^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) ==================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) Memory Attribute Aliasing on IA-64
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) ==================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) Bjorn Helgaas <bjorn.helgaas@hp.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) May 4, 2006
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) Memory Attributes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) =================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) Itanium supports several attributes for virtual memory references.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) The attribute is part of the virtual translation, i.e., it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) contained in the TLB entry. The ones of most interest to the Linux
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) kernel are:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) == ======================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) WB Write-back (cacheable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) UC Uncacheable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) WC Write-coalescing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) == ======================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) System memory typically uses the WB attribute. The UC attribute is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) used for memory-mapped I/O devices. The WC attribute is uncacheable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) like UC is, but writes may be delayed and combined to increase
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) performance for things like frame buffers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) The Itanium architecture requires that we avoid accessing the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) page with both a cacheable mapping and an uncacheable mapping[1].
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) The design of the chipset determines which attributes are supported
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) on which regions of the address space. For example, some chipsets
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) support either WB or UC access to main memory, while others support
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) only WB access.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) Memory Map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) ==========
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) Platform firmware describes the physical memory map and the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) supported attributes for each region. At boot-time, the kernel uses
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) the EFI GetMemoryMap() interface. ACPI can also describe memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) devices and the attributes they support, but Linux/ia64 currently
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) doesn't use this information.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) The kernel uses the efi_memmap table returned from GetMemoryMap() to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) learn the attributes supported by each region of physical address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) space. Unfortunately, this table does not completely describe the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) address space because some machines omit some or all of the MMIO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) regions from the map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) The kernel maintains another table, kern_memmap, which describes the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) memory Linux is actually using and the attribute for each region.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) This contains only system memory; it does not contain MMIO space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) The kern_memmap table typically contains only a subset of the system
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) memory described by the efi_memmap. Linux/ia64 can't use all memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) in the system because of constraints imposed by the identity mapping
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) scheme.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) The efi_memmap table is preserved unmodified because the original
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) boot-time information is required for kexec.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) Kernel Identify Mappings
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) ========================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) Linux/ia64 identity mappings are done with large pages, currently
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) either 16MB or 64MB, referred to as "granules." Cacheable mappings
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) are speculative[2], so the processor can read any location in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) page at any time, independent of the programmer's intentions. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) means that to avoid attribute aliasing, Linux can create a cacheable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) identity mapping only when the entire granule supports cacheable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) access.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) Therefore, kern_memmap contains only full granule-sized regions that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) can referenced safely by an identity mapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) Uncacheable mappings are not speculative, so the processor will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) generate UC accesses only to locations explicitly referenced by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) software. This allows UC identity mappings to cover granules that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) are only partially populated, or populated with a combination of UC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) and WB regions.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) User Mappings
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) =============
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) User mappings are typically done with 16K or 64K pages. The smaller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) page size allows more flexibility because only 16K or 64K has to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) homogeneous with respect to memory attributes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) Potential Attribute Aliasing Cases
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) ==================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) There are several ways the kernel creates new mappings:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) mmap of /dev/mem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) ----------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) This uses remap_pfn_range(), which creates user mappings. These
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) mappings may be either WB or UC. If the region being mapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) happens to be in kern_memmap, meaning that it may also be mapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) by a kernel identity mapping, the user mapping must use the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) attribute as the kernel mapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) If the region is not in kern_memmap, the user mapping should use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) an attribute reported as being supported in the EFI memory map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) Since the EFI memory map does not describe MMIO on some
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) machines, this should use an uncacheable mapping as a fallback.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) mmap of /sys/class/pci_bus/.../legacy_mem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) -----------------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) This is very similar to mmap of /dev/mem, except that legacy_mem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) only allows mmap of the one megabyte "legacy MMIO" area for a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) specific PCI bus. Typically this is the first megabyte of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) physical address space, but it may be different on machines with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) several VGA devices.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) "X" uses this to access VGA frame buffers. Using legacy_mem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) rather than /dev/mem allows multiple instances of X to talk to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) different VGA cards.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) The /dev/mem mmap constraints apply.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) mmap of /proc/bus/pci/.../??.?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) ------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) This is an MMIO mmap of PCI functions, which additionally may or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) may not be requested as using the WC attribute.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) If WC is requested, and the region in kern_memmap is either WC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) or UC, and the EFI memory map designates the region as WC, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) the WC mapping is allowed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) Otherwise, the user mapping must use the same attribute as the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) kernel mapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) read/write of /dev/mem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) ----------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) This uses copy_from_user(), which implicitly uses a kernel
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) identity mapping. This is obviously safe for things in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) kern_memmap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) There may be corner cases of things that are not in kern_memmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) but could be accessed this way. For example, registers in MMIO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) space are not in kern_memmap, but could be accessed with a UC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) mapping. This would not cause attribute aliasing. But
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) registers typically can be accessed only with four-byte or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) eight-byte accesses, and the copy_from_user() path doesn't allow
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) any control over the access size, so this would be dangerous.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) ioremap()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) ---------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) This returns a mapping for use inside the kernel.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) If the region is in kern_memmap, we should use the attribute
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) specified there.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) If the EFI memory map reports that the entire granule supports
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) WB, we should use that (granules that are partially reserved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) or occupied by firmware do not appear in kern_memmap).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) If the granule contains non-WB memory, but we can cover the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) region safely with kernel page table mappings, we can use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) ioremap_page_range() as most other architectures do.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) Failing all of the above, we have to fall back to a UC mapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) Past Problem Cases
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) ==================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) mmap of various MMIO regions from /dev/mem by "X" on Intel platforms
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) --------------------------------------------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) The EFI memory map may not report these MMIO regions.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) These must be allowed so that X will work. This means that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) when the EFI memory map is incomplete, every /dev/mem mmap must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) succeed. It may create either WB or UC user mappings, depending
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) on whether the region is in kern_memmap or the EFI memory map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) mmap of 0x0-0x9FFFF /dev/mem by "hwinfo" on HP sx1000 with VGA enabled
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) ----------------------------------------------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) The EFI memory map reports the following attributes:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) =============== ======= ==================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 0x00000-0x9FFFF WB only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 0xA0000-0xBFFFF UC only (VGA frame buffer)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) 0xC0000-0xFFFFF WB only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) =============== ======= ==================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) This mmap is done with user pages, not kernel identity mappings,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) so it is safe to use WB mappings.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) The kernel VGA driver may ioremap the VGA frame buffer at 0xA0000,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) which uses a granule-sized UC mapping. This granule will cover some
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) WB-only memory, but since UC is non-speculative, the processor will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) never generate an uncacheable reference to the WB-only areas unless
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) the driver explicitly touches them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) mmap of 0x0-0xFFFFF legacy_mem by "X"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) -------------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) If the EFI memory map reports that the entire range supports the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) same attributes, we can allow the mmap (and we will prefer WB if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) supported, as is the case with HP sx[12]000 machines with VGA
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) disabled).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) If EFI reports the range as partly WB and partly UC (as on sx[12]000
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) machines with VGA enabled), we must fail the mmap because there's no
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) safe attribute to use.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) If EFI reports some of the range but not all (as on Intel firmware
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) that doesn't report the VGA frame buffer at all), we should fail the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) mmap and force the user to map just the specific region of interest.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) mmap of 0xA0000-0xBFFFF legacy_mem by "X" on HP sx1000 with VGA disabled
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) ------------------------------------------------------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) The EFI memory map reports the following attributes::
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) 0x00000-0xFFFFF WB only (no VGA MMIO hole)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) This is a special case of the previous case, and the mmap should
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) fail for the same reason as above.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) read of /sys/devices/.../rom
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) ----------------------------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) For VGA devices, this may cause an ioremap() of 0xC0000. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) used to be done with a UC mapping, because the VGA frame buffer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) at 0xA0000 prevents use of a WB granule. The UC mapping causes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) an MCA on HP sx[12]000 chipsets.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) We should use WB page table mappings to avoid covering the VGA
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) frame buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) Notes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) =====
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) [1] SDM rev 2.2, vol 2, sec 4.4.1.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) [2] SDM rev 2.2, vol 2, sec 4.4.6.
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags