From 33dd94ae1ccbfb7bf0fb6c692bc3d1c4269e6177 Mon Sep 17 00:00:00 2001 From: Nelson Elhage Date: Thu, 2 Dec 2010 14:31:21 -0800 Subject: do_exit(): make sure that we run with get_fs() == USER_DS If a user manages to trigger an oops with fs set to KERNEL_DS, fs is not otherwise reset before do_exit(). do_exit may later (via mm_release in fork.c) do a put_user to a user-controlled address, potentially allowing a user to leverage an oops into a controlled write into kernel memory. This is only triggerable in the presence of another bug, but this potentially turns a lot of DoS bugs into privilege escalations, so it's worth fixing. I have proof-of-concept code which uses this bug along with CVE-2010-3849 to write a zero to an arbitrary kernel address, so I've tested that this is not theoretical. A more logical place to put this fix might be when we know an oops has occurred, before we call do_exit(), but that would involve changing every architecture, in multiple places. Let's just stick it in do_exit instead. [akpm@linux-foundation.org: update code comment] Signed-off-by: Nelson Elhage Cc: KOSAKI Motohiro Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'kernel') diff --git a/kernel/exit.c b/kernel/exit.c index 21aa7b3001fb..676149a4ac5f 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -914,6 +914,15 @@ NORET_TYPE void do_exit(long code) if (unlikely(!tsk->pid)) panic("Attempted to kill the idle task!"); + /* + * If do_exit is called because this processes oopsed, it's possible + * that get_fs() was left as KERNEL_DS, so reset it to USER_DS before + * continuing. Amongst other possible reasons, this is to prevent + * mm_release()->clear_child_tid() from writing to a user-controlled + * kernel address. + */ + set_fs(USER_DS); + tracehook_report_exit(&code); validate_creds_for_do_exit(tsk); -- cgit v1.2.3 From 9f339caf8454f0c21983111350ede93983db4340 Mon Sep 17 00:00:00 2001 From: Bojan Smojver Date: Thu, 25 Nov 2010 23:41:39 +0100 Subject: PM / Hibernate: Use async I/O when reading compressed hibernation image This is a fix for reading LZO compressed image using async I/O. Essentially, instead of having just one page into which we keep reading blocks from swap, we allocate enough of them to cover the largest compressed size and then let block I/O pick them all up. Once we have them all (and here we wait), we decompress them, as usual. Obviously, the very first block we still pick up synchronously, because we need to know the size of the lot before we pick up the rest. Also fixed the copyright line, which I've forgotten before. Signed-off-by: Bojan Smojver Signed-off-by: Rafael J. Wysocki --- kernel/power/swap.c | 53 ++++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 38 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/kernel/power/swap.c b/kernel/power/swap.c index a0e4a86ccf94..baf667bb2794 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -6,6 +6,7 @@ * * Copyright (C) 1998,2001-2005 Pavel Machek * Copyright (C) 2006 Rafael J. Wysocki + * Copyright (C) 2010 Bojan Smojver * * This file is released under the GPLv2. * @@ -753,30 +754,43 @@ static int load_image_lzo(struct swap_map_handle *handle, { unsigned int m; int error = 0; + struct bio *bio; struct timeval start; struct timeval stop; unsigned nr_pages; - size_t off, unc_len, cmp_len; - unsigned char *unc, *cmp, *page; + size_t i, off, unc_len, cmp_len; + unsigned char *unc, *cmp, *page[LZO_CMP_PAGES]; - page = (void *)__get_free_page(__GFP_WAIT | __GFP_HIGH); - if (!page) { - printk(KERN_ERR "PM: Failed to allocate LZO page\n"); - return -ENOMEM; + for (i = 0; i < LZO_CMP_PAGES; i++) { + page[i] = (void *)__get_free_page(__GFP_WAIT | __GFP_HIGH); + if (!page[i]) { + printk(KERN_ERR "PM: Failed to allocate LZO page\n"); + + while (i) + free_page((unsigned long)page[--i]); + + return -ENOMEM; + } } unc = vmalloc(LZO_UNC_SIZE); if (!unc) { printk(KERN_ERR "PM: Failed to allocate LZO uncompressed\n"); - free_page((unsigned long)page); + + for (i = 0; i < LZO_CMP_PAGES; i++) + free_page((unsigned long)page[i]); + return -ENOMEM; } cmp = vmalloc(LZO_CMP_SIZE); if (!cmp) { printk(KERN_ERR "PM: Failed to allocate LZO compressed\n"); + vfree(unc); - free_page((unsigned long)page); + for (i = 0; i < LZO_CMP_PAGES; i++) + free_page((unsigned long)page[i]); + return -ENOMEM; } @@ -787,6 +801,7 @@ static int load_image_lzo(struct swap_map_handle *handle, if (!m) m = 1; nr_pages = 0; + bio = NULL; do_gettimeofday(&start); error = snapshot_write_next(snapshot); @@ -794,11 +809,11 @@ static int load_image_lzo(struct swap_map_handle *handle, goto out_finish; for (;;) { - error = swap_read_page(handle, page, NULL); /* sync */ + error = swap_read_page(handle, page[0], NULL); /* sync */ if (error) break; - cmp_len = *(size_t *)page; + cmp_len = *(size_t *)page[0]; if (unlikely(!cmp_len || cmp_len > lzo1x_worst_compress(LZO_UNC_SIZE))) { printk(KERN_ERR "PM: Invalid LZO compressed length\n"); @@ -806,13 +821,20 @@ static int load_image_lzo(struct swap_map_handle *handle, break; } - memcpy(cmp, page, PAGE_SIZE); - for (off = PAGE_SIZE; off < LZO_HEADER + cmp_len; off += PAGE_SIZE) { - error = swap_read_page(handle, page, NULL); /* sync */ + for (off = PAGE_SIZE, i = 1; + off < LZO_HEADER + cmp_len; off += PAGE_SIZE, i++) { + error = swap_read_page(handle, page[i], &bio); if (error) goto out_finish; + } - memcpy(cmp + off, page, PAGE_SIZE); + error = hib_wait_on_bio_chain(&bio); /* need all data now */ + if (error) + goto out_finish; + + for (off = 0, i = 0; + off < LZO_HEADER + cmp_len; off += PAGE_SIZE, i++) { + memcpy(cmp + off, page[i], PAGE_SIZE); } unc_len = LZO_UNC_SIZE; @@ -857,7 +879,8 @@ out_finish: vfree(cmp); vfree(unc); - free_page((unsigned long)page); + for (i = 0; i < LZO_CMP_PAGES; i++) + free_page((unsigned long)page[i]); return error; } -- cgit v1.2.3 From c9e664f1fdf34aa8cede047b206deaa8f1945af0 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 3 Dec 2010 22:57:45 +0100 Subject: PM / Hibernate: Fix memory corruption related to swap There is a problem that swap pages allocated before the creation of a hibernation image can be released and used for storing the contents of different memory pages while the image is being saved. Since the kernel stored in the image doesn't know of that, it causes memory corruption to occur after resume from hibernation, especially on systems with relatively small RAM that need to swap often. This issue can be addressed by keeping the GFP_IOFS bits clear in gfp_allowed_mask during the entire hibernation, including the saving of the image, until the system is finally turned off or the hibernation is aborted. Unfortunately, for this purpose it's necessary to rework the way in which the hibernate and suspend code manipulates gfp_allowed_mask. This change is based on an earlier patch from Hugh Dickins. Signed-off-by: Rafael J. Wysocki Reported-by: Ondrej Zary Acked-by: Hugh Dickins Reviewed-by: KAMEZAWA Hiroyuki Cc: stable@kernel.org --- kernel/power/hibernate.c | 22 ++++++++++++---------- kernel/power/suspend.c | 5 ++--- kernel/power/user.c | 2 ++ 3 files changed, 16 insertions(+), 13 deletions(-) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index 657272e91d0a..048d0b514831 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -327,7 +327,6 @@ static int create_image(int platform_mode) int hibernation_snapshot(int platform_mode) { int error; - gfp_t saved_mask; error = platform_begin(platform_mode); if (error) @@ -339,7 +338,7 @@ int hibernation_snapshot(int platform_mode) goto Close; suspend_console(); - saved_mask = clear_gfp_allowed_mask(GFP_IOFS); + pm_restrict_gfp_mask(); error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Recover_platform; @@ -348,7 +347,10 @@ int hibernation_snapshot(int platform_mode) goto Recover_platform; error = create_image(platform_mode); - /* Control returns here after successful restore */ + /* + * Control returns here (1) after the image has been created or the + * image creation has failed and (2) after a successful restore. + */ Resume_devices: /* We may need to release the preallocated image pages here. */ @@ -357,7 +359,10 @@ int hibernation_snapshot(int platform_mode) dpm_resume_end(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); - set_gfp_allowed_mask(saved_mask); + + if (error || !in_suspend) + pm_restore_gfp_mask(); + resume_console(); Close: platform_end(platform_mode); @@ -452,17 +457,16 @@ static int resume_target_kernel(bool platform_mode) int hibernation_restore(int platform_mode) { int error; - gfp_t saved_mask; pm_prepare_console(); suspend_console(); - saved_mask = clear_gfp_allowed_mask(GFP_IOFS); + pm_restrict_gfp_mask(); error = dpm_suspend_start(PMSG_QUIESCE); if (!error) { error = resume_target_kernel(platform_mode); dpm_resume_end(PMSG_RECOVER); } - set_gfp_allowed_mask(saved_mask); + pm_restore_gfp_mask(); resume_console(); pm_restore_console(); return error; @@ -476,7 +480,6 @@ int hibernation_restore(int platform_mode) int hibernation_platform_enter(void) { int error; - gfp_t saved_mask; if (!hibernation_ops) return -ENOSYS; @@ -492,7 +495,6 @@ int hibernation_platform_enter(void) entering_platform_hibernation = true; suspend_console(); - saved_mask = clear_gfp_allowed_mask(GFP_IOFS); error = dpm_suspend_start(PMSG_HIBERNATE); if (error) { if (hibernation_ops->recover) @@ -536,7 +538,6 @@ int hibernation_platform_enter(void) Resume_devices: entering_platform_hibernation = false; dpm_resume_end(PMSG_RESTORE); - set_gfp_allowed_mask(saved_mask); resume_console(); Close: @@ -646,6 +647,7 @@ int hibernate(void) swsusp_free(); if (!error) power_down(); + pm_restore_gfp_mask(); } else { pr_debug("PM: Image restored successfully.\n"); } diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index 7335952ee473..ecf770509d0d 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -197,7 +197,6 @@ static int suspend_enter(suspend_state_t state) int suspend_devices_and_enter(suspend_state_t state) { int error; - gfp_t saved_mask; if (!suspend_ops) return -ENOSYS; @@ -208,7 +207,7 @@ int suspend_devices_and_enter(suspend_state_t state) goto Close; } suspend_console(); - saved_mask = clear_gfp_allowed_mask(GFP_IOFS); + pm_restrict_gfp_mask(); suspend_test_start(); error = dpm_suspend_start(PMSG_SUSPEND); if (error) { @@ -225,7 +224,7 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); dpm_resume_end(PMSG_RESUME); suspend_test_finish("resume devices"); - set_gfp_allowed_mask(saved_mask); + pm_restore_gfp_mask(); resume_console(); Close: if (suspend_ops->end) diff --git a/kernel/power/user.c b/kernel/power/user.c index e819e17877ca..1b2ea31e6bd8 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -263,6 +263,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, case SNAPSHOT_UNFREEZE: if (!data->frozen || data->ready) break; + pm_restore_gfp_mask(); thaw_processes(); usermodehelper_enable(); data->frozen = 0; @@ -275,6 +276,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, error = -EPERM; break; } + pm_restore_gfp_mask(); error = hibernation_snapshot(data->platform_support); if (!error) error = put_user(in_suspend, (int __user *)arg); -- cgit v1.2.3