| Message ID | 20210115083255.12744-2-rppt@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | x86/setup: consolidate early memory reservations | expand |
On 15.01.21 09:32, Mike Rapoport wrote: > From: Mike Rapoport <rppt@linux.ibm.com> > > The early reservations of memory areas used by the firmware, bootloader, > kernel text and data are spread over setup_arch(). Moreover, some of them > happen *after* memblock allocations, e.g trim_platform_memory_ranges() and > trim_low_memory_range() are called after reserve_real_mode() that allocates > memory. > > We did not observe corruption of these memory regions because memblock > always allocates memory either from the end of memory (in top-down mode) or > above the kernel image (in bottom-up mode). However, the bottom up mode is > going to be updated to span the entire memory [1] to avoid limitations > caused by KASLR. > > Consolidate early memory reservations in a dedicated function to improve > robustness against future changes. Having the early reservations in one > place also makes it clearer what memory must be reserved before we allow > memblock allocations. > > [1] https://lore.kernel.org/lkml/20201217201214.3414100-2-guro@fb.com > > Signed-off-by: Mike Rapoport <rppt@linux.ibm.com> > --- > arch/x86/kernel/setup.c | 80 ++++++++++++++++++++++------------------- > 1 file changed, 44 insertions(+), 36 deletions(-) > > diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c > index 3412c4595efd..32cd2e790a0a 100644 > --- a/arch/x86/kernel/setup.c > +++ b/arch/x86/kernel/setup.c > @@ -728,7 +728,38 @@ static void __init trim_low_memory_range(void) > */ > memblock_reserve(0, ALIGN(reserve_low, PAGE_SIZE)); > } > - > + > +static void __init early_reserve_memory(void) > +{ > + /* > + * Reserve the memory occupied by the kernel between _text and > + * __end_of_kernel_reserve symbols. Any kernel sections after the > + * __end_of_kernel_reserve symbol must be explicitly reserved with a > + * separate memblock_reserve() or they will be discarded. > + */ > + memblock_reserve(__pa_symbol(_text), > + (unsigned long)__end_of_kernel_reserve - (unsigned long)_text); > + > + /* > + * Make sure page 0 is always reserved because on systems with > + * L1TF its contents can be leaked to user processes. > + */ > + memblock_reserve(0, PAGE_SIZE); > + > + early_reserve_initrd(); > + > + if (efi_enabled(EFI_BOOT)) > + efi_memblock_x86_reserve_range(); > + > + memblock_x86_reserve_range_setup_data(); > + > + reserve_ibft_region(); > + reserve_bios_regions(); > + > + trim_platform_memory_ranges(); > + trim_low_memory_range(); > +} > + > /* > * Dump out kernel offset information on panic. > */ > @@ -763,29 +794,6 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p) > > void __init setup_arch(char **cmdline_p) > { > - /* > - * Reserve the memory occupied by the kernel between _text and > - * __end_of_kernel_reserve symbols. Any kernel sections after the > - * __end_of_kernel_reserve symbol must be explicitly reserved with a > - * separate memblock_reserve() or they will be discarded. > - */ > - memblock_reserve(__pa_symbol(_text), > - (unsigned long)__end_of_kernel_reserve - (unsigned long)_text); > - > - /* > - * Make sure page 0 is always reserved because on systems with > - * L1TF its contents can be leaked to user processes. > - */ > - memblock_reserve(0, PAGE_SIZE); > - > - early_reserve_initrd(); > - > - /* > - * At this point everything still needed from the boot loader > - * or BIOS or kernel text should be early reserved or marked not > - * RAM in e820. All other memory is free game. > - */ > - > #ifdef CONFIG_X86_32 > memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data)); > > @@ -909,8 +917,18 @@ void __init setup_arch(char **cmdline_p) > > parse_early_param(); > > - if (efi_enabled(EFI_BOOT)) > - efi_memblock_x86_reserve_range(); > + /* > + * Do some memory reservations *before* memory is added to > + * memblock, so memblock allocations won't overwrite it. > + * Do it after early param, so we could get (unlikely) panic from > + * serial. > + * > + * After this point everything still needed from the boot loader or > + * firmware or kernel text should be early reserved or marked not > + * RAM in e820. All other memory is free game. > + */ > + early_reserve_memory(); > + > #ifdef CONFIG_MEMORY_HOTPLUG > /* > * Memory used by the kernel cannot be hot-removed because Linux > @@ -937,9 +955,6 @@ void __init setup_arch(char **cmdline_p) > > x86_report_nx(); > > - /* after early param, so could get panic from serial */ > - memblock_x86_reserve_range_setup_data(); > - > if (acpi_mps_check()) { > #ifdef CONFIG_X86_LOCAL_APIC > disable_apic = 1; > @@ -1031,8 +1046,6 @@ void __init setup_arch(char **cmdline_p) > */ > find_smp_config(); > > - reserve_ibft_region(); > - > early_alloc_pgt_buf(); > > /* > @@ -1053,8 +1066,6 @@ void __init setup_arch(char **cmdline_p) > */ > sev_setup_arch(); > > - reserve_bios_regions(); > - > efi_fake_memmap(); > efi_find_mirror(); > efi_esrt_init(); > @@ -1080,9 +1091,6 @@ void __init setup_arch(char **cmdline_p) > > reserve_real_mode(); > > - trim_platform_memory_ranges(); > - trim_low_memory_range(); > - > init_mem_mapping(); > > idt_setup_early_pf(); > Did not fully review for side-effects, but looks like a reasonable thing to do!
On Fri, Jan 15, 2021 at 10:32:54AM +0200, Mike Rapoport wrote: > From: Mike Rapoport <rppt@linux.ibm.com> > > The early reservations of memory areas used by the firmware, bootloader, > kernel text and data are spread over setup_arch(). Moreover, some of them > happen *after* memblock allocations, e.g trim_platform_memory_ranges() and > trim_low_memory_range() are called after reserve_real_mode() that allocates > memory. > > We did not observe corruption of these memory regions because memblock Make that "We" impersonal, passive voice pls. > always allocates memory either from the end of memory (in top-down mode) or > above the kernel image (in bottom-up mode). However, the bottom up mode is > going to be updated to span the entire memory [1] to avoid limitations > caused by KASLR. > > Consolidate early memory reservations in a dedicated function to improve > robustness against future changes. Having the early reservations in one > place also makes it clearer what memory must be reserved before we allow > memblock allocations. Would it make sense to have a check with a WARN or so to catch early reservations which get added after memblock allocations have been allowed? To catch people who don't pay attention...
On Fri, Jan 15, 2021 at 10:32:54AM +0200, Mike Rapoport wrote:
> + trim_low_memory_range();
Btw, you can get rid of that one too:
/*
* Here we put platform-specific memory range workarounds, i.e.
* memory known to be corrupt or otherwise in need to be reserved on
* specific platforms.
*
* If this gets used more widely it could use a real dispatch mechanism.
*/
static void __init trim_platform_memory_ranges(void)
{
trim_snb_memory();
}
yeah, yeah, we can do a real dispatch mechanism but we didn't need one
since 2012 so I guess we can get rid of trim_platform_memory_ranges()
and call trim_snb_memory() directly and simplify it even more.
Thx.
On Mon, Jan 25, 2021 at 03:50:41PM +0100, Borislav Petkov wrote: > On Fri, Jan 15, 2021 at 10:32:54AM +0200, Mike Rapoport wrote: > > From: Mike Rapoport <rppt@linux.ibm.com> > > > > The early reservations of memory areas used by the firmware, bootloader, > > kernel text and data are spread over setup_arch(). Moreover, some of them > > happen *after* memblock allocations, e.g trim_platform_memory_ranges() and > > trim_low_memory_range() are called after reserve_real_mode() that allocates > > memory. > > > > We did not observe corruption of these memory regions because memblock > > Make that "We" impersonal, passive voice pls. Ok. > > always allocates memory either from the end of memory (in top-down mode) or > > above the kernel image (in bottom-up mode). However, the bottom up mode is > > going to be updated to span the entire memory [1] to avoid limitations > > caused by KASLR. > > > > Consolidate early memory reservations in a dedicated function to improve > > robustness against future changes. Having the early reservations in one > > place also makes it clearer what memory must be reserved before we allow > > memblock allocations. > > Would it make sense to have a check with a WARN or so to catch early > reservations which get added after memblock allocations have been > allowed? To catch people who don't pay attention... This would make sense but it's tricky. From memblock perspective, allocations are always allowed and it is the user responsibility to ensure all the early reservations are done before allocating memory. So adding such a WARN would require a new memblock API and it's adoption by all architectures, which is way beyond the scope of this series :)
On Mon, Jan 25, 2021 at 03:59:11PM +0100, Borislav Petkov wrote: > On Fri, Jan 15, 2021 at 10:32:54AM +0200, Mike Rapoport wrote: > > + trim_low_memory_range(); > > Btw, you can get rid of that one too: > > /* > * Here we put platform-specific memory range workarounds, i.e. > * memory known to be corrupt or otherwise in need to be reserved on > * specific platforms. > * > * If this gets used more widely it could use a real dispatch mechanism. > */ > static void __init trim_platform_memory_ranges(void) > { > trim_snb_memory(); > } > > yeah, yeah, we can do a real dispatch mechanism but we didn't need one > since 2012 so I guess we can get rid of trim_platform_memory_ranges() > and call trim_snb_memory() directly and simplify it even more. Ok. > Thx. >
On Mon, Jan 25, 2021 at 05:31:14PM +0200, Mike Rapoport wrote: > This would make sense but it's tricky. From memblock perspective, > allocations are always allowed and it is the user responsibility to ensure > all the early reservations are done before allocating memory. Yah, I don't trust my users to know that for sure... > So adding such a WARN would require a new memblock API and it's adoption by > all architectures, which is way beyond the scope of this series :) So definitely not for those series but I could imagine something like memblock_reserve() if (memblock_allocations_allowed()) WARN or so. This way you don't need to touch the archtectures. It all depends on what the other arches need/use. Or you could even make that a new memblock_reserve_warn() thing or so and wrap that functionality in it and have x86 call it only... Anyway, something to that effect. As to those two patches, you can add Acked-by: Borislav Petkov <bp@suse.de> to the next revision since akpm is going to take them. Thx.
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 3412c4595efd..32cd2e790a0a 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -728,7 +728,38 @@ static void __init trim_low_memory_range(void) */ memblock_reserve(0, ALIGN(reserve_low, PAGE_SIZE)); } - + +static void __init early_reserve_memory(void) +{ + /* + * Reserve the memory occupied by the kernel between _text and + * __end_of_kernel_reserve symbols. Any kernel sections after the + * __end_of_kernel_reserve symbol must be explicitly reserved with a + * separate memblock_reserve() or they will be discarded. + */ + memblock_reserve(__pa_symbol(_text), + (unsigned long)__end_of_kernel_reserve - (unsigned long)_text); + + /* + * Make sure page 0 is always reserved because on systems with + * L1TF its contents can be leaked to user processes. + */ + memblock_reserve(0, PAGE_SIZE); + + early_reserve_initrd(); + + if (efi_enabled(EFI_BOOT)) + efi_memblock_x86_reserve_range(); + + memblock_x86_reserve_range_setup_data(); + + reserve_ibft_region(); + reserve_bios_regions(); + + trim_platform_memory_ranges(); + trim_low_memory_range(); +} + /* * Dump out kernel offset information on panic. */ @@ -763,29 +794,6 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p) void __init setup_arch(char **cmdline_p) { - /* - * Reserve the memory occupied by the kernel between _text and - * __end_of_kernel_reserve symbols. Any kernel sections after the - * __end_of_kernel_reserve symbol must be explicitly reserved with a - * separate memblock_reserve() or they will be discarded. - */ - memblock_reserve(__pa_symbol(_text), - (unsigned long)__end_of_kernel_reserve - (unsigned long)_text); - - /* - * Make sure page 0 is always reserved because on systems with - * L1TF its contents can be leaked to user processes. - */ - memblock_reserve(0, PAGE_SIZE); - - early_reserve_initrd(); - - /* - * At this point everything still needed from the boot loader - * or BIOS or kernel text should be early reserved or marked not - * RAM in e820. All other memory is free game. - */ - #ifdef CONFIG_X86_32 memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data)); @@ -909,8 +917,18 @@ void __init setup_arch(char **cmdline_p) parse_early_param(); - if (efi_enabled(EFI_BOOT)) - efi_memblock_x86_reserve_range(); + /* + * Do some memory reservations *before* memory is added to + * memblock, so memblock allocations won't overwrite it. + * Do it after early param, so we could get (unlikely) panic from + * serial. + * + * After this point everything still needed from the boot loader or + * firmware or kernel text should be early reserved or marked not + * RAM in e820. All other memory is free game. + */ + early_reserve_memory(); + #ifdef CONFIG_MEMORY_HOTPLUG /* * Memory used by the kernel cannot be hot-removed because Linux @@ -937,9 +955,6 @@ void __init setup_arch(char **cmdline_p) x86_report_nx(); - /* after early param, so could get panic from serial */ - memblock_x86_reserve_range_setup_data(); - if (acpi_mps_check()) { #ifdef CONFIG_X86_LOCAL_APIC disable_apic = 1; @@ -1031,8 +1046,6 @@ void __init setup_arch(char **cmdline_p) */ find_smp_config(); - reserve_ibft_region(); - early_alloc_pgt_buf(); /* @@ -1053,8 +1066,6 @@ void __init setup_arch(char **cmdline_p) */ sev_setup_arch(); - reserve_bios_regions(); - efi_fake_memmap(); efi_find_mirror(); efi_esrt_init(); @@ -1080,9 +1091,6 @@ void __init setup_arch(char **cmdline_p) reserve_real_mode(); - trim_platform_memory_ranges(); - trim_low_memory_range(); - init_mem_mapping(); idt_setup_early_pf();