/*      $NetBSD: linux_machdep.c,v 1.62 2024/10/01 16:35:42 riastradh Exp $ */

/*-
* Copyright (c) 2005 Emmanuel Dreyfus, all rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
*    must display the following acknowledgement:
*      This product includes software developed by Emmanuel Dreyfus
* 4. The name of the author may not be used to endorse or promote
*    products derived from this software without specific prior written
*    permission.
*
* THIS SOFTWARE IS PROVIDED BY THE THE AUTHOR AND CONTRIBUTORS ``AS IS''
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
* THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/

#include <sys/cdefs.h>

__KERNEL_RCSID(0, "$NetBSD: linux_machdep.c,v 1.62 2024/10/01 16:35:42 riastradh Exp $");

#include <sys/param.h>
#include <sys/types.h>
#include <sys/systm.h>
#include <sys/signal.h>
#include <sys/exec.h>
#include <sys/proc.h>
#include <sys/ptrace.h> /* for process_read_fpregs() */
#include <sys/ucontext.h>
#include <sys/conf.h>

#include <machine/reg.h>
#include <machine/pcb.h>
#include <machine/mcontext.h>
#include <machine/specialreg.h>
#include <machine/vmparam.h>
#include <machine/cpufunc.h>
#include <x86/include/sysarch.h>

/*
* To see whether wscons is configured (for virtual console ioctl calls).
*/
#if defined(_KERNEL_OPT)
#include "opt_user_ldt.h"
#include "wsdisplay.h"
#endif

#if (NWSDISPLAY > 0)
#include <dev/wscons/wsconsio.h>
#include <dev/wscons/wsdisplay_usl_io.h>
#endif


#include <compat/linux/common/linux_signal.h>
#include <compat/linux/common/linux_errno.h>
#include <compat/linux/common/linux_exec.h>
#include <compat/linux/common/linux_ioctl.h>
#include <compat/linux/common/linux_prctl.h>
#include <compat/linux/common/linux_machdep.h>
#include <compat/linux/common/linux_ipc.h>
#include <compat/linux/common/linux_sem.h>
#include <compat/linux/linux_syscall.h>
#include <compat/linux/linux_syscallargs.h>

static void linux_buildcontext(struct lwp *, void *, void *);

void
linux_setregs(struct lwp *l, struct exec_package *epp, vaddr_t stack)
{
       struct pcb *pcb = lwp_getpcb(l);
       struct trapframe *tf;

#ifdef USER_LDT
       pmap_ldt_cleanup(l);
#endif

       fpu_clear(l, __NetBSD_NPXCW__);

       kpreempt_disable();
       pcb->pcb_flags = 0;
       l->l_proc->p_flag &= ~PK_32;
       l->l_md.md_flags = MDL_IRET;
       cpu_segregs64_zero(l);
       kpreempt_enable();

       tf = l->l_md.md_regs;
       tf->tf_rax = 0;
       tf->tf_rbx = 0;
       tf->tf_rcx = epp->ep_entry;
       tf->tf_rdx = 0;
       tf->tf_rsi = 0;
       tf->tf_rdi = 0;
       tf->tf_rbp = 0;
       tf->tf_rsp = stack;
       tf->tf_r8 = 0;
       tf->tf_r9 = 0;
       tf->tf_r10 = 0;
       tf->tf_r11 = 0;
       tf->tf_r12 = 0;
       tf->tf_r13 = 0;
       tf->tf_r14 = 0;
       tf->tf_r15 = 0;
       tf->tf_rip = epp->ep_entry;
       tf->tf_rflags = PSL_USERSET;
       tf->tf_cs = GSEL(GUCODE_SEL, SEL_UPL);
       tf->tf_ss = GSEL(GUDATA_SEL, SEL_UPL);
       tf->tf_ds = GSEL(GUDATA_SEL, SEL_UPL);
       tf->tf_es = 0;

       return;
}

void
linux_sendsig(const ksiginfo_t *ksi, const sigset_t *mask)
{
       struct lwp *l = curlwp;
       struct proc *p = l->l_proc;
       struct pcb *pcb = lwp_getpcb(l);
       struct sigacts *ps = p->p_sigacts;
       int onstack, error;
       int sig = ksi->ksi_signo;
       struct linux_rt_sigframe *sfp, sigframe;
       struct linux__fpstate *fpsp;
       struct fpreg fpregs;
       struct trapframe *tf = l->l_md.md_regs;
       sig_t catcher = SIGACTION(p, sig).sa_handler;
       linux_sigset_t lmask;
       char *sp;

       /* Do we need to jump onto the signal stack? */
       onstack =
           (l->l_sigstk.ss_flags & (SS_DISABLE | SS_ONSTACK)) == 0 &&
           (SIGACTION(p, sig).sa_flags & SA_ONSTACK) != 0;

       /* Allocate space for the signal handler context. */
       if (onstack)
               sp = ((char *)l->l_sigstk.ss_sp +
                   l->l_sigstk.ss_size);
       else
               sp = (char *)tf->tf_rsp - 128;

       /* Save FPU state */
       sp = (char *) (((long)sp - sizeof (*fpsp)) & ~0xfUL);
       fpsp = (struct linux__fpstate *)sp;

       /*
        * Populate the rt_sigframe
        */
       sp = (char *)
           ((((long)sp - sizeof(struct linux_rt_sigframe)) & ~0xfUL) - 8);
       sfp = (struct linux_rt_sigframe *)sp;

       memset(&sigframe, 0, sizeof(sigframe));
       if (ps->sa_sigdesc[sig].sd_vers != __SIGTRAMP_SIGCODE_VERSION)
               sigframe.pretcode =
                   (char *)(u_long)ps->sa_sigdesc[sig].sd_tramp;
       else
               sigframe.pretcode = NULL;

       /*
        * The user context
        */
       sigframe.uc.luc_flags = 0;
       sigframe.uc.luc_link = NULL;

       /* This is used regardless of SA_ONSTACK in Linux */
       sigframe.uc.luc_stack.ss_sp = l->l_sigstk.ss_sp;
       sigframe.uc.luc_stack.ss_size = l->l_sigstk.ss_size;
       sigframe.uc.luc_stack.ss_flags = 0;
       if (l->l_sigstk.ss_flags & SS_ONSTACK)
               sigframe.uc.luc_stack.ss_flags |= LINUX_SS_ONSTACK;
       if (l->l_sigstk.ss_flags & SS_DISABLE)
               sigframe.uc.luc_stack.ss_flags |= LINUX_SS_DISABLE;

       sigframe.uc.luc_mcontext.r8 = tf->tf_r8;
       sigframe.uc.luc_mcontext.r9 = tf->tf_r9;
       sigframe.uc.luc_mcontext.r10 = tf->tf_r10;
       sigframe.uc.luc_mcontext.r11 = tf->tf_r11;
       sigframe.uc.luc_mcontext.r12 = tf->tf_r12;
       sigframe.uc.luc_mcontext.r13 = tf->tf_r13;
       sigframe.uc.luc_mcontext.r14 = tf->tf_r14;
       sigframe.uc.luc_mcontext.r15 = tf->tf_r15;
       sigframe.uc.luc_mcontext.rdi = tf->tf_rdi;
       sigframe.uc.luc_mcontext.rsi = tf->tf_rsi;
       sigframe.uc.luc_mcontext.rbp = tf->tf_rbp;
       sigframe.uc.luc_mcontext.rbx = tf->tf_rbx;
       sigframe.uc.luc_mcontext.rdx = tf->tf_rdx;
       sigframe.uc.luc_mcontext.rax = tf->tf_rax;
       sigframe.uc.luc_mcontext.rcx = tf->tf_rcx;
       sigframe.uc.luc_mcontext.rsp = tf->tf_rsp;
       sigframe.uc.luc_mcontext.rip = tf->tf_rip;
       sigframe.uc.luc_mcontext.eflags = tf->tf_rflags;
       sigframe.uc.luc_mcontext.cs = GSEL(GUCODE_SEL, SEL_UPL);
       sigframe.uc.luc_mcontext.gs = tf->tf_gs & 0xFFFF;
       sigframe.uc.luc_mcontext.fs = tf->tf_fs & 0xFFFF;
       sigframe.uc.luc_mcontext.err = tf->tf_err;
       sigframe.uc.luc_mcontext.trapno = tf->tf_trapno;
       native_to_linux_sigset(&lmask, mask);
       sigframe.uc.luc_mcontext.oldmask = lmask.sig[0];
       sigframe.uc.luc_mcontext.cr2 = (long)pcb->pcb_onfault;
       sigframe.uc.luc_mcontext.fpstate = fpsp;
       native_to_linux_sigset(&sigframe.uc.luc_sigmask, mask);
       native_to_linux_siginfo(&sigframe.info, &ksi->ksi_info);
       sendsig_reset(l, sig);
       mutex_exit(p->p_lock);
       error = 0;

       /*
        * Save FPU state, if any
        */
       if (fpsp != NULL) {
               size_t fp_size = sizeof fpregs;
               /* The netbsd and linux structures both match the fxsave data */
               memset(&fpregs, 0, sizeof(fpregs));
               (void)process_read_fpregs(l, &fpregs, &fp_size);
               error = copyout(&fpregs, fpsp, sizeof(*fpsp));
       }

       if (error == 0)
               error = copyout(&sigframe, sp, sizeof(sigframe));

       mutex_enter(p->p_lock);

       if (error != 0) {
               sigexit(l, SIGILL);
               return;
       }

       if ((vaddr_t)catcher >= VM_MAXUSER_ADDRESS) {
               sigexit(l, SIGILL);
               return;
       }

       linux_buildcontext(l, catcher, sp);
       tf->tf_rdi = sigframe.info.lsi_signo;
       tf->tf_rax = 0;
       tf->tf_rsi = (long)&sfp->info;
       tf->tf_rdx = (long)&sfp->uc;

       /*
        * Remember we use signal stack
        */
       if (onstack)
               l->l_sigstk.ss_flags |= SS_ONSTACK;
       return;
}

int
linux_sys_modify_ldt(struct lwp *l, const struct linux_sys_modify_ldt_args *v, register_t *retval)
{
       return 0;
}

int
linux_sys_iopl(struct lwp *l, const struct linux_sys_iopl_args *v, register_t *retval)
{
       return 0;
}

int
linux_sys_ioperm(struct lwp *l, const struct linux_sys_ioperm_args *v, register_t *retval)
{
       return 0;
}

dev_t
linux_fakedev(dev_t dev, int raw)
{

      extern const struct cdevsw ptc_cdevsw, pts_cdevsw;
      const struct cdevsw *cd = cdevsw_lookup(dev);

      if (raw) {
#if (NWSDISPLAY > 0)
              extern const struct cdevsw wsdisplay_cdevsw;
              if (cd == &wsdisplay_cdevsw)
                      return makedev(LINUX_CONS_MAJOR, (minor(dev) + 1));
#endif
      }

      if (cd == &ptc_cdevsw)
              return makedev(LINUX_PTC_MAJOR, minor(dev));
      if (cd == &pts_cdevsw)
              return makedev(LINUX_PTS_MAJOR, minor(dev));

       return ((minor(dev) & 0xff) | ((major(dev) & 0xfff) << 8)
           | (((unsigned long long int) (minor(dev) & ~0xff)) << 12)
           | (((unsigned long long int) (major(dev) & ~0xfff)) << 32));
}

int
linux_machdepioctl(struct lwp *l, const struct linux_sys_ioctl_args *v, register_t *retval)
{
       return 0;
}

int
linux_sys_rt_sigreturn(struct lwp *l, const void *v, register_t *retval)
{
       struct linux_ucontext *luctx;
       struct trapframe *tf = l->l_md.md_regs;
       struct linux_sigcontext *lsigctx;
       struct linux_rt_sigframe frame, *fp;
       ucontext_t uctx;
       mcontext_t *mctx;
       struct fxsave *fxarea;
       int error;

       fp = (struct linux_rt_sigframe *)(tf->tf_rsp - 8);
       if ((error = copyin(fp, &frame, sizeof(frame))) != 0) {
               mutex_enter(l->l_proc->p_lock);
               sigexit(l, SIGILL);
               return error;
       }
       luctx = &frame.uc;
       lsigctx = &luctx->luc_mcontext;

       memset(&uctx, 0, sizeof(uctx));
       mctx = (mcontext_t *)&uctx.uc_mcontext;
       fxarea = (struct fxsave *)&mctx->__fpregs;

       /*
        * Set the flags. Linux always have CPU, stack and signal state,
        * FPU is optional. uc_flags is not used to tell what we have.
        */
       uctx.uc_flags = (_UC_SIGMASK|_UC_CPU|_UC_STACK|_UC_CLRSTACK);
       if (lsigctx->fpstate != NULL)
               uctx.uc_flags |= _UC_FPU;
       uctx.uc_link = NULL;

       /*
        * Signal set
        */
       linux_to_native_sigset(&uctx.uc_sigmask, &luctx->luc_sigmask);

       /*
        * CPU state
        */
       mctx->__gregs[_REG_R8] = lsigctx->r8;
       mctx->__gregs[_REG_R9] = lsigctx->r9;
       mctx->__gregs[_REG_R10] = lsigctx->r10;
       mctx->__gregs[_REG_R11] = lsigctx->r11;
       mctx->__gregs[_REG_R12] = lsigctx->r12;
       mctx->__gregs[_REG_R13] = lsigctx->r13;
       mctx->__gregs[_REG_R14] = lsigctx->r14;
       mctx->__gregs[_REG_R15] = lsigctx->r15;
       mctx->__gregs[_REG_RDI] = lsigctx->rdi;
       mctx->__gregs[_REG_RSI] = lsigctx->rsi;
       mctx->__gregs[_REG_RBP] = lsigctx->rbp;
       mctx->__gregs[_REG_RBX] = lsigctx->rbx;
       mctx->__gregs[_REG_RAX] = lsigctx->rax;
       mctx->__gregs[_REG_RDX] = lsigctx->rdx;
       mctx->__gregs[_REG_RCX] = lsigctx->rcx;
       mctx->__gregs[_REG_RIP] = lsigctx->rip;
       mctx->__gregs[_REG_RFLAGS] = lsigctx->eflags;
       mctx->__gregs[_REG_CS] = lsigctx->cs & 0xFFFF;
       mctx->__gregs[_REG_GS] = lsigctx->gs & 0xFFFF;
       mctx->__gregs[_REG_FS] = lsigctx->fs & 0xFFFF;
       mctx->__gregs[_REG_ERR] = lsigctx->err;
       mctx->__gregs[_REG_TRAPNO] = lsigctx->trapno;
       mctx->__gregs[_REG_ES] = tf->tf_es & 0xFFFF;
       mctx->__gregs[_REG_DS] = tf->tf_ds & 0xFFFF;
       mctx->__gregs[_REG_RSP] = lsigctx->rsp; /* XXX */
       mctx->__gregs[_REG_SS] = tf->tf_ss & 0xFFFF;

       /*
        * FPU state
        */
       if (lsigctx->fpstate != NULL) {
               /* Both structures match the fxstate data */
               error = copyin(lsigctx->fpstate, fxarea, sizeof(*fxarea));
               if (error != 0) {
                       mutex_enter(l->l_proc->p_lock);
                       sigexit(l, SIGILL);
                       return error;
               }
       }

       /*
        * And the stack
        */
       uctx.uc_stack.ss_flags = 0;
       if (luctx->luc_stack.ss_flags & LINUX_SS_ONSTACK)
               uctx.uc_stack.ss_flags |= SS_ONSTACK;

       if (luctx->luc_stack.ss_flags & LINUX_SS_DISABLE)
               uctx.uc_stack.ss_flags |= SS_DISABLE;

       uctx.uc_stack.ss_sp = luctx->luc_stack.ss_sp;
       uctx.uc_stack.ss_size = luctx->luc_stack.ss_size;

       /*
        * And let setucontext deal with that.
        */
       mutex_enter(l->l_proc->p_lock);
       error = setucontext(l, &uctx);
       mutex_exit(l->l_proc->p_lock);
       if (error)
               return error;

       return EJUSTRETURN;
}

int
linux_sys_arch_prctl(struct lwp *l,
   const struct linux_sys_arch_prctl_args *uap, register_t *retval)
{
       /* {
               syscallarg(int) code;
               syscallarg(unsigned long) addr;
       } */
       void *addr = (void *)SCARG(uap, addr);

       switch(SCARG(uap, code)) {
       case LINUX_ARCH_SET_GS:
               return x86_set_sdbase(addr, 'g', l, true);

       case LINUX_ARCH_GET_GS:
               return x86_get_sdbase(addr, 'g');

       case LINUX_ARCH_SET_FS:
               return x86_set_sdbase(addr, 'f', l, true);

       case LINUX_ARCH_GET_FS:
               return x86_get_sdbase(addr, 'f');

       default:
#ifdef DEBUG_LINUX
               printf("linux_sys_arch_prctl: unexpected code %d\n",
                   SCARG(uap, code));
#endif
               return EINVAL;
       }
       /* NOTREACHED */
}

const int linux_vsyscall_to_syscall[] = {
       LINUX_SYS_gettimeofday,
       LINUX_SYS_time,
       LINUX_SYS_nosys,        /* nosys */
       LINUX_SYS_nosys,        /* nosys */
};

int
linux_usertrap(struct lwp *l, vaddr_t trapaddr, void *arg)
{
       struct trapframe *tf = arg;
       uint64_t retaddr;
       size_t vsyscallnr;

       /*
        * Check for a vsyscall. %rip must be the fault address,
        * and the address must be in the Linux vsyscall area.
        * Also, vsyscalls are only done at 1024-byte boundaries.
        */

       if (__predict_true(trapaddr < LINUX_VSYSCALL_START))
               return 0;

       if (trapaddr != tf->tf_rip)
               return 0;

       if ((tf->tf_rip & (LINUX_VSYSCALL_SIZE - 1)) != 0)
               return 0;

       vsyscallnr = (tf->tf_rip - LINUX_VSYSCALL_START) / LINUX_VSYSCALL_SIZE;

       if (vsyscallnr > LINUX_VSYSCALL_MAXNR)
               return 0;

       /*
        * Get the return address from the top of the stack,
        * and fix up the return address.
        * This assumes the faulting instruction was callq *reg,
        * which is the only way that vsyscalls are ever entered.
        */
       if (copyin((void *)tf->tf_rsp, &retaddr, sizeof retaddr) != 0)
               return 0;
       if ((vaddr_t)retaddr >= VM_MAXUSER_ADDRESS)
               return 0;
       tf->tf_rip = retaddr;
       tf->tf_rax = linux_vsyscall_to_syscall[vsyscallnr];
       tf->tf_rsp += 8;        /* "pop" the return address */

#if 0
       printf("usertrap: rip %p rsp %p retaddr %p vsys %d sys %d\n",
           (void *)tf->tf_rip, (void *)tf->tf_rsp, (void *)retaddr,
           vsyscallnr, (int)tf->tf_rax);
#endif

       (*l->l_proc->p_md.md_syscall)(tf);

       return 1;
}

static void
linux_buildcontext(struct lwp *l, void *catcher, void *f)
{
       struct trapframe *tf = l->l_md.md_regs;

       tf->tf_ds = GSEL(GUDATA_SEL, SEL_UPL);
       tf->tf_rip = (u_int64_t)catcher;
       tf->tf_cs = GSEL(GUCODE_SEL, SEL_UPL);
       tf->tf_rflags &= ~PSL_CLEARSIG;
       tf->tf_rsp = (u_int64_t)f;
       tf->tf_ss = GSEL(GUDATA_SEL, SEL_UPL);
}