Fixed env_segment_free bug

[akaros.git] / kern / src / env.c

/* See COPYRIGHT for copyright information. */

#ifdef __SHARC__
#pragma nosharc
#endif

#include <arch/arch.h>
#include <arch/mmu.h>
#include <arch/bitmask.h>
#include <elf.h>
#include <smp.h>

#include <atomic.h>
#include <string.h>
#include <assert.h>
#include <process.h>
#include <pmap.h>
#include <trap.h>
#include <monitor.h>
#include <manager.h>
#include <stdio.h>
#include <schedule.h>
#include <kmalloc.h>

#include <ros/syscall.h>
#include <ros/error.h>

env_t *envs = NULL;             // All environments
atomic_t num_envs;
// TODO: make this a struct of info including the pointer and cacheline-align it
// This lets the kernel know what process is running on the core it traps into.
// A lot of the Env business, including this and its usage, will change when we
// redesign the env as a multi-process.
env_t* (RO curenvs)[MAX_NUM_CPUS] = {[0 ... (MAX_NUM_CPUS-1)] NULL};

#define ENVGENSHIFT     12              // >= LOGNENV

//
// Converts an envid to an env pointer.
//
// RETURNS
//   0 on success, -EBADENV on error.
//   On success, sets *env_store to the environment.
//   On error, sets *env_store to NULL.
//
int
envid2env(envid_t envid, env_t **env_store, bool checkperm)
{
        env_t *e;

        // If envid is zero, return the current environment.
        if (envid == 0) {
                *env_store = current;
                return 0;
        }

        // Look up the Env structure via the index part of the envid,
        // then check the env_id field in that env_t
        // to ensure that the envid is not stale
        // (i.e., does not refer to a _previous_ environment
        // that used the same slot in the envs[] array).
        e = &envs[ENVX(envid)];
        if (e->state == ENV_FREE || e->env_id != envid) {
                *env_store = 0;
                return -EBADENV;
        }

        // Check that the calling environment has legitimate permission
        // to manipulate the specified environment.
        // If checkperm is set, the specified environment
        // must be either the current environment
        // or an immediate child of the current environment.
        // TODO: should check for current being null
        if (checkperm && e != current && e->env_parent_id != current->env_id) {
                *env_store = 0;
                return -EBADENV;
        }

        *env_store = e;
        return 0;
}

//
// Mark all environments in 'envs' as free, set their env_ids to 0,
// and insert them into the proc_freelist.
// Insert in reverse order, so that the first call to env_alloc()
// returns envs[0].
// TODO: get rid of this whole array bullshit
//
void
env_init(void)
{
        int i;

        schedule_init();
        // core 0 is not idle, all others are (for now)
        spin_lock(&idle_lock);
        num_idlecores = num_cpus - 1;
        for (i = 0; i < num_idlecores; i++)
                idlecoremap[i] = i + 1;
        spin_unlock(&idle_lock);
        atomic_init(&num_envs, 0);
        TAILQ_INIT(&proc_freelist);
        assert(envs != NULL);
        for (i = NENV-1; i >= 0; i--) {
                // these should already be set from when i memset'd the array to 0
                envs[i].state = ENV_FREE;
                envs[i].end_text_segment = (void*)UTEXT;
                envs[i].end_data_segment = (void*)UTEXT;
                envs[i].env_id = 0;
                TAILQ_INSERT_HEAD(&proc_freelist, &envs[i], proc_link);
        }
}

//
// Initialize the kernel virtual memory layout for environment e.
// Allocate a page directory, set e->env_pgdir and e->env_cr3 accordingly,
// and initialize the kernel portion of the new environment's address space.
// Do NOT (yet) map anything into the user portion
// of the environment's virtual address space.
//
// Returns 0 on success, < 0 on error.  Errors include:
//      -ENOMEM if page directory or table could not be allocated.
//
static int
env_setup_vm(env_t *e)
WRITES(e->env_pgdir, e->env_cr3, e->env_procinfo, e->env_procdata)
{
        int i, r;
        page_t *pgdir = NULL;
        page_t *pginfo[PROCINFO_NUM_PAGES] = {NULL};
        page_t *pgdata[PROCDATA_NUM_PAGES] = {NULL};
        static page_t * RO shared_page = 0;

        /*
         * First, allocate a page for the pgdir of this process and up
         * its reference count since this will never be done elsewhere
         */
        r = page_alloc(&pgdir);
        if(r < 0) return r;
        page_incref(pgdir);

        /*
         * Next, set up the e->env_pgdir and e->env_cr3 pointers to point
         * to this newly allocated page and clear its contents
         */
        memset(page2kva(pgdir), 0, PGSIZE);
        e->env_pgdir = (pde_t *COUNT(NPDENTRIES)) TC(page2kva(pgdir));
        e->env_cr3 =   (physaddr_t) TC(page2pa(pgdir));

        /*
         * Now start filling in the pgdir with mappings required by all newly
         * created address spaces
         */

        // Map in the kernel to the top of every address space
        // should be able to do this so long as boot_pgdir never has
        // anything put below UTOP
        // TODO check on this!  had a nasty bug because of it
        // this is a bit wonky, since if it's not PGSIZE, lots of other things are
        // screwed up...
        memcpy(e->env_pgdir, boot_pgdir, NPDENTRIES*sizeof(pde_t));

        // VPT and UVPT map the env's own page table, with
        // different permissions.
        e->env_pgdir[PDX(VPT)]  = PTE(LA2PPN(e->env_cr3), PTE_P | PTE_KERN_RW);
        e->env_pgdir[PDX(UVPT)] = PTE(LA2PPN(e->env_cr3), PTE_P | PTE_USER_RO);

        /*
         * Now allocate and insert all pages required for the shared
         * procinfo structure into the page table
         */
        for(int i=0; i<PROCINFO_NUM_PAGES; i++) {
                if(page_alloc(&pginfo[i]) < 0)
                        goto env_setup_vm_error;
                if(page_insert(e->env_pgdir, pginfo[i], (void*SNT)(UINFO + i*PGSIZE),
                               PTE_USER_RO) < 0)
                        goto env_setup_vm_error;
        }

        /*
         * Now allocate and insert all pages required for the shared
         * procdata structure into the page table
         */
        for(int i=0; i<PROCDATA_NUM_PAGES; i++) {
                if(page_alloc(&pgdata[i]) < 0)
                        goto env_setup_vm_error;
                if(page_insert(e->env_pgdir, pgdata[i], (void*SNT)(UDATA + i*PGSIZE),
                               PTE_USER_RW) < 0)
                        goto env_setup_vm_error;
        }

        /*
         * Now, set e->env_procinfo, and e->env_procdata to point to
         * the proper pages just allocated and clear them out.
         */
        e->env_procinfo = (procinfo_t *SAFE) TC(page2kva(pginfo[0]));
        e->env_procdata = (procdata_t *SAFE) TC(page2kva(pgdata[0]));

        memset(e->env_procinfo, 0, sizeof(procinfo_t));
        memset(e->env_procdata, 0, sizeof(procdata_t));

        /* Finally, set up the Global Shared Data page for all processes.
         * Can't be trusted, but still very useful at this stage for us.
         * Consider removing when we have real processes.
         * (TODO).  Note the page is alloced only the first time through
         */
        if (!shared_page) {
                if(page_alloc(&shared_page) < 0)
                        goto env_setup_vm_error;
                // Up it, so it never goes away.  One per user, plus one from page_alloc
                // This is necessary, since it's in the per-process range of memory that
                // gets freed during page_free.
                page_incref(shared_page);
        }

        // Inserted into every process's address space at UGDATA
        if(page_insert(e->env_pgdir, shared_page, (void*SNT)UGDATA, PTE_USER_RW) < 0)
                goto env_setup_vm_error;

        return 0;

env_setup_vm_error:
        page_free(shared_page);
        for(int i=0; i< PROCDATA_NUM_PAGES; i++) {
                page_free(pgdata[i]);
        }
        for(int i=0; i< PROCINFO_NUM_PAGES; i++) {
                page_free(pginfo[i]);
        }
        env_user_mem_free(e);
        page_free(pgdir);
        return -ENOMEM;
}

static void
proc_init_procinfo(struct proc* p)
{
        p->env_procinfo->id = (p->env_id & 0x3FF);

        // TODO: maybe do something smarter here
        p->env_procinfo->max_harts = num_cpus-1;
}

//
// Allocates and initializes a new environment.
// On success, the new environment is stored in *newenv_store.
//
// Returns 0 on success, < 0 on failure.  Errors include:
//      -ENOFREEENV if all NENVS environments are allocated
//      -ENOMEM on memory exhaustion
//
int
env_alloc(env_t **newenv_store, envid_t parent_id)
{
        int32_t generation;
        int r;
        env_t *e;

        spin_lock(&freelist_lock);
        e = TAILQ_FIRST(&proc_freelist);
        if (e) {
                TAILQ_REMOVE(&proc_freelist, e, proc_link);
                spin_unlock(&freelist_lock);
        } else {
                spin_unlock(&freelist_lock);
                return -ENOFREEENV;
        }

    { INITSTRUCT(*e)

        // Allocate and set up the page directory for this environment.
        if ((r = env_setup_vm(e)) < 0) {
                spin_lock(&freelist_lock);
                TAILQ_INSERT_HEAD(&proc_freelist, e, proc_link);
                spin_unlock(&freelist_lock);
                return r;
        }

        // Generate an env_id for this environment.
        generation = (e->env_id + (1 << ENVGENSHIFT)) & ~(NENV - 1);
        if (generation <= 0)    // Don't create a negative env_id.
                generation = 1 << ENVGENSHIFT;
        e->env_id = generation | (e - envs);

        // Set the basic status variables.
        e->proc_lock = 0;
        e->env_parent_id = parent_id;
        proc_set_state(e, PROC_CREATED);
        e->env_runs = 0;
        e->env_refcnt = 1;
        e->env_flags = 0;
        e->env_entry = 0; // cheating.  this really gets set in load_icode
        e->num_vcores = 0;
        for (int i = 0; i < MAX_NUM_CPUS; i++)
                e->vcoremap[i] = -1;
        e->cache_colors_map = kmalloc(llc_cache->num_colors, 0);
        CLR_BITMASK(e->cache_colors_map, llc_cache->num_colors);
        memset(&e->resources, 0, sizeof(e->resources));

        memset(&e->env_ancillary_state, 0, sizeof(e->env_ancillary_state));
        memset(&e->env_tf, 0, sizeof(e->env_tf));
        proc_init_trapframe(&e->env_tf);

        proc_init_procinfo(e);

        /*
         * Initialize the contents of the e->env_procdata structure
         */
        // Initialize the generic syscall ring buffer
        SHARED_RING_INIT(&e->env_procdata->syscallring);
        // Initialize the backend of the syscall ring buffer
        BACK_RING_INIT(&e->syscallbackring,
                       &e->env_procdata->syscallring,
                       SYSCALLRINGSIZE);

        // Initialize the generic sysevent ring buffer
        SHARED_RING_INIT(&e->env_procdata->syseventring);
        // Initialize the frontend of the sysevent ring buffer
        FRONT_RING_INIT(&e->syseventfrontring,
                        &e->env_procdata->syseventring,
                        SYSEVENTRINGSIZE);

        *newenv_store = e;
        atomic_inc(&num_envs);

        printk("[%08x] new env %08x\n", current ? current->env_id : 0, e->env_id);
        } // INIT_STRUCT
        return 0;
}

//
// Allocate len bytes of physical memory for environment env,
// and map it at virtual address va in the environment's address space.
// Does not zero or otherwise initialize the mapped pages in any way.
// Pages should be writable by user and kernel.
// Panic if any allocation attempt fails.
//
void
env_segment_alloc(env_t *e, void *SNT va, size_t len)
{
        void *SNT start, *SNT end;
        size_t num_pages;
        int i, r;
        page_t *page;
        pte_t *pte;

        start = ROUNDDOWN(va, PGSIZE);
        end = ROUNDUP(va + len, PGSIZE);
        if (start >= end)
                panic("Wrap-around in memory allocation addresses!");
        if ((uintptr_t)end > UTOP)
                panic("Attempting to map above UTOP!");
        // page_insert/pgdir_walk alloc a page and read/write to it via its address
        // starting from pgdir (e's), so we need to be using e's pgdir
        assert(e->env_cr3 == rcr3());
        num_pages = LA2PPN(end - start);

        for (i = 0; i < num_pages; i++, start += PGSIZE) {
                // skip if a page is already mapped.  yes, page_insert will page_remove
                // whatever page was already there, but if we are seg allocing adjacent
                // regions, we don't want to destroy that old mapping/page
                // though later on we are told we can ignore this...
                pte = pgdir_walk(e->env_pgdir, start, 0);
                if (pte && *pte & PTE_P)
                        continue;
                if ((r = page_alloc(&page)) < 0)
                        panic("env_segment_alloc: %e", r);
                page_insert(e->env_pgdir, page, start, PTE_USER_RW);
        }
}

void
env_segment_free(env_t *e, void *SNT va, size_t len)
{
        void *SNT start, *SNT end;
        size_t num_pages;
        page_t *page;
        pte_t *pte;

        // Round this up this time so we don't free the page that va is actually on
        start = ROUNDUP(va, PGSIZE);
        end = ROUNDUP(va + len, PGSIZE);
        if (start >= end)
                panic("Wrap-around in memory free addresses!");
        if ((uintptr_t)end > UTOP)
                panic("Attempting to unmap above UTOP!");
        // page_insert/pgdir_walk alloc a page and read/write to it via its address
        // starting from pgdir (e's), so we need to be using e's pgdir
        assert(e->env_cr3 == rcr3());
        num_pages = LA2PPN(end - start);

        for (int i = 0; i < num_pages; i++, start += PGSIZE) {
                // skip if a page is already unmapped. 
                pte = pgdir_walk(e->env_pgdir, start, 0);
                if (pte && *pte & PTE_P)
                        page_remove(e->env_pgdir,start);
        }
}

//
// Set up the initial program binary, stack, and processor flags
// for a user process.
//
// This function loads all loadable segments from the ELF binary image
// into the environment's user memory, starting at the appropriate
// virtual addresses indicated in the ELF program header.
// At the same time it clears to zero any portions of these segments
// that are marked in the program header as being mapped
// but not actually present in the ELF file - i.e., the program's bss section.
//
// Finally, this function maps one page for the program's initial stack.
static void*
load_icode(env_t *SAFE e, uint8_t *COUNT(size) binary, size_t size)
{
        // asw: copy the headers because they might not be aligned.
        elf_t elfhdr;
        proghdr_t phdr;
        void* _end = 0;
        memcpy(&elfhdr, binary, sizeof(elfhdr));

        int i, r;

        // is this an elf?
        assert(elfhdr.e_magic == ELF_MAGIC);
        // make sure we have proghdrs to load
        assert(elfhdr.e_phnum);

        // to actually access any pages alloc'd for this environment, we
        // need to have the hardware use this environment's page tables.
        uintreg_t old_cr3 = rcr3();
        /*
         * Even though we'll decref later and no one should be killing us at this
         * stage, we're still going to wrap the lcr3s with incref/decref.
         *
         * Note we never decref on the old_cr3, since we aren't willing to let it
         * die.  It's also not clear who the previous process is - sometimes it
         * isn't even a process (when the kernel loads on its own, and not in
         * response to a syscall).  Probably need to think more about this (TODO)
         *
         * This can get a bit tricky if this code blocks (will need to think about a
         * decref then), if we try to change states, etc.
         */
        proc_incref(e);
        lcr3(e->env_cr3);

        // TODO: how do we do a runtime COUNT?
        {TRUSTEDBLOCK // zra: TRUSTEDBLOCK until validation is done.
        for (i = 0; i < elfhdr.e_phnum; i++) {
                memcpy(&phdr, binary + elfhdr.e_phoff + i*sizeof(phdr), sizeof(phdr));
                if (phdr.p_type != ELF_PROG_LOAD)
                        continue;
                // TODO: validate elf header fields!
                // seg alloc creates PTE_U|PTE_W pages.  if you ever want to change
                // this, there will be issues with overlapping sections
                _end = MAX(_end, (void*)(phdr.p_va + phdr.p_memsz));
                env_segment_alloc(e, (void*SNT)phdr.p_va, phdr.p_memsz);
                memcpy((void*)phdr.p_va, binary + phdr.p_offset, phdr.p_filesz);
                memset((void*)phdr.p_va + phdr.p_filesz, 0, 
                              phdr.p_memsz - phdr.p_filesz);
        }}

        proc_set_program_counter(&e->env_tf, elfhdr.e_entry);
        e->env_entry = elfhdr.e_entry;

        // Now map USTACK_NUM_PAGES pages for the program's initial stack
        // starting at virtual address USTACKTOP - USTACK_NUM_PAGES*PGSIZE.
        env_segment_alloc(e, (void*SNT)(USTACKTOP - USTACK_NUM_PAGES*PGSIZE), 
                          USTACK_NUM_PAGES*PGSIZE);

        // reload the original address space
        lcr3(old_cr3);
        proc_decref(e);
        
        return _end;
}

//
// Allocates a new env and loads the named elf binary into it.
//
env_t* env_create(uint8_t *binary, size_t size)
{
        env_t *e;
        int r;
        envid_t curid;

        curid = (current ? current->env_id : 0);
        if ((r = env_alloc(&e, curid)) < 0)
                panic("env_create: %e", r);
        
        /* Load the binary and set the current locations of the elf segments.
         * All end-of-segment pointers are page aligned (invariant) */
        e->end_text_segment = load_icode(e, binary, size);
        e->end_data_segment = e->end_text_segment;

        return e;
}

//
// Frees env e and all memory it uses.
//
void
env_free(env_t *e)
{
        physaddr_t pa;

        // Note the environment's demise.
        printk("[%08x] free env %08x\n", current ? current->env_id : 0, e->env_id);
        // All parts of the kernel should have decref'd before env_free was called.
        assert(e->env_refcnt == 0);

        // Flush all mapped pages in the user portion of the address space
        env_user_mem_free(e);

        // free the page directory
        pa = e->env_cr3;
        e->env_pgdir = 0;
        e->env_cr3 = 0;
        page_decref(pa2page(pa));

        //Free any memory allocated by this process
        kfree(e->cache_colors_map);

        // return the environment to the free list
        e->state = ENV_FREE;
        spin_lock(&freelist_lock);
        TAILQ_INSERT_HEAD(&proc_freelist, e, proc_link);
        spin_unlock(&freelist_lock);
}


#define PER_CPU_THING(type,name)\
type SLOCKED(name##_lock) * RWPROTECT name;\
type SLOCKED(name##_lock) *\
(get_per_cpu_##name)()\
{\
        { R_PERMITTED(global(name))\
                return &name[core_id()];\
        }\
}


/* This is the top-half of an interrupt handler, where the bottom half is
 * proc_run (which never returns).  Just add it to the delayed work queue,
 * which (incidentally) can only hold one item at this point.
 *
 * Note this is rather old, and meant to run a RUNNABLE_S on a worker core.
 */
#ifdef __IVY__
void run_env_handler(trapframe_t *tf, env_t * data)
#else
void run_env_handler(trapframe_t *tf, void * data)
#endif
{
        assert(data);
        struct work TP(env_t *) job;
        struct workqueue TP(env_t *) *CT(1) workqueue =
            TC(&per_cpu_info[core_id()].workqueue);
        // this doesn't work, and making it a TP(env_t) is wrong
        // zra: When you want to use other types, let me know, and I can help
    // make something that Ivy is happy with. 
#ifdef __IVY__
        job.func = proc_run;
#else
        job.func = (func_t)proc_run;
#endif
        job.data = data;
        if (enqueue_work(workqueue, &job))
                panic("Failed to enqueue work!");
}