From f5919b65d4c671fd5083838c7a445f319f9a13c8 Mon Sep 17 00:00:00 2001 From: Stijn Tintel Date: Sat, 10 Nov 2018 13:03:18 +0200 Subject: brcm2708: add kernel 4.14 support Patch generation process: - rebase rpi/rpi-4.14.y on v4.14.89 from linux-stable - git format-patch v4.14.89 Patches skipped during rebase: - lan78xx: Read MAC address from DT if present - lan78xx: Enable LEDs and auto-negotiation - Revert "softirq: Let ksoftirqd do its job" - sc16is7xx: Fix for multi-channel stall - lan78xx: Ignore DT MAC address if already valid - lan78xx: Simple patch to prevent some crashes - tcp_write_queue_purge clears all the SKBs in the write queue - Revert "lan78xx: Simple patch to prevent some crashes" - lan78xx: Connect phy early - Arm: mm: ftrace: Only set text back to ro after kernel has been marked ro - Revert "Revert "softirq: Let ksoftirqd do its job"" - ASoC: cs4265: SOC_SINGLE register value error fix - Revert "ASoC: cs4265: SOC_SINGLE register value error fix" - Revert "net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends" - Revert "Revert "net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends"" Patches dropped after rebase: - net: Add non-mainline source for rtl8192cu wlan - net: Fix rtl8192cu build errors on other platforms - brcm: adds support for BCM43341 wifi - brcmfmac: Mute expected startup 'errors' - ARM64: Fix build break for RTL8187/RTL8192CU wifi - ARM64: Enable RTL8187/RTL8192CU wifi in build config - This is the driver for Sony CXD2880 DVB-T2/T tuner + demodulator - brcmfmac: add CLM download support - brcmfmac: request_firmware_direct is quieter - Sets the BCDC priority to constant 0 - brcmfmac: Disable ARP offloading when promiscuous - brcmfmac: Avoid possible out-of-bounds read - brcmfmac: Delete redundant length check - net: rtl8192cu: Normalize indentation - net: rtl8192cu: Fix implicit fallthrough warnings - Revert "Sets the BCDC priority to constant 0" - media: cxd2880: Bump to match 4.18.y version - media: cxd2880-spi: Bump to match 4.18.y version - Revert "mm: alloc_contig: re-allow CMA to compact FS pages" - Revert "Revert "mm: alloc_contig: re-allow CMA to compact FS pages"" - cxd2880: CXD2880_SPI_DRV should select DVB_CXD2880 with MEDIA_SUBDRV_AUTOSELECT - 950-0421-HID-hid-bigbenff-driver-for-BigBen-Interactive-PS3OF.patch - 950-0453-Add-hid-bigbenff-to-list-of-have_special_driver-for-.patch Make I2C built-in instead of modular as in upstream defconfig; also the easiest way to get MFD_ARIZONA enabled, which is required by kmod-sound-soc-rpi-cirrus. Add missing compatible strings from 4.9/960-add-rasbperrypi-compatible.patch, using upstream names for compute modules. Add extra patch to enable the LEDs on lan78xx. Compile-tested: bcm2708, bcm2709, bcm2710 (with CONFIG_ALL_KMODS=y) Runtime-tested: bcm2708, bcm2710 Signed-off-by: Stijn Tintel --- .../patches-4.14/950-0031-Update-vfpmodule.c.patch | 138 +++++++++++++++++++++ 1 file changed, 138 insertions(+) create mode 100644 target/linux/brcm2708/patches-4.14/950-0031-Update-vfpmodule.c.patch (limited to 'target/linux/brcm2708/patches-4.14/950-0031-Update-vfpmodule.c.patch') diff --git a/target/linux/brcm2708/patches-4.14/950-0031-Update-vfpmodule.c.patch b/target/linux/brcm2708/patches-4.14/950-0031-Update-vfpmodule.c.patch new file mode 100644 index 0000000000..6647e5a4f8 --- /dev/null +++ b/target/linux/brcm2708/patches-4.14/950-0031-Update-vfpmodule.c.patch @@ -0,0 +1,138 @@ +From 5abe6a0ebe150f320a4ca0ed73b2e15fca1b16fd Mon Sep 17 00:00:00 2001 +From: Claggy3 +Date: Sat, 11 Feb 2017 14:00:30 +0000 +Subject: [PATCH 031/454] Update vfpmodule.c + +Christopher Alexander Tobias Schulze - May 2, 2015, 11:57 a.m. +This patch fixes a problem with VFP state save and restore related +to exception handling (panic with message "BUG: unsupported FP +instruction in kernel mode") present on VFP11 floating point units +(as used with ARM1176JZF-S CPUs, e.g. on first generation Raspberry +Pi boards). This patch was developed and discussed on + + https://github.com/raspberrypi/linux/issues/859 + +A precondition to see the crashes is that floating point exception +traps are enabled. In this case, the VFP11 might determine that a FPU +operation needs to trap at a point in time when it is not possible to +signal this to the ARM11 core any more. The VFP11 will then set the +FPEXC.EX bit and store the trapped opcode in FPINST. (In some cases, +a second opcode might have been accepted by the VFP11 before the +exception was detected and could be reported to the ARM11 - in this +case, the VFP11 also sets FPEXC.FP2V and stores the second opcode in +FPINST2.) + +If FPEXC.EX is set, the VFP11 will "bounce" the next FPU opcode issued +by the ARM11 CPU, which will be seen by the ARM11 as an undefined opcode +trap. The VFP support code examines the FPEXC.EX and FPEXC.FP2V bits +to decide what actions to take, i.e., whether to emulate the opcodes +found in FPINST and FPINST2, and whether to retry the bounced instruction. + +If a user space application has left the VFP11 in this "pending trap" +state, the next FPU opcode issued to the VFP11 might actually be the +VSTMIA operation vfp_save_state() uses to store the FPU registers +to memory (in our test cases, when building the signal stack frame). +In this case, the kernel crashes as described above. + +This patch fixes the problem by making sure that vfp_save_state() is +always entered with FPEXC.EX cleared. (The current value of FPEXC has +already been saved, so this does not corrupt the context. Clearing +FPEXC.EX has no effects on FPINST or FPINST2. Also note that many +callers already modify FPEXC by setting FPEXC.EN before invoking +vfp_save_state().) + +This patch also addresses a second problem related to FPEXC.EX: After +returning from signal handling, the kernel reloads the VFP context +from the user mode stack. However, the current code explicitly clears +both FPEXC.EX and FPEXC.FP2V during reload. As VFP11 requires these +bits to be preserved, this patch disables clearing them for VFP +implementations belonging to architecture 1. There should be no +negative side effects: the user can set both bits by executing FPU +opcodes anyway, and while user code may now place arbitrary values +into FPINST and FPINST2 (e.g., non-VFP ARM opcodes) the VFP support +code knows which instructions can be emulated, and rejects other +opcodes with "unhandled bounce" messages, so there should be no +security impact from allowing reloading FPEXC.EX and FPEXC.FP2V. + +Signed-off-by: Christopher Alexander Tobias Schulze +--- + arch/arm/vfp/vfpmodule.c | 26 ++++++++++++++++++++------ + 1 file changed, 20 insertions(+), 6 deletions(-) + +--- a/arch/arm/vfp/vfpmodule.c ++++ b/arch/arm/vfp/vfpmodule.c +@@ -179,8 +179,11 @@ static int vfp_notifier(struct notifier_ + * case the thread migrates to a different CPU. The + * restoring is done lazily. + */ +- if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) ++ if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) { ++ /* vfp_save_state oopses on VFP11 if EX bit set */ ++ fmxr(FPEXC, fpexc & ~FPEXC_EX); + vfp_save_state(vfp_current_hw_state[cpu], fpexc); ++ } + #endif + + /* +@@ -463,13 +466,16 @@ static int vfp_pm_suspend(void) + /* if vfp is on, then save state for resumption */ + if (fpexc & FPEXC_EN) { + pr_debug("%s: saving vfp state\n", __func__); ++ /* vfp_save_state oopses on VFP11 if EX bit set */ ++ fmxr(FPEXC, fpexc & ~FPEXC_EX); + vfp_save_state(&ti->vfpstate, fpexc); + + /* disable, just in case */ + fmxr(FPEXC, fmrx(FPEXC) & ~FPEXC_EN); + } else if (vfp_current_hw_state[ti->cpu]) { + #ifndef CONFIG_SMP +- fmxr(FPEXC, fpexc | FPEXC_EN); ++ /* vfp_save_state oopses on VFP11 if EX bit set */ ++ fmxr(FPEXC, (fpexc & ~FPEXC_EX) | FPEXC_EN); + vfp_save_state(vfp_current_hw_state[ti->cpu], fpexc); + fmxr(FPEXC, fpexc); + #endif +@@ -532,7 +538,8 @@ void vfp_sync_hwstate(struct thread_info + /* + * Save the last VFP state on this CPU. + */ +- fmxr(FPEXC, fpexc | FPEXC_EN); ++ /* vfp_save_state oopses on VFP11 if EX bit set */ ++ fmxr(FPEXC, (fpexc & ~FPEXC_EX) | FPEXC_EN); + vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN); + fmxr(FPEXC, fpexc); + } +@@ -602,6 +609,8 @@ int vfp_restore_user_hwstate(struct user + struct thread_info *thread = current_thread_info(); + struct vfp_hard_struct *hwstate = &thread->vfpstate.hard; + unsigned long fpexc; ++ int err = 0; ++ u32 fpsid = fmrx(FPSID); + + /* Disable VFP to avoid corrupting the new thread state. */ + vfp_flush_hwstate(thread); +@@ -624,8 +633,12 @@ int vfp_restore_user_hwstate(struct user + /* Ensure the VFP is enabled. */ + fpexc |= FPEXC_EN; + +- /* Ensure FPINST2 is invalid and the exception flag is cleared. */ +- fpexc &= ~(FPEXC_EX | FPEXC_FP2V); ++ /* Mask FPXEC_EX and FPEXC_FP2V if not required by VFP arch */ ++ if ((fpsid & FPSID_ARCH_MASK) != (1 << FPSID_ARCH_BIT)) { ++ /* Ensure FPINST2 is invalid and the exception flag is cleared. */ ++ fpexc &= ~(FPEXC_EX | FPEXC_FP2V); ++ } ++ + hwstate->fpexc = fpexc; + + hwstate->fpinst = ufp_exc->fpinst; +@@ -695,7 +708,8 @@ void kernel_neon_begin(void) + cpu = get_cpu(); + + fpexc = fmrx(FPEXC) | FPEXC_EN; +- fmxr(FPEXC, fpexc); ++ /* vfp_save_state oopses on VFP11 if EX bit set */ ++ fmxr(FPEXC, fpexc & ~FPEXC_EX); + + /* + * Save the userland NEON/VFP state. Under UP, -- cgit v1.2.3