgdb will also now work within a PV guest without using the 'nokgdbroundup' workaround. Note that the 32-bit version is different and this patch does not enable that. CC: Lisa Nguyen <lisa@xenapiadmin.com> CC: Ben Guthro <benjamin.guthro@citrix.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [v1: Fixed up per David Vrabel comments] Reviewed-by: Ben Guthro <benjamin.guthro@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
2013-07-19 23:51:31 +08:00
#ifdef CONFIG_X86_64
extern asmlinkage void nmi(void);
xen: Support 64-bit PV guest receiving NMIs This is based on a patch that Zhenzhong Duan had sent - which was missing some of the remaining pieces. The kernel has the logic to handle Xen-type-exceptions using the paravirt interface in the assembler code (see PARAVIRT_ADJUST_EXCEPTION_FRAME - pv_irq_ops.adjust_exception_frame and and INTERRUPT_RETURN - pv_cpu_ops.iret). That means the nmi handler (and other exception handlers) use the hypervisor iret. The other changes that would be neccessary for this would be to translate the NMI_VECTOR to one of the entries on the ipi_vector and make xen_send_IPI_mask_allbutself use different events. Fortunately for us commit 1db01b4903639fcfaec213701a494fe3fb2c490b (xen: Clean up apic ipi interface) implemented this and we piggyback on the cleanup such that the apic IPI interface will pass the right vector value for NMI. With this patch we can trigger NMIs within a PV guest (only tested x86_64). For this to work with normal PV guests (not initial domain) we need the domain to be able to use the APIC ops - they are already implemented to use the Xen event channels. For that to be turned on in a PV domU we need to remove the masking of X86_FEATURE_APIC. Incidentally that means kgdb will also now work within a PV guest without using the 'nokgdbroundup' workaround. Note that the 32-bit version is different and this patch does not enable that. CC: Lisa Nguyen <lisa@xenapiadmin.com> CC: Ben Guthro <benjamin.guthro@citrix.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [v1: Fixed up per David Vrabel comments] Reviewed-by: Ben Guthro <benjamin.guthro@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
2013-07-19 23:51:31 +08:00
#endif
extern void xen_sysenter_target(void);
extern void xen_syscall_target(void);
extern void xen_syscall32_target(void);
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
/* Amount of extra memory space we add to the e820 ranges */
struct xen_memory_region xen_extra_mem[XEN_EXTRA_MEM_MAX_REGIONS] __initdata;
/* Number of pages released from the initial allocation. */
unsigned long xen_released_pages;
/*
* The maximum amount of extra memory compared to the base size. The
* main scaling factor is the size of struct page. At extreme ratios
* of base:extra, all the base memory can be filled with page
* structures for the extra memory, leaving no space for anything
* else.
*
* 10x seems like a reasonable balance between scaling flexibility and
* leaving a practically usable system.
*/
#define EXTRA_MEM_RATIO (10)
static void __init xen_add_extra_mem(u64 start, u64 size)
{
unsigned long pfn;
int i;
for (i = 0; i < XEN_EXTRA_MEM_MAX_REGIONS; i++) {
/* Add new region. */
if (xen_extra_mem[i].size == 0) {
xen_extra_mem[i].start = start;
xen_extra_mem[i].size = size;
break;
}
/* Append to existing region. */
if (xen_extra_mem[i].start + xen_extra_mem[i].size == start) {
xen_extra_mem[i].size += size;
break;
}
}
if (i == XEN_EXTRA_MEM_MAX_REGIONS)
printk(KERN_WARNING "Warning: not enough extra memory regions\n");
memblock_reserve(start, size);
if (xen_feature(XENFEAT_auto_translated_physmap))
return;
xen_max_p2m_pfn = PFN_DOWN(start + size);
for (pfn = PFN_DOWN(start); pfn < xen_max_p2m_pfn; pfn++) {
unsigned long mfn = pfn_to_mfn(pfn);
if (WARN(mfn == pfn, "Trying to over-write 1-1 mapping (pfn: %lx)\n", pfn))
continue;
WARN(mfn != INVALID_P2M_ENTRY, "Trying to remove %lx which has %lx mfn!\n",
pfn, mfn);
__set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
}
}
static unsigned long __init xen_do_chunk(unsigned long start,
unsigned long end, bool release)
{
struct xen_memory_reservation reservation = {
.address_bits = 0,
.extent_order = 0,
.domid = DOMID_SELF
};
unsigned long len = 0;
int xlated_phys = xen_feature(XENFEAT_auto_translated_physmap);
unsigned long pfn;
int ret;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
for (pfn = start; pfn < end; pfn++) {
unsigned long frame;
unsigned long mfn = pfn_to_mfn(pfn);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
if (release) {
/* Make sure pfn exists to start with */
if (mfn == INVALID_P2M_ENTRY || mfn_to_pfn(mfn) != pfn)
continue;
frame = mfn;
} else {
if (!xlated_phys && mfn != INVALID_P2M_ENTRY)
continue;
frame = pfn;
}
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
set_xen_guest_handle(reservation.extent_start, &frame);
reservation.nr_extents = 1;
ret = HYPERVISOR_memory_op(release ? XENMEM_decrease_reservation : XENMEM_populate_physmap,
&reservation);
WARN(ret != 1, "Failed to %s pfn %lx err=%d\n",
release ? "release" : "populate", pfn, ret);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
if (ret == 1) {
if (!early_set_phys_to_machine(pfn, release ? INVALID_P2M_ENTRY : frame)) {
if (release)
break;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
set_xen_guest_handle(reservation.extent_start, &frame);
reservation.nr_extents = 1;
ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
&reservation);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
break;
}
len++;
} else
break;
}
if (len)
printk(KERN_INFO "%s %lx-%lx pfn range: %lu pages %s\n",
release ? "Freeing" : "Populating",
start, end, len,
release ? "freed" : "added");
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
return len;
}
static unsigned long __init xen_release_chunk(unsigned long start,
unsigned long end)
{
/*
* Xen already ballooned out the E820 non RAM regions for us
* and set them up properly in EPT.
*/
if (xen_feature(XENFEAT_auto_translated_physmap))
return end - start;
return xen_do_chunk(start, end, true);
}
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
static unsigned long __init xen_populate_chunk(
const struct e820entry *list, size_t map_size,
unsigned long max_pfn, unsigned long *last_pfn,
unsigned long credits_left)
{
const struct e820entry *entry;
unsigned int i;
unsigned long done = 0;
unsigned long dest_pfn;
for (i = 0, entry = list; i < map_size; i++, entry++) {
unsigned long s_pfn;
unsigned long e_pfn;
unsigned long pfns;
long capacity;
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
if (credits_left <= 0)
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
break;
if (entry->type != E820_RAM)
continue;
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
e_pfn = PFN_DOWN(entry->addr + entry->size);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
/* We only care about E820 after the xen_start_info->nr_pages */
if (e_pfn <= max_pfn)
continue;
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
s_pfn = PFN_UP(entry->addr);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
/* If the E820 falls within the nr_pages, we want to start
* at the nr_pages PFN.
* If that would mean going past the E820 entry, skip it
*/
if (s_pfn <= max_pfn) {
capacity = e_pfn - max_pfn;
dest_pfn = max_pfn;
} else {
capacity = e_pfn - s_pfn;
dest_pfn = s_pfn;
}
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
if (credits_left < capacity)
capacity = credits_left;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
pfns = xen_do_chunk(dest_pfn, dest_pfn + capacity, false);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
done += pfns;
*last_pfn = (dest_pfn + pfns);
xen: populate correct number of pages when across mem boundary (v2) When populate pages across a mem boundary at bootup, the page count populated isn't correct. This is due to mem populated to non-mem region and ignored. Pfn range is also wrongly aligned when mem boundary isn't page aligned. For a dom0 booted with dom_mem=3368952K(0xcd9ff000-4k) dmesg diff is: [ 0.000000] Freeing 9e-100 pfn range: 98 pages freed [ 0.000000] 1-1 mapping on 9e->100 [ 0.000000] 1-1 mapping on cd9ff->100000 [ 0.000000] Released 98 pages of unused memory [ 0.000000] Set 206435 page(s) to 1-1 mapping -[ 0.000000] Populating cd9fe-cda00 pfn range: 1 pages added +[ 0.000000] Populating cd9fe-cd9ff pfn range: 1 pages added +[ 0.000000] Populating 100000-100061 pfn range: 97 pages added [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 000000000009e000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) [ 0.000000] Xen: 0000000000100000 - 00000000cd9ff000 (usable) [ 0.000000] Xen: 00000000cd9ffc00 - 00000000cda53c00 (ACPI NVS) ... [ 0.000000] Xen: 0000000100000000 - 0000000100061000 (usable) [ 0.000000] Xen: 0000000100061000 - 000000012c000000 (unusable) ... [ 0.000000] MEMBLOCK configuration: ... -[ 0.000000] reserved[0x4] [0x000000cd9ff000-0x000000cd9ffbff], 0xc00 bytes -[ 0.000000] reserved[0x5] [0x00000100000000-0x00000100060fff], 0x61000 bytes Related xen memory layout: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009ec00 (usable) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000cd9ffc00 (usable) Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com> [v2: If xen_do_chunk fail(populate), abort this chunk and any others] Suggested by David, thanks. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-07-18 13:06:39 +08:00
if (pfns < capacity)
break;
credits_left -= pfns;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
}
return done;
}
static void __init xen_set_identity_and_release_chunk(
unsigned long start_pfn, unsigned long end_pfn, unsigned long nr_pages,
unsigned long *released, unsigned long *identity)
{
unsigned long pfn;
/*
* If the PFNs are currently mapped, clear the mappings
* (except for the ISA region which must be 1:1 mapped) to
* release the refcounts (in Xen) on the original frames.
*/
/*
* PVH E820 matches the hypervisor's P2M which means we need to
* account for the proper values of *release and *identity.
*/
for (pfn = start_pfn; !xen_feature(XENFEAT_auto_translated_physmap) &&
pfn <= max_pfn_mapped && pfn < end_pfn; pfn++) {
pte_t pte = __pte_ma(0);
if (pfn < PFN_UP(ISA_END_ADDRESS))
pte = mfn_pte(pfn, PAGE_KERNEL_IO);
(void)HYPERVISOR_update_va_mapping(
(unsigned long)__va(pfn << PAGE_SHIFT), pte, 0);
}
if (start_pfn < nr_pages)
*released += xen_release_chunk(
start_pfn, min(end_pfn, nr_pages));
*identity += set_phys_range_identity(start_pfn, end_pfn);
}
static unsigned long __init xen_set_identity_and_release(
const struct e820entry *list, size_t map_size, unsigned long nr_pages)
{
phys_addr_t start = 0;
unsigned long released = 0;
unsigned long identity = 0;
const struct e820entry *entry;
int i;
/*
* Combine non-RAM regions and gaps until a RAM region (or the
* end of the map) is reached, then set the 1:1 map and
* release the pages (if available) in those non-RAM regions.
*
* The combined non-RAM regions are rounded to a whole number
* of pages so any partial pages are accessible via the 1:1
* mapping. This is needed for some BIOSes that put (for
* example) the DMI tables in a reserved region that begins on
* a non-page boundary.
*/
for (i = 0, entry = list; i < map_size; i++, entry++) {
phys_addr_t end = entry->addr + entry->size;
if (entry->type == E820_RAM || i == map_size - 1) {
unsigned long start_pfn = PFN_DOWN(start);
unsigned long end_pfn = PFN_UP(end);
if (entry->type == E820_RAM)
end_pfn = PFN_UP(entry->addr);
if (start_pfn < end_pfn)
xen_set_identity_and_release_chunk(
start_pfn, end_pfn, nr_pages,
&released, &identity);
start = end;
}
}
if (released)
printk(KERN_INFO "Released %lu pages of unused memory\n", released);
if (identity)
printk(KERN_INFO "Set %ld page(s) to 1-1 mapping\n", identity);
return released;
}
static unsigned long __init xen_get_max_pages(void)
{
unsigned long max_pages = MAX_DOMAIN_PAGES;
domid_t domid = DOMID_SELF;
int ret;
xen: only limit memory map to maximum reservation for domain 0. d312ae878b6a "xen: use maximum reservation to limit amount of usable RAM" clamped the total amount of RAM to the current maximum reservation. This is correct for dom0 but is not correct for guest domains. In order to boot a guest "pre-ballooned" (e.g. with memory=1G but maxmem=2G) in order to allow for future memory expansion the guest must derive max_pfn from the e820 provided by the toolstack and not the current maximum reservation (which can reflect only the current maximum, not the guest lifetime max). The existing algorithm already behaves this correctly if we do not artificially limit the maximum number of pages for the guest case. For a guest booted with maxmem=512, memory=128 this results in: [ 0.000000] BIOS-provided physical RAM map: [ 0.000000] Xen: 0000000000000000 - 00000000000a0000 (usable) [ 0.000000] Xen: 00000000000a0000 - 0000000000100000 (reserved) -[ 0.000000] Xen: 0000000000100000 - 0000000008100000 (usable) -[ 0.000000] Xen: 0000000008100000 - 0000000020800000 (unusable) +[ 0.000000] Xen: 0000000000100000 - 0000000020800000 (usable) ... [ 0.000000] NX (Execute Disable) protection: active [ 0.000000] DMI not present or invalid. [ 0.000000] e820 update range: 0000000000000000 - 0000000000010000 (usable) ==> (reserved) [ 0.000000] e820 remove range: 00000000000a0000 - 0000000000100000 (usable) -[ 0.000000] last_pfn = 0x8100 max_arch_pfn = 0x1000000 +[ 0.000000] last_pfn = 0x20800 max_arch_pfn = 0x1000000 [ 0.000000] initial memory mapped : 0 - 027ff000 [ 0.000000] Base memory trampoline at [c009f000] 9f000 size 4096 -[ 0.000000] init_memory_mapping: 0000000000000000-0000000008100000 -[ 0.000000] 0000000000 - 0008100000 page 4k -[ 0.000000] kernel direct mapping tables up to 8100000 @ 27bb000-27ff000 +[ 0.000000] init_memory_mapping: 0000000000000000-0000000020800000 +[ 0.000000] 0000000000 - 0020800000 page 4k +[ 0.000000] kernel direct mapping tables up to 20800000 @ 26f8000-27ff000 [ 0.000000] xen: setting RW the range 27e8000 - 27ff000 [ 0.000000] 0MB HIGHMEM available. -[ 0.000000] 129MB LOWMEM available. -[ 0.000000] mapped low ram: 0 - 08100000 -[ 0.000000] low ram: 0 - 08100000 +[ 0.000000] 520MB LOWMEM available. +[ 0.000000] mapped low ram: 0 - 20800000 +[ 0.000000] low ram: 0 - 20800000 With this change "xl mem-set <domain> 512M" will successfully increase the guest RAM (by reducing the balloon). There is no change for dom0. Reported-and-Tested-by: George Shuklin <george.shuklin@gmail.com> Signed-off-by: Ian Campbell <ian.campbell@citrix.com> Cc: stable@kernel.org Reviewed-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2011-12-14 20:16:08 +08:00
/*
* For the initial domain we use the maximum reservation as
* the maximum page.
*
* For guest domains the current maximum reservation reflects
* the current maximum rather than the static maximum. In this
* case the e820 map provided to us will cover the static
* maximum region.
*/
if (xen_initial_domain()) {
ret = HYPERVISOR_memory_op(XENMEM_maximum_reservation, &domid);
if (ret > 0)
max_pages = ret;
}
return min(max_pages, MAX_DOMAIN_PAGES);
}
static void xen_align_and_add_e820_region(u64 start, u64 size, int type)
{
u64 end = start + size;
/* Align RAM regions to page boundaries. */
if (type == E820_RAM) {
start = PAGE_ALIGN(start);
end &= ~((u64)PAGE_SIZE - 1);
}
e820_add_region(start, end - start, type);
}
x86/xen: do not identity map UNUSABLE regions in the machine E820 If there are UNUSABLE regions in the machine memory map, dom0 will attempt to map them 1:1 which is not permitted by Xen and the kernel will crash. There isn't anything interesting in the UNUSABLE region that the dom0 kernel needs access to so we can avoid making the 1:1 mapping and treat it as RAM. We only do this for dom0, as that is where tboot case shows up. A PV domU could have an UNUSABLE region in its pseudo-physical map and would need to be handled in another patch. This fixes a boot failure on hosts with tboot. tboot marks a region in the e820 map as unusable and the dom0 kernel would attempt to map this region and Xen does not permit unusable regions to be mapped by guests. (XEN) 0000000000000000 - 0000000000060000 (usable) (XEN) 0000000000060000 - 0000000000068000 (reserved) (XEN) 0000000000068000 - 000000000009e000 (usable) (XEN) 0000000000100000 - 0000000000800000 (usable) (XEN) 0000000000800000 - 0000000000972000 (unusable) tboot marked this region as unusable. (XEN) 0000000000972000 - 00000000cf200000 (usable) (XEN) 00000000cf200000 - 00000000cf38f000 (reserved) (XEN) 00000000cf38f000 - 00000000cf3ce000 (ACPI data) (XEN) 00000000cf3ce000 - 00000000d0000000 (reserved) (XEN) 00000000e0000000 - 00000000f0000000 (reserved) (XEN) 00000000fe000000 - 0000000100000000 (reserved) (XEN) 0000000100000000 - 0000000630000000 (usable) Signed-off-by: David Vrabel <david.vrabel@citrix.com> [v1: Altered the patch and description with domU's with UNUSABLE regions] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2013-08-16 22:42:55 +08:00
void xen_ignore_unusable(struct e820entry *list, size_t map_size)
{
struct e820entry *entry;
unsigned int i;
for (i = 0, entry = list; i < map_size; i++, entry++) {
if (entry->type == E820_UNUSABLE)
entry->type = E820_RAM;
}
}
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
/**
* machine_specific_memory_setup - Hook for machine specific memory setup.
**/
char * __init xen_memory_setup(void)
{
static struct e820entry map[E820MAX] __initdata;
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
unsigned long max_pfn = xen_start_info->nr_pages;
unsigned long long mem_end;
int rc;
struct xen_memory_map memmap;
unsigned long max_pages;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
unsigned long last_pfn = 0;
unsigned long extra_pages = 0;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
unsigned long populated;
int i;
int op;
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
max_pfn = min(MAX_DOMAIN_PAGES, max_pfn);
mem_end = PFN_PHYS(max_pfn);
memmap.nr_entries = E820MAX;
set_xen_guest_handle(memmap.buffer, map);
op = xen_initial_domain() ?
XENMEM_machine_memory_map :
XENMEM_memory_map;
rc = HYPERVISOR_memory_op(op, &memmap);
if (rc == -ENOSYS) {
BUG_ON(xen_initial_domain());
memmap.nr_entries = 1;
map[0].addr = 0ULL;
map[0].size = mem_end;
/* 8MB slack (to balance backend allocations). */
map[0].size += 8ULL << 20;
map[0].type = E820_RAM;
rc = 0;
}
BUG_ON(rc);
x86/xen: do not identity map UNUSABLE regions in the machine E820 If there are UNUSABLE regions in the machine memory map, dom0 will attempt to map them 1:1 which is not permitted by Xen and the kernel will crash. There isn't anything interesting in the UNUSABLE region that the dom0 kernel needs access to so we can avoid making the 1:1 mapping and treat it as RAM. We only do this for dom0, as that is where tboot case shows up. A PV domU could have an UNUSABLE region in its pseudo-physical map and would need to be handled in another patch. This fixes a boot failure on hosts with tboot. tboot marks a region in the e820 map as unusable and the dom0 kernel would attempt to map this region and Xen does not permit unusable regions to be mapped by guests. (XEN) 0000000000000000 - 0000000000060000 (usable) (XEN) 0000000000060000 - 0000000000068000 (reserved) (XEN) 0000000000068000 - 000000000009e000 (usable) (XEN) 0000000000100000 - 0000000000800000 (usable) (XEN) 0000000000800000 - 0000000000972000 (unusable) tboot marked this region as unusable. (XEN) 0000000000972000 - 00000000cf200000 (usable) (XEN) 00000000cf200000 - 00000000cf38f000 (reserved) (XEN) 00000000cf38f000 - 00000000cf3ce000 (ACPI data) (XEN) 00000000cf3ce000 - 00000000d0000000 (reserved) (XEN) 00000000e0000000 - 00000000f0000000 (reserved) (XEN) 00000000fe000000 - 0000000100000000 (reserved) (XEN) 0000000100000000 - 0000000630000000 (usable) Signed-off-by: David Vrabel <david.vrabel@citrix.com> [v1: Altered the patch and description with domU's with UNUSABLE regions] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2013-08-16 22:42:55 +08:00
/*
* Xen won't allow a 1:1 mapping to be created to UNUSABLE
* regions, so if we're using the machine memory map leave the
* region as RAM as it is in the pseudo-physical map.
*
* UNUSABLE regions in domUs are not handled and will need
* a patch in the future.
*/
if (xen_initial_domain())
xen_ignore_unusable(map, memmap.nr_entries);
/* Make sure the Xen-supplied memory map is well-ordered. */
sanitize_e820_map(map, memmap.nr_entries, &memmap.nr_entries);
max_pages = xen_get_max_pages();
if (max_pages > max_pfn)
extra_pages += max_pages - max_pfn;
/*
* Set P2M for all non-RAM pages and E820 gaps to be identity
* type PFNs. Any RAM pages that would be made inaccesible by
* this are first released.
*/
xen_released_pages = xen_set_identity_and_release(
map, memmap.nr_entries, max_pfn);
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
/*
* Populate back the non-RAM pages and E820 gaps that had been
* released. */
populated = xen_populate_chunk(map, memmap.nr_entries,
max_pfn, &last_pfn, xen_released_pages);
xen_released_pages -= populated;
extra_pages += xen_released_pages;
xen/setup: Populate freed MFNs from non-RAM E820 entries and gaps to E820 RAM When the Xen hypervisor boots a PV kernel it hands it two pieces of information: nr_pages and a made up E820 entry. The nr_pages value defines the range from zero to nr_pages of PFNs which have a valid Machine Frame Number (MFN) underneath it. The E820 mirrors that (with the VGA hole): BIOS-provided physical RAM map: Xen: 0000000000000000 - 00000000000a0000 (usable) Xen: 00000000000a0000 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000080800000 (usable) The fun comes when a PV guest that is run with a machine E820 - that can either be the initial domain or a PCI PV guest, where the E820 looks like the normal thing: BIOS-provided physical RAM map: Xen: 0000000000000000 - 000000000009e000 (usable) Xen: 000000000009ec00 - 0000000000100000 (reserved) Xen: 0000000000100000 - 0000000020000000 (usable) Xen: 0000000020000000 - 0000000020200000 (reserved) Xen: 0000000020200000 - 0000000040000000 (usable) Xen: 0000000040000000 - 0000000040200000 (reserved) Xen: 0000000040200000 - 00000000bad80000 (usable) Xen: 00000000bad80000 - 00000000badc9000 (ACPI NVS) .. With that overlaying the nr_pages directly on the E820 does not work as there are gaps and non-RAM regions that won't be used by the memory allocator. The 'xen_release_chunk' helps with that by punching holes in the P2M (PFN to MFN lookup tree) for those regions and tells us that: Freeing 20000-20200 pfn range: 512 pages freed Freeing 40000-40200 pfn range: 512 pages freed Freeing bad80-badf4 pfn range: 116 pages freed Freeing badf6-bae7f pfn range: 137 pages freed Freeing bb000-100000 pfn range: 282624 pages freed Released 283999 pages of unused memory Those 283999 pages are subtracted from the nr_pages and are returned to the hypervisor. The end result is that the initial domain boots with 1GB less memory as the nr_pages has been subtracted by the amount of pages residing within the PCI hole. It can balloon up to that if desired using 'xl mem-set 0 8092', but the balloon driver is not always compiled in for the initial domain. This patch, implements the populate hypercall (XENMEM_populate_physmap) which increases the the domain with the same amount of pages that were released. The other solution (that did not work) was to transplant the MFN in the P2M tree - the ones that were going to be freed were put in the E820_RAM regions past the nr_pages. But the modifications to the M2P array (the other side of creating PTEs) were not carried away. As the hypervisor is the only one capable of modifying that and the only two hypercalls that would do this are: the update_va_mapping (which won't work, as during initial bootup only PFNs up to nr_pages are mapped in the guest) or via the populate hypercall. The end result is that the kernel can now boot with the nr_pages without having to subtract the 283999 pages. On a 8GB machine, with various dom0_mem= parameters this is what we get: no dom0_mem -Memory: 6485264k/9435136k available (5817k kernel code, 1136060k absent, 1813812k reserved, 2899k data, 696k init) +Memory: 7619036k/9435136k available (5817k kernel code, 1136060k absent, 680040k reserved, 2899k data, 696k init) dom0_mem=3G -Memory: 2616536k/9435136k available (5817k kernel code, 1136060k absent, 5682540k reserved, 2899k data, 696k init) +Memory: 2703776k/9435136k available (5817k kernel code, 1136060k absent, 5595300k reserved, 2899k data, 696k init) dom0_mem=max:3G -Memory: 2696732k/4281724k available (5817k kernel code, 1136060k absent, 448932k reserved, 2899k data, 696k init) +Memory: 2702204k/4281724k available (5817k kernel code, 1136060k absent, 443460k reserved, 2899k data, 696k init) And the 'xm list' or 'xl list' now reflect what the dom0_mem= argument is. Acked-by: David Vrabel <david.vrabel@citrix.com> [v2: Use populate hypercall] [v3: Remove debug printks] [v4: Simplify code] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-04-06 22:07:11 +08:00
if (last_pfn > max_pfn) {
max_pfn = min(MAX_DOMAIN_PAGES, last_pfn);
mem_end = PFN_PHYS(max_pfn);
}
/*
* Clamp the amount of extra memory to a EXTRA_MEM_RATIO
* factor the base size. On non-highmem systems, the base
* size is the full initial memory allocation; on highmem it
* is limited to the max size of lowmem, so that it doesn't
* get completely filled.
*
* In principle there could be a problem in lowmem systems if
* the initial memory is also very large with respect to
* lowmem, but we won't try to deal with that here.
*/
extra_pages = min(EXTRA_MEM_RATIO * min(max_pfn, PFN_DOWN(MAXMEM)),
extra_pages);
i = 0;
while (i < memmap.nr_entries) {
u64 addr = map[i].addr;
u64 size = map[i].size;
u32 type = map[i].type;
if (type == E820_RAM) {
if (addr < mem_end) {
size = min(size, mem_end - addr);
} else if (extra_pages) {
size = min(size, (u64)extra_pages * PAGE_SIZE);
extra_pages -= size / PAGE_SIZE;
xen_add_extra_mem(addr, size);
} else
type = E820_UNUSABLE;
}
xen_align_and_add_e820_region(addr, size, type);
map[i].addr += size;
map[i].size -= size;
if (map[i].size == 0)
i++;
}
/*
* In domU, the ISA region is normal, usable memory, but we
* reserve ISA memory anyway because too many things poke
* about in there.
*/
e820_add_region(ISA_START_ADDRESS, ISA_END_ADDRESS - ISA_START_ADDRESS,
E820_RESERVED);
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
/*
* Reserve Xen bits:
* - mfn_list
* - xen_start_info
* See comment above "struct start_info" in <xen/interface/xen.h>
* We tried to make the the memblock_reserve more selective so
* that it would be clear what region is reserved. Sadly we ran
* in the problem wherein on a 64-bit hypervisor with a 32-bit
* initial domain, the pt_base has the cr3 value which is not
* neccessarily where the pagetable starts! As Jan put it: "
* Actually, the adjustment turns out to be correct: The page
* tables for a 32-on-64 dom0 get allocated in the order "first L1",
* "first L2", "first L3", so the offset to the page table base is
* indeed 2. When reading xen/include/public/xen.h's comment
* very strictly, this is not a violation (since there nothing is said
* that the first thing in the page table space is pointed to by
* pt_base; I admit that this seems to be implied though, namely
* do I think that it is implied that the page table space is the
* range [pt_base, pt_base + nt_pt_frames), whereas that
* range here indeed is [pt_base - 2, pt_base - 2 + nt_pt_frames),
* which - without a priori knowledge - the kernel would have
* difficulty to figure out)." - so lets just fall back to the
* easy way and reserve the whole region.
*/
memblock_reserve(__pa(xen_start_info->mfn_list),
xen_start_info->pt_base - xen_start_info->mfn_list);
sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
return "Xen";
}
/*
* Set the bit indicating "nosegneg" library variants should be used.
* We only need to bother in pure 32-bit mode; compat 32-bit processes
* can have un-truncated segments, so wrapping around is allowed.
*/
static void __init fiddle_vdso(void)
{
#ifdef CONFIG_X86_32
u32 *mask;
mask = VDSO32_SYMBOL(&vdso32_int80_start, NOTE_MASK);
*mask |= 1 << VDSO_NOTE_NONEGSEG_BIT;
mask = VDSO32_SYMBOL(&vdso32_sysenter_start, NOTE_MASK);
*mask |= 1 << VDSO_NOTE_NONEGSEG_BIT;
#endif
}
x86: delete __cpuinit usage from all x86 files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. Note that some harmless section mismatch warnings may result, since notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c) are flagged as __cpuinit -- so if we remove the __cpuinit from arch specific callers, we will also get section mismatch warnings. As an intermediate step, we intend to turn the linux/init.h cpuinit content into no-ops as early as possible, since that will get rid of these warnings. In any case, they are temporary and harmless. This removes all the arch/x86 uses of the __cpuinit macros from all C files. x86 only had the one __CPUINIT used in assembly files, and it wasn't paired off with a .previous or a __FINIT, so we can delete it directly w/o any corresponding additional change there. [1] https://lkml.org/lkml/2013/5/20/589 Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: x86@kernel.org Acked-by: Ingo Molnar <mingo@kernel.org> Acked-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
static int register_callback(unsigned type, const void *func)
{
struct callback_register callback = {
.type = type,
.address = XEN_CALLBACK(__KERNEL_CS, func),
.flags = CALLBACKF_mask_events,
};
return HYPERVISOR_callback_op(CALLBACKOP_register, &callback);
}
x86: delete __cpuinit usage from all x86 files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. Note that some harmless section mismatch warnings may result, since notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c) are flagged as __cpuinit -- so if we remove the __cpuinit from arch specific callers, we will also get section mismatch warnings. As an intermediate step, we intend to turn the linux/init.h cpuinit content into no-ops as early as possible, since that will get rid of these warnings. In any case, they are temporary and harmless. This removes all the arch/x86 uses of the __cpuinit macros from all C files. x86 only had the one __CPUINIT used in assembly files, and it wasn't paired off with a .previous or a __FINIT, so we can delete it directly w/o any corresponding additional change there. [1] https://lkml.org/lkml/2013/5/20/589 Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: x86@kernel.org Acked-by: Ingo Molnar <mingo@kernel.org> Acked-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
void xen_enable_sysenter(void)
{
int ret;
unsigned sysenter_feature;
#ifdef CONFIG_X86_32
sysenter_feature = X86_FEATURE_SEP;
#else
sysenter_feature = X86_FEATURE_SYSENTER32;
#endif
if (!boot_cpu_has(sysenter_feature))
return;
ret = register_callback(CALLBACKTYPE_sysenter, xen_sysenter_target);
if(ret != 0)
setup_clear_cpu_cap(sysenter_feature);
}
x86: delete __cpuinit usage from all x86 files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. Note that some harmless section mismatch warnings may result, since notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c) are flagged as __cpuinit -- so if we remove the __cpuinit from arch specific callers, we will also get section mismatch warnings. As an intermediate step, we intend to turn the linux/init.h cpuinit content into no-ops as early as possible, since that will get rid of these warnings. In any case, they are temporary and harmless. This removes all the arch/x86 uses of the __cpuinit macros from all C files. x86 only had the one __CPUINIT used in assembly files, and it wasn't paired off with a .previous or a __FINIT, so we can delete it directly w/o any corresponding additional change there. [1] https://lkml.org/lkml/2013/5/20/589 Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: x86@kernel.org Acked-by: Ingo Molnar <mingo@kernel.org> Acked-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2013-06-19 06:23:59 +08:00
void xen_enable_syscall(void)
{
#ifdef CONFIG_X86_64
int ret;
ret = register_callback(CALLBACKTYPE_syscall, xen_syscall_target);
if (ret != 0) {
printk(KERN_ERR "Failed to set syscall callback: %d\n", ret);
/* Pretty fatal; 64-bit userspace has no other
mechanism for syscalls. */
}
if (boot_cpu_has(X86_FEATURE_SYSCALL32)) {
ret = register_callback(CALLBACKTYPE_syscall32,
xen_syscall32_target);
if (ret != 0)
setup_clear_cpu_cap(X86_FEATURE_SYSCALL32);
}
#endif /* CONFIG_X86_64 */
}
void xen_enable_nmi(void)
xen: Support 64-bit PV guest receiving NMIs This is based on a patch that Zhenzhong Duan had sent - which was missing some of the remaining pieces. The kernel has the logic to handle Xen-type-exceptions using the paravirt interface in the assembler code (see PARAVIRT_ADJUST_EXCEPTION_FRAME - pv_irq_ops.adjust_exception_frame and and INTERRUPT_RETURN - pv_cpu_ops.iret). That means the nmi handler (and other exception handlers) use the hypervisor iret. The other changes that would be neccessary for this would be to translate the NMI_VECTOR to one of the entries on the ipi_vector and make xen_send_IPI_mask_allbutself use different events. Fortunately for us commit 1db01b4903639fcfaec213701a494fe3fb2c490b (xen: Clean up apic ipi interface) implemented this and we piggyback on the cleanup such that the apic IPI interface will pass the right vector value for NMI. With this patch we can trigger NMIs within a PV guest (only tested x86_64). For this to work with normal PV guests (not initial domain) we need the domain to be able to use the APIC ops - they are already implemented to use the Xen event channels. For that to be turned on in a PV domU we need to remove the masking of X86_FEATURE_APIC. Incidentally that means kgdb will also now work within a PV guest without using the 'nokgdbroundup' workaround. Note that the 32-bit version is different and this patch does not enable that. CC: Lisa Nguyen <lisa@xenapiadmin.com> CC: Ben Guthro <benjamin.guthro@citrix.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [v1: Fixed up per David Vrabel comments] Reviewed-by: Ben Guthro <benjamin.guthro@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
2013-07-19 23:51:31 +08:00
{
#ifdef CONFIG_X86_64
if (register_callback(CALLBACKTYPE_nmi, (char *)nmi))
xen: Support 64-bit PV guest receiving NMIs This is based on a patch that Zhenzhong Duan had sent - which was missing some of the remaining pieces. The kernel has the logic to handle Xen-type-exceptions using the paravirt interface in the assembler code (see PARAVIRT_ADJUST_EXCEPTION_FRAME - pv_irq_ops.adjust_exception_frame and and INTERRUPT_RETURN - pv_cpu_ops.iret). That means the nmi handler (and other exception handlers) use the hypervisor iret. The other changes that would be neccessary for this would be to translate the NMI_VECTOR to one of the entries on the ipi_vector and make xen_send_IPI_mask_allbutself use different events. Fortunately for us commit 1db01b4903639fcfaec213701a494fe3fb2c490b (xen: Clean up apic ipi interface) implemented this and we piggyback on the cleanup such that the apic IPI interface will pass the right vector value for NMI. With this patch we can trigger NMIs within a PV guest (only tested x86_64). For this to work with normal PV guests (not initial domain) we need the domain to be able to use the APIC ops - they are already implemented to use the Xen event channels. For that to be turned on in a PV domU we need to remove the masking of X86_FEATURE_APIC. Incidentally that means kgdb will also now work within a PV guest without using the 'nokgdbroundup' workaround. Note that the 32-bit version is different and this patch does not enable that. CC: Lisa Nguyen <lisa@xenapiadmin.com> CC: Ben Guthro <benjamin.guthro@citrix.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [v1: Fixed up per David Vrabel comments] Reviewed-by: Ben Guthro <benjamin.guthro@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
2013-07-19 23:51:31 +08:00
BUG();
#endif
}
void __init xen_pvmmu_arch_setup(void)
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
{
HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_4gb_segments);
HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_writable_pagetables);
HYPERVISOR_vm_assist(VMASST_CMD_enable,
VMASST_TYPE_pae_extended_cr3);
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
if (register_callback(CALLBACKTYPE_event, xen_hypervisor_callback) ||
register_callback(CALLBACKTYPE_failsafe, xen_failsafe_callback))
BUG();
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
xen_enable_sysenter();
xen_enable_syscall();
xen: Support 64-bit PV guest receiving NMIs This is based on a patch that Zhenzhong Duan had sent - which was missing some of the remaining pieces. The kernel has the logic to handle Xen-type-exceptions using the paravirt interface in the assembler code (see PARAVIRT_ADJUST_EXCEPTION_FRAME - pv_irq_ops.adjust_exception_frame and and INTERRUPT_RETURN - pv_cpu_ops.iret). That means the nmi handler (and other exception handlers) use the hypervisor iret. The other changes that would be neccessary for this would be to translate the NMI_VECTOR to one of the entries on the ipi_vector and make xen_send_IPI_mask_allbutself use different events. Fortunately for us commit 1db01b4903639fcfaec213701a494fe3fb2c490b (xen: Clean up apic ipi interface) implemented this and we piggyback on the cleanup such that the apic IPI interface will pass the right vector value for NMI. With this patch we can trigger NMIs within a PV guest (only tested x86_64). For this to work with normal PV guests (not initial domain) we need the domain to be able to use the APIC ops - they are already implemented to use the Xen event channels. For that to be turned on in a PV domU we need to remove the masking of X86_FEATURE_APIC. Incidentally that means kgdb will also now work within a PV guest without using the 'nokgdbroundup' workaround. Note that the 32-bit version is different and this patch does not enable that. CC: Lisa Nguyen <lisa@xenapiadmin.com> CC: Ben Guthro <benjamin.guthro@citrix.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [v1: Fixed up per David Vrabel comments] Reviewed-by: Ben Guthro <benjamin.guthro@citrix.com> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
2013-07-19 23:51:31 +08:00
xen_enable_nmi();
}
/* This function is not called for HVM domains */
void __init xen_arch_setup(void)
{
xen_panic_handler_init();
if (!xen_feature(XENFEAT_auto_translated_physmap))
xen_pvmmu_arch_setup();
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
#ifdef CONFIG_ACPI
if (!(xen_start_info->flags & SIF_INITDOMAIN)) {
printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
disable_acpi();
}
#endif
memcpy(boot_command_line, xen_start_info->cmd_line,
MAX_GUEST_CMDLINE > COMMAND_LINE_SIZE ?
COMMAND_LINE_SIZE : MAX_GUEST_CMDLINE);
/* Set up idle, making sure it calls safe_halt() pvop */
disable_cpuidle();
disable_cpufreq();
WARN_ON(xen_set_default_idle());
fiddle_vdso();
xen/boot: Disable NUMA for PV guests. The hypervisor is in charge of allocating the proper "NUMA" memory and dealing with the CPU scheduler to keep them bound to the proper NUMA node. The PV guests (and PVHVM) have no inkling of where they run and do not need to know that right now. In the future we will need to inject NUMA configuration data (if a guest spans two or more NUMA nodes) so that the kernel can make the right choices. But those patches are not yet present. In the meantime, disable the NUMA capability in the PV guest, which also fixes a bootup issue. Andre says: "we see Dom0 crashes due to the kernel detecting the NUMA topology not by ACPI, but directly from the northbridge (CONFIG_AMD_NUMA). This will detect the actual NUMA config of the physical machine, but will crash about the mismatch with Dom0's virtual memory. Variation of the theme: Dom0 sees what it's not supposed to see. This happens with the said config option enabled and on a machine where this scanning is still enabled (K8 and Fam10h, not Bulldozer class) We have this dump then: NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10 Scanning NUMA topology in Northbridge 24 Number of physical nodes 4 Node 0 MemBase 0000000000000000 Limit 0000000040000000 Node 1 MemBase 0000000040000000 Limit 0000000138000000 Node 2 MemBase 0000000138000000 Limit 00000001f8000000 Node 3 MemBase 00000001f8000000 Limit 0000000238000000 Initmem setup node 0 0000000000000000-0000000040000000 NODE_DATA [000000003ffd9000 - 000000003fffffff] Initmem setup node 1 0000000040000000-0000000138000000 NODE_DATA [0000000137fd9000 - 0000000137ffffff] Initmem setup node 2 0000000138000000-00000001f8000000 NODE_DATA [00000001f095e000 - 00000001f0984fff] Initmem setup node 3 00000001f8000000-0000000238000000 Cannot find 159744 bytes in node 3 BUG: unable to handle kernel NULL pointer dereference at (null) IP: [<ffffffff81d220e6>] __alloc_bootmem_node+0x43/0x96 Pid: 0, comm: swapper Not tainted 3.3.6 #1 AMD Dinar/Dinar RIP: e030:[<ffffffff81d220e6>] [<ffffffff81d220e6>] __alloc_bootmem_node+0x43/0x96 .. snip.. [<ffffffff81d23024>] sparse_early_usemaps_alloc_node+0x64/0x178 [<ffffffff81d23348>] sparse_init+0xe4/0x25a [<ffffffff81d16840>] paging_init+0x13/0x22 [<ffffffff81d07fbb>] setup_arch+0x9c6/0xa9b [<ffffffff81683954>] ? printk+0x3c/0x3e [<ffffffff81d01a38>] start_kernel+0xe5/0x468 [<ffffffff81d012cf>] x86_64_start_reservations+0xba/0xc1 [<ffffffff81007153>] ? xen_setup_runstate_info+0x2c/0x36 [<ffffffff81d050ee>] xen_start_kernel+0x565/0x56c " so we just disable NUMA scanning by setting numa_off=1. CC: stable@vger.kernel.org Reported-and-Tested-by: Andre Przywara <andre.przywara@amd.com> Acked-by: Andre Przywara <andre.przywara@amd.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
2012-08-17 22:22:37 +08:00
#ifdef CONFIG_NUMA
numa_off = 1;
#endif
xen: Core Xen implementation This patch is a rollup of all the core pieces of the Xen implementation, including: - booting and setup - pagetable setup - privileged instructions - segmentation - interrupt flags - upcalls - multicall batching BOOTING AND SETUP The vmlinux image is decorated with ELF notes which tell the Xen domain builder what the kernel's requirements are; the domain builder then constructs the address space accordingly and starts the kernel. Xen has its own entrypoint for the kernel (contained in an ELF note). The ELF notes are set up by xen-head.S, which is included into head.S. In principle it could be linked separately, but it seems to provoke lots of binutils bugs. Because the domain builder starts the kernel in a fairly sane state (32-bit protected mode, paging enabled, flat segments set up), there's not a lot of setup needed before starting the kernel proper. The main steps are: 1. Install the Xen paravirt_ops, which is simply a matter of a structure assignment. 2. Set init_mm to use the Xen-supplied pagetables (analogous to the head.S generated pagetables in a native boot). 3. Reserve address space for Xen, since it takes a chunk at the top of the address space for its own use. 4. Call start_kernel() PAGETABLE SETUP Once we hit the main kernel boot sequence, it will end up calling back via paravirt_ops to set up various pieces of Xen specific state. One of the critical things which requires a bit of extra care is the construction of the initial init_mm pagetable. Because Xen places tight constraints on pagetables (an active pagetable must always be valid, and must always be mapped read-only to the guest domain), we need to be careful when constructing the new pagetable to keep these constraints in mind. It turns out that the easiest way to do this is use the initial Xen-provided pagetable as a template, and then just insert new mappings for memory where a mapping doesn't already exist. This means that during pagetable setup, it uses a special version of xen_set_pte which ignores any attempt to remap a read-only page as read-write (since Xen will map its own initial pagetable as RO), but lets other changes to the ptes happen, so that things like NX are set properly. PRIVILEGED INSTRUCTIONS AND SEGMENTATION When the kernel runs under Xen, it runs in ring 1 rather than ring 0. This means that it is more privileged than user-mode in ring 3, but it still can't run privileged instructions directly. Non-performance critical instructions are dealt with by taking a privilege exception and trapping into the hypervisor and emulating the instruction, but more performance-critical instructions have their own specific paravirt_ops. In many cases we can avoid having to do any hypercalls for these instructions, or the Xen implementation is quite different from the normal native version. The privileged instructions fall into the broad classes of: Segmentation: setting up the GDT and the GDT entries, LDT, TLS and so on. Xen doesn't allow the GDT to be directly modified; all GDT updates are done via hypercalls where the new entries can be validated. This is important because Xen uses segment limits to prevent the guest kernel from damaging the hypervisor itself. Traps and exceptions: Xen uses a special format for trap entrypoints, so when the kernel wants to set an IDT entry, it needs to be converted to the form Xen expects. Xen sets int 0x80 up specially so that the trap goes straight from userspace into the guest kernel without going via the hypervisor. sysenter isn't supported. Kernel stack: The esp0 entry is extracted from the tss and provided to Xen. TLB operations: the various TLB calls are mapped into corresponding Xen hypercalls. Control registers: all the control registers are privileged. The most important is cr3, which points to the base of the current pagetable, and we handle it specially. Another instruction we treat specially is CPUID, even though its not privileged. We want to control what CPU features are visible to the rest of the kernel, and so CPUID ends up going into a paravirt_op. Xen implements this mainly to disable the ACPI and APIC subsystems. INTERRUPT FLAGS Xen maintains its own separate flag for masking events, which is contained within the per-cpu vcpu_info structure. Because the guest kernel runs in ring 1 and not 0, the IF flag in EFLAGS is completely ignored (and must be, because even if a guest domain disables interrupts for itself, it can't disable them overall). (A note on terminology: "events" and interrupts are effectively synonymous. However, rather than using an "enable flag", Xen uses a "mask flag", which blocks event delivery when it is non-zero.) There are paravirt_ops for each of cli/sti/save_fl/restore_fl, which are implemented to manage the Xen event mask state. The only thing worth noting is that when events are unmasked, we need to explicitly see if there's a pending event and call into the hypervisor to make sure it gets delivered. UPCALLS Xen needs a couple of upcall (or callback) functions to be implemented by each guest. One is the event upcalls, which is how events (interrupts, effectively) are delivered to the guests. The other is the failsafe callback, which is used to report errors in either reloading a segment register, or caused by iret. These are implemented in i386/kernel/entry.S so they can jump into the normal iret_exc path when necessary. MULTICALL BATCHING Xen provides a multicall mechanism, which allows multiple hypercalls to be issued at once in order to mitigate the cost of trapping into the hypervisor. This is particularly useful for context switches, since the 4-5 hypercalls they would normally need (reload cr3, update TLS, maybe update LDT) can be reduced to one. This patch implements a generic batching mechanism for hypercalls, which gets used in many places in the Xen code. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Ian Pratt <ian.pratt@xensource.com> Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Cc: Adrian Bunk <bunk@stusta.de>
2007-07-18 09:37:04 +08:00
}