From: Alexey Sheplyakov <asheplyakov@basealt.ru> To: devel-kernel@lists.altlinux.org Cc: "Роман Ставцев" <rst@basealt.ru>, "Игорь Чудов" <nir@basealt.ru>, "Евгений Синельников" <sin@basealt.ru>, "Дмитрий Терёхин" <jqt4@basealt.ru> Subject: [d-kernel] [PATCH 08/32] arm64-stub: fixed secondary cores boot on Baikal-M SoC Date: Wed, 14 Dec 2022 17:18:55 +0400 Message-ID: <20221214131919.681481-8-asheplyakov@basealt.ru> (raw) In-Reply-To: <20221214131919.681481-1-asheplyakov@basealt.ru> Old versions of Baikal-M firmware (ARM-TF) deny execution attempts outside of the (physical) address ranges [0x80000000, 0x8FFFFFFF] and [0xA0000000, 0xBFFFFFFF] Thus PSCI calls to boot secondary cores fail unless the kernel image resides in one of these address ranges. However UEFI PE/COFF loader puts the kernel image into the forbidden range. Since the alignment is good enough EFI stub does not try to relocate the kernel. As a result secondary CPUs fail to boot. Relocation to a random address is not going to work either. Therefore automatically disable kaslr on "known bad" systems (for now only Baikal-M) and forcibly relocate the kernel to a low(er) address. This patch is necessary only for old firmware (pre SDK-M 5.1) and prevents kalsr from working on Baikal-M systems. Signed-off-by: Alexey Sheplyakov <asheplyakov@basealt.ru> X-DONTUPSTREAM X-legacy X-feature-Baikal-M --- drivers/firmware/efi/libstub/arm64-stub.c | 62 ++++++++++++++++++++++- 1 file changed, 61 insertions(+), 1 deletion(-) diff --git a/drivers/firmware/efi/libstub/arm64-stub.c b/drivers/firmware/efi/libstub/arm64-stub.c index f9de5217ea65..665a59a287dc 100644 --- a/drivers/firmware/efi/libstub/arm64-stub.c +++ b/drivers/firmware/efi/libstub/arm64-stub.c @@ -11,6 +11,7 @@ #include <asm/efi.h> #include <asm/memory.h> #include <asm/sections.h> +#include <linux/libfdt.h> #include <asm/sysreg.h> #include "efistub.h" @@ -57,6 +58,32 @@ efi_status_t check_platform_features(void) return EFI_SUCCESS; } +static const char* machines_need_low_alloc[] = { + "baikal,baikal-m", + "baikal,bm1000", +}; + +static bool need_low_alloc(void) { + size_t i; + const void *fdt; + const char *match; + + fdt = get_efi_config_table(DEVICE_TREE_GUID); + if (!fdt) { + efi_info("failed to retrive FDT from EFI\n"); + return false; + } + + for (i = 0; i < ARRAY_SIZE(machines_need_low_alloc); i++) { + match = machines_need_low_alloc[i]; + if (fdt_node_check_compatible(fdt, 0, match) == 0) { + efi_info("machine %s: forcing kernel relocation to low address\n", match); + return true; + } + } + return false; +} + /* * Distro versions of GRUB may ignore the BSS allocation entirely (i.e., fail * to provide space, and fail to zero it). Check for this condition by double @@ -93,6 +120,18 @@ static bool check_image_region(u64 base, u64 size) return ret; } +static inline efi_status_t efi_low_alloc(unsigned long size, unsigned long align, + unsigned long *addr) +{ + /* + * Don't allocate at 0x0. It will confuse code that + * checks pointers against NULL. Skip the first 8 + * bytes so we start at a nice even number. + */ + return efi_low_alloc_above(size, align, addr, 0x8); +} + + efi_status_t handle_kernel_image(unsigned long *image_addr, unsigned long *image_size, unsigned long *reserve_addr, @@ -114,12 +153,21 @@ efi_status_t handle_kernel_image(unsigned long *image_addr, */ u64 min_kimg_align = efi_nokaslr ? MIN_KIMG_ALIGN : EFI_KIMG_ALIGN; + bool force_low_reloc = need_low_alloc(); + if (force_low_reloc) { + if (!efi_nokaslr) { + efi_info("booting on a broken firmware, KASLR will be disabled\n"); + efi_nokaslr = true; + } + } + if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) { efi_guid_t li_fixed_proto = LINUX_EFI_LOADED_IMAGE_FIXED_GUID; void *p; if (efi_nokaslr) { - efi_info("KASLR disabled on kernel command line\n"); + if (!force_low_reloc) + efi_info("KASLR disabled on kernel command line\n"); } else if (efi_bs_call(handle_protocol, image_handle, &li_fixed_proto, &p) == EFI_SUCCESS) { efi_info("Image placement fixed by loader\n"); @@ -161,6 +209,15 @@ efi_status_t handle_kernel_image(unsigned long *image_addr, status = EFI_OUT_OF_RESOURCES; } + if (force_low_reloc) { + status = efi_low_alloc(*reserve_size, + min_kimg_align, + reserve_addr); + if (status != EFI_SUCCESS) { + efi_err("Failed to relocate kernel, expect secondary CPUs boot failure\n"); + } + } + if (status != EFI_SUCCESS) { if (!check_image_region((u64)_text, kernel_memsize)) { efi_err("FIRMWARE BUG: Image BSS overlaps adjacent EFI memory region\n"); @@ -185,6 +242,9 @@ efi_status_t handle_kernel_image(unsigned long *image_addr, } *image_addr = *reserve_addr; + if (efi_nokaslr) { + efi_info("relocating kernel to 0x%lx\n", *image_addr); + } memcpy((void *)*image_addr, _text, kernel_size); return EFI_SUCCESS; -- 2.33.5
next prev parent reply other threads:[~2022-12-14 13:18 UTC|newest] Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-12-14 13:18 [d-kernel] [PATCH 01/32] clk: added Baikal-M clock management unit driver Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 02/32] cpufreq-dt: don't load on Baikal-M SoC Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 03/32] serial: 8250_dw: verify clock rate in dw8250_set_termios Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 04/32] usb: dwc3: of-simple: added compatible string for Baikal-M SoC Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 05/32] dw-pcie: refuse to load on Baikal-M with recent firmware Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 06/32] arm64: Enable armv8 based Baikal-M SoC support Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 07/32] efi-rtc: avoid calling efi.get_time on Baikal-M SoC Alexey Sheplyakov 2022-12-14 13:18 ` Alexey Sheplyakov [this message] 2022-12-14 13:18 ` [d-kernel] [PATCH 09/32] pm: disable all sleep states on Baikal-M based boards Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 10/32] net: fwnode_get_phy_id: consider all compatible strings Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 11/32] net: stmmac: inital support of Baikal-T1/M SoCs GMAC Alexey Sheplyakov 2022-12-14 13:18 ` [d-kernel] [PATCH 12/32] dt-bindings: dwmac: Add bindings for Baikal-T1/M SoCs Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 13/32] net: dwmac-baikal: added compatible strings Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 14/32] Added TF307/TF306 board management controller driver Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 15/32] hwmon: bt1-pvt: access registers via pvt_{readl, writel} helpers Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 16/32] hwmon: bt1-pvt: define pvt_readl/pvt_writel for Baikal-M SoC Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 17/32] hwmon: bt1-pvt: adjusted probing " Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 18/32] hwmon: bt1-pvt: added compatible baikal, pvt Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 19/32] drm: new bridge driver - stdp4028 Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 20/32] drm: added Baikal-M SoC video display unit driver Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 21/32] drm/bridge: dw-hdmi: support ahb audio hw revision 0x2a Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 22/32] dt-bindings: dw-hdmi: added ahb-audio-regshift Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 23/32] drm/bridge: dw-hdmi: force ahb audio register offset for Baikal-M Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 24/32] drm/panfrost: forcibly set dma-coherent on Baikal-M Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 25/32] drm/panfrost: disable devfreq " Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 26/32] ALSA: hda: Baikal-M support Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 27/32] PCI: pcie-baikal: driver for Baikal-M with new firmware Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 28/32] (BROKEN) dwc-i2s: support Baikal-M SoC Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 29/32] input: added TF307 serio PS/2 emulator driver Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 30/32] input: new driver - serdev-serio Alexey Sheplyakov 2022-12-14 13:19 ` [d-kernel] [PATCH 31/32] phy: realtek: leds configuration for RTL8211f Alexey Sheplyakov 2022-12-14 15:06 ` [d-kernel] [PATCH 01/32] clk: added Baikal-M clock management unit driver Vitaly Chikunov 2022-12-16 9:54 ` Alexey Sheplyakov 2022-12-16 12:34 ` Vitaly Chikunov 2022-12-16 12:40 ` Vitaly Chikunov
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20221214131919.681481-8-asheplyakov@basealt.ru \ --to=asheplyakov@basealt.ru \ --cc=devel-kernel@lists.altlinux.org \ --cc=jqt4@basealt.ru \ --cc=nir@basealt.ru \ --cc=rst@basealt.ru \ --cc=sin@basealt.ru \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
ALT Linux kernel packages development This inbox may be cloned and mirrored by anyone: git clone --mirror http://lore.altlinux.org/devel-kernel/0 devel-kernel/git/0.git # If you have public-inbox 1.1+ installed, you may # initialize and index your mirror using the following commands: public-inbox-init -V2 devel-kernel devel-kernel/ http://lore.altlinux.org/devel-kernel \ devel-kernel@altlinux.org devel-kernel@altlinux.ru devel-kernel@altlinux.com public-inbox-index devel-kernel Example config snippet for mirrors. Newsgroup available over NNTP: nntp://lore.altlinux.org/org.altlinux.lists.devel-kernel AGPL code for this site: git clone https://public-inbox.org/public-inbox.git