June 2021 - Kernel - mailweb.openeuler.org

[PATCH openEuler-1.0-LTS] io_uring: NULL files dereference by SQPOLL
by Yang Yingliang 30 Jun '21

30 Jun '21

From: Pavel Begunkov <asml.silence(a)gmail.com> mainline inclusion from mainline-5.11-rc1 commit 1a38ffc9cbca361cc274d6e234f5ef8922f0b6d9 category: bugfix bugzilla: 167380 CVE: NA --------------------------- SQPOLL task may find sqo_task->files == NULL and __io_sq_thread_acquire_files() would leave it unset, so following fget_many() and others try to dereference NULL and fault. Propagate an error files are missing. [ 118.962785] BUG: kernel NULL pointer dereference, address: 0000000000000020 [ 118.963812] #PF: supervisor read access in kernel mode [ 118.964534] #PF: error_code(0x0000) - not-present page [ 118.969029] RIP: 0010:__fget_files+0xb/0x80 [ 119.005409] Call Trace: [ 119.005651] fget_many+0x2b/0x30 [ 119.005964] io_file_get+0xcf/0x180 [ 119.006315] io_submit_sqes+0x3a4/0x950 [ 119.007481] io_sq_thread+0x1de/0x6a0 [ 119.007828] kthread+0x114/0x150 [ 119.008963] ret_from_fork+0x22/0x30 Reported-by: Josef Grieb <josef.grieb(a)gmail.com> Signed-off-by: Pavel Begunkov <asml.silence(a)gmail.com> Signed-off-by: Jens Axboe <axboe(a)kernel.dk> Conflicts: fs/io_uring.c [ aa1a90087b("io_uring: allow non-fixed files with SQPOLL") merge conflicts. ] Signed-off-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- fs/io_uring.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index f661231cdfad6..c944a27178a48 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -978,7 +978,7 @@ static void io_sq_thread_drop_mm_files(void) } } -static void __io_sq_thread_acquire_files(struct io_ring_ctx *ctx) +static int __io_sq_thread_acquire_files(struct io_ring_ctx *ctx) { if (!current->files) { struct files_struct *files; @@ -988,7 +988,7 @@ static void __io_sq_thread_acquire_files(struct io_ring_ctx *ctx) files = ctx->sqo_task->files; if (!files) { task_unlock(ctx->sqo_task); - return; + return -EOWNERDEAD; } atomic_inc(&files->count); get_nsproxy(ctx->sqo_task->nsproxy); @@ -1000,6 +1000,7 @@ static void __io_sq_thread_acquire_files(struct io_ring_ctx *ctx) current->nsproxy = nsproxy; task_unlock(current); } + return 0; } static int __io_sq_thread_acquire_mm(struct io_ring_ctx *ctx) @@ -1031,15 +1032,19 @@ static int io_sq_thread_acquire_mm_files(struct io_ring_ctx *ctx, struct io_kiocb *req) { const struct io_op_def *def = &io_op_defs[req->opcode]; + int ret; if (def->needs_mm) { - int ret = __io_sq_thread_acquire_mm(ctx); + ret = __io_sq_thread_acquire_mm(ctx); if (unlikely(ret)) return ret; } - if (def->needs_file || def->file_table) - __io_sq_thread_acquire_files(ctx); + if (def->needs_file || def->file_table) { + ret = __io_sq_thread_acquire_files(ctx); + if (unlikely(ret)) + return ret; + } return 0; } @@ -1878,8 +1883,8 @@ static void __io_req_task_submit(struct io_kiocb *req) { struct io_ring_ctx *ctx = req->ctx; - if (!__io_sq_thread_acquire_mm(ctx)) { - __io_sq_thread_acquire_files(ctx); + if (!__io_sq_thread_acquire_mm(ctx) && + !__io_sq_thread_acquire_files(ctx)) { mutex_lock(&ctx->uring_lock); __io_queue_sqe(req, NULL, NULL); mutex_unlock(&ctx->uring_lock); -- 2.25.1

1 0

[PATCH openEuler-1.0-LTS] vgacon: remove software scrollback support
by Yang Yingliang 29 Jun '21

29 Jun '21

From: Linus Torvalds <torvalds(a)linux-foundation.org> stable inclusion from linux-4.19.146 commit f5fa64c8daf7b97280865c73903edc0a3eea819e CVE: CVE-2020-28097 -------------------------------- commit 973c096f6a85e5b5f2a295126ba6928d9a6afd45 upstream. Yunhai Zhang recently fixed a VGA software scrollback bug in commit ebfdfeeae8c0 ("vgacon: Fix for missing check in scrollback handling"), but that then made people look more closely at some of this code, and there were more problems on the vgacon side, but also the fbcon software scrollback. We don't really have anybody who maintains this code - probably because nobody actually _uses_ it any more. Sure, people still use both VGA and the framebuffer consoles, but they are no longer the main user interfaces to the kernel, and haven't been for decades, so these kinds of extra features end up bitrotting and not really being used. So rather than try to maintain a likely unused set of code, I'll just aggressively remove it, and see if anybody even notices. Maybe there are people who haven't jumped on the whole GUI badnwagon yet, and think it's just a fad. And maybe those people use the scrollback code. If that turns out to be the case, we can resurrect this again, once we've found the sucker^Wmaintainer for it who actually uses it. Reported-by: NopNop Nop <nopitydays(a)gmail.com> Tested-by: Willy Tarreau <w(a)1wt.eu> Cc: 张云海 <zhangyunhai(a)nsfocus.com> Acked-by: Andy Lutomirski <luto(a)amacapital.net> Acked-by: Willy Tarreau <w(a)1wt.eu> Reviewed-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- arch/powerpc/configs/pasemi_defconfig | 1 - arch/powerpc/configs/ppc6xx_defconfig | 1 - arch/x86/configs/i386_defconfig | 1 - arch/x86/configs/x86_64_defconfig | 1 - drivers/video/console/Kconfig | 46 ------ drivers/video/console/vgacon.c | 221 +------------------------- 6 files changed, 1 insertion(+), 270 deletions(-) diff --git a/arch/powerpc/configs/pasemi_defconfig b/arch/powerpc/configs/pasemi_defconfig index 6daa56f8895cb..5f743db4c9f8c 100644 --- a/arch/powerpc/configs/pasemi_defconfig +++ b/arch/powerpc/configs/pasemi_defconfig @@ -111,7 +111,6 @@ CONFIG_FB_NVIDIA=y CONFIG_FB_NVIDIA_I2C=y CONFIG_FB_RADEON=y # CONFIG_LCD_CLASS_DEVICE is not set -CONFIG_VGACON_SOFT_SCROLLBACK=y CONFIG_LOGO=y CONFIG_SOUND=y CONFIG_SND=y diff --git a/arch/powerpc/configs/ppc6xx_defconfig b/arch/powerpc/configs/ppc6xx_defconfig index 7ee736f207740..533db43431727 100644 --- a/arch/powerpc/configs/ppc6xx_defconfig +++ b/arch/powerpc/configs/ppc6xx_defconfig @@ -780,7 +780,6 @@ CONFIG_FB_TRIDENT=m CONFIG_FB_SM501=m CONFIG_FB_IBM_GXT4500=y CONFIG_LCD_PLATFORM=m -CONFIG_VGACON_SOFT_SCROLLBACK=y CONFIG_FRAMEBUFFER_CONSOLE=y CONFIG_FRAMEBUFFER_CONSOLE_ROTATION=y CONFIG_LOGO=y diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig index 0eb9f92f37179..7aab9c4f86a9f 100644 --- a/arch/x86/configs/i386_defconfig +++ b/arch/x86/configs/i386_defconfig @@ -217,7 +217,6 @@ CONFIG_FB_MODE_HELPERS=y CONFIG_FB_TILEBLITTING=y CONFIG_FB_EFI=y # CONFIG_LCD_CLASS_DEVICE is not set -CONFIG_VGACON_SOFT_SCROLLBACK=y CONFIG_LOGO=y # CONFIG_LOGO_LINUX_MONO is not set # CONFIG_LOGO_LINUX_VGA16 is not set diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig index e32fc1f274d85..7484a27861d4c 100644 --- a/arch/x86/configs/x86_64_defconfig +++ b/arch/x86/configs/x86_64_defconfig @@ -213,7 +213,6 @@ CONFIG_FB_MODE_HELPERS=y CONFIG_FB_TILEBLITTING=y CONFIG_FB_EFI=y # CONFIG_LCD_CLASS_DEVICE is not set -CONFIG_VGACON_SOFT_SCROLLBACK=y CONFIG_LOGO=y # CONFIG_LOGO_LINUX_MONO is not set # CONFIG_LOGO_LINUX_VGA16 is not set diff --git a/drivers/video/console/Kconfig b/drivers/video/console/Kconfig index 787792c3d08d6..40d5fea8513cf 100644 --- a/drivers/video/console/Kconfig +++ b/drivers/video/console/Kconfig @@ -21,52 +21,6 @@ config VGA_CONSOLE Say Y. -config VGACON_SOFT_SCROLLBACK - bool "Enable Scrollback Buffer in System RAM" - depends on VGA_CONSOLE - default n - help - The scrollback buffer of the standard VGA console is located in - the VGA RAM. The size of this RAM is fixed and is quite small. - If you require a larger scrollback buffer, this can be placed in - System RAM which is dynamically allocated during initialization. - Placing the scrollback buffer in System RAM will slightly slow - down the console. - - If you want this feature, say 'Y' here and enter the amount of - RAM to allocate for this buffer. If unsure, say 'N'. - -config VGACON_SOFT_SCROLLBACK_SIZE - int "Scrollback Buffer Size (in KB)" - depends on VGACON_SOFT_SCROLLBACK - range 1 1024 - default "64" - help - Enter the amount of System RAM to allocate for scrollback - buffers of VGA consoles. Each 64KB will give you approximately - 16 80x25 screenfuls of scrollback buffer. - -config VGACON_SOFT_SCROLLBACK_PERSISTENT_ENABLE_BY_DEFAULT - bool "Persistent Scrollback History for each console by default" - depends on VGACON_SOFT_SCROLLBACK - default n - help - Say Y here if the scrollback history should persist by default when - switching between consoles. Otherwise, the scrollback history will be - flushed each time the console is switched. This feature can also be - enabled using the boot command line parameter - 'vgacon.scrollback_persistent=1'. - - This feature might break your tool of choice to flush the scrollback - buffer, e.g. clear(1) will work fine but Debian's clear_console(1) - will be broken, which might cause security issues. - You can use the escape sequence \e[3J instead if this feature is - activated. - - Note that a buffer of VGACON_SOFT_SCROLLBACK_SIZE is taken for each - created tty device. - So if you use a RAM-constrained system, say N here. - config MDA_CONSOLE depends on !M68K && !PARISC && ISA tristate "MDA text console (dual-headed)" diff --git a/drivers/video/console/vgacon.c b/drivers/video/console/vgacon.c index e079b910feb24..55507df335bdd 100644 --- a/drivers/video/console/vgacon.c +++ b/drivers/video/console/vgacon.c @@ -165,214 +165,6 @@ static inline void vga_set_mem_top(struct vc_data *c) write_vga(12, (c->vc_visible_origin - vga_vram_base) / 2); } -#ifdef CONFIG_VGACON_SOFT_SCROLLBACK -/* software scrollback */ -struct vgacon_scrollback_info { - void *data; - int tail; - int size; - int rows; - int cnt; - int cur; - int save; - int restore; -}; - -static struct vgacon_scrollback_info *vgacon_scrollback_cur; -static struct vgacon_scrollback_info vgacon_scrollbacks[MAX_NR_CONSOLES]; -static bool scrollback_persistent = \ - IS_ENABLED(CONFIG_VGACON_SOFT_SCROLLBACK_PERSISTENT_ENABLE_BY_DEFAULT); -module_param_named(scrollback_persistent, scrollback_persistent, bool, 0000); -MODULE_PARM_DESC(scrollback_persistent, "Enable persistent scrollback for all vga consoles"); - -static void vgacon_scrollback_reset(int vc_num, size_t reset_size) -{ - struct vgacon_scrollback_info *scrollback = &vgacon_scrollbacks[vc_num]; - - if (scrollback->data && reset_size > 0) - memset(scrollback->data, 0, reset_size); - - scrollback->cnt = 0; - scrollback->tail = 0; - scrollback->cur = 0; -} - -static void vgacon_scrollback_init(int vc_num) -{ - int pitch = vga_video_num_columns * 2; - size_t size = CONFIG_VGACON_SOFT_SCROLLBACK_SIZE * 1024; - int rows = size / pitch; - void *data; - - data = kmalloc_array(CONFIG_VGACON_SOFT_SCROLLBACK_SIZE, 1024, - GFP_NOWAIT); - - vgacon_scrollbacks[vc_num].data = data; - vgacon_scrollback_cur = &vgacon_scrollbacks[vc_num]; - - vgacon_scrollback_cur->rows = rows - 1; - vgacon_scrollback_cur->size = rows * pitch; - - vgacon_scrollback_reset(vc_num, size); -} - -static void vgacon_scrollback_switch(int vc_num) -{ - if (!scrollback_persistent) - vc_num = 0; - - if (!vgacon_scrollbacks[vc_num].data) { - vgacon_scrollback_init(vc_num); - } else { - if (scrollback_persistent) { - vgacon_scrollback_cur = &vgacon_scrollbacks[vc_num]; - } else { - size_t size = CONFIG_VGACON_SOFT_SCROLLBACK_SIZE * 1024; - - vgacon_scrollback_reset(vc_num, size); - } - } -} - -static void vgacon_scrollback_startup(void) -{ - vgacon_scrollback_cur = &vgacon_scrollbacks[0]; - vgacon_scrollback_init(0); -} - -static void vgacon_scrollback_update(struct vc_data *c, int t, int count) -{ - void *p; - - if (!vgacon_scrollback_cur->data || !vgacon_scrollback_cur->size || - c->vc_num != fg_console) - return; - - p = (void *) (c->vc_origin + t * c->vc_size_row); - - while (count--) { - if ((vgacon_scrollback_cur->tail + c->vc_size_row) > - vgacon_scrollback_cur->size) - vgacon_scrollback_cur->tail = 0; - - scr_memcpyw(vgacon_scrollback_cur->data + - vgacon_scrollback_cur->tail, - p, c->vc_size_row); - - vgacon_scrollback_cur->cnt++; - p += c->vc_size_row; - vgacon_scrollback_cur->tail += c->vc_size_row; - - if (vgacon_scrollback_cur->tail >= vgacon_scrollback_cur->size) - vgacon_scrollback_cur->tail = 0; - - if (vgacon_scrollback_cur->cnt > vgacon_scrollback_cur->rows) - vgacon_scrollback_cur->cnt = vgacon_scrollback_cur->rows; - - vgacon_scrollback_cur->cur = vgacon_scrollback_cur->cnt; - } -} - -static void vgacon_restore_screen(struct vc_data *c) -{ - c->vc_origin = c->vc_visible_origin; - vgacon_scrollback_cur->save = 0; - - if (!vga_is_gfx && !vgacon_scrollback_cur->restore) { - scr_memcpyw((u16 *) c->vc_origin, (u16 *) c->vc_screenbuf, - c->vc_screenbuf_size > vga_vram_size ? - vga_vram_size : c->vc_screenbuf_size); - vgacon_scrollback_cur->restore = 1; - vgacon_scrollback_cur->cur = vgacon_scrollback_cur->cnt; - } -} - -static void vgacon_scrolldelta(struct vc_data *c, int lines) -{ - int start, end, count, soff; - - if (!lines) { - vgacon_restore_screen(c); - return; - } - - if (!vgacon_scrollback_cur->data) - return; - - if (!vgacon_scrollback_cur->save) { - vgacon_cursor(c, CM_ERASE); - vgacon_save_screen(c); - c->vc_origin = (unsigned long)c->vc_screenbuf; - vgacon_scrollback_cur->save = 1; - } - - vgacon_scrollback_cur->restore = 0; - start = vgacon_scrollback_cur->cur + lines; - end = start + abs(lines); - - if (start < 0) - start = 0; - - if (start > vgacon_scrollback_cur->cnt) - start = vgacon_scrollback_cur->cnt; - - if (end < 0) - end = 0; - - if (end > vgacon_scrollback_cur->cnt) - end = vgacon_scrollback_cur->cnt; - - vgacon_scrollback_cur->cur = start; - count = end - start; - soff = vgacon_scrollback_cur->tail - - ((vgacon_scrollback_cur->cnt - end) * c->vc_size_row); - soff -= count * c->vc_size_row; - - if (soff < 0) - soff += vgacon_scrollback_cur->size; - - count = vgacon_scrollback_cur->cnt - start; - - if (count > c->vc_rows) - count = c->vc_rows; - - if (count) { - int copysize; - - int diff = c->vc_rows - count; - void *d = (void *) c->vc_visible_origin; - void *s = (void *) c->vc_screenbuf; - - count *= c->vc_size_row; - /* how much memory to end of buffer left? */ - copysize = min(count, vgacon_scrollback_cur->size - soff); - scr_memcpyw(d, vgacon_scrollback_cur->data + soff, copysize); - d += copysize; - count -= copysize; - - if (count) { - scr_memcpyw(d, vgacon_scrollback_cur->data, count); - d += count; - } - - if (diff) - scr_memcpyw(d, s, diff * c->vc_size_row); - } else - vgacon_cursor(c, CM_MOVE); -} - -static void vgacon_flush_scrollback(struct vc_data *c) -{ - size_t size = CONFIG_VGACON_SOFT_SCROLLBACK_SIZE * 1024; - - vgacon_scrollback_reset(c->vc_num, size); -} -#else -#define vgacon_scrollback_startup(...) do { } while (0) -#define vgacon_scrollback_init(...) do { } while (0) -#define vgacon_scrollback_update(...) do { } while (0) -#define vgacon_scrollback_switch(...) do { } while (0) - static void vgacon_restore_screen(struct vc_data *c) { if (c->vc_origin != c->vc_visible_origin) @@ -386,11 +178,6 @@ static void vgacon_scrolldelta(struct vc_data *c, int lines) vga_set_mem_top(c); } -static void vgacon_flush_scrollback(struct vc_data *c) -{ -} -#endif /* CONFIG_VGACON_SOFT_SCROLLBACK */ - static const char *vgacon_startup(void) { const char *display_desc = NULL; @@ -573,10 +360,7 @@ static const char *vgacon_startup(void) vgacon_xres = screen_info.orig_video_cols * VGA_FONTWIDTH; vgacon_yres = vga_scan_lines; - if (!vga_init_done) { - vgacon_scrollback_startup(); - vga_init_done = true; - } + vga_init_done = true; return display_desc; } @@ -867,7 +651,6 @@ static int vgacon_switch(struct vc_data *c) vgacon_doresize(c, c->vc_cols, c->vc_rows); } - vgacon_scrollback_switch(c->vc_num); return 0; /* Redrawing not needed */ } @@ -1384,7 +1167,6 @@ static bool vgacon_scroll(struct vc_data *c, unsigned int t, unsigned int b, oldo = c->vc_origin; delta = lines * c->vc_size_row; if (dir == SM_UP) { - vgacon_scrollback_update(c, t, lines); if (c->vc_scr_end + delta >= vga_vram_end) { scr_memcpyw((u16 *) vga_vram_base, (u16 *) (oldo + delta), @@ -1448,7 +1230,6 @@ const struct consw vga_con = { .con_save_screen = vgacon_save_screen, .con_build_attr = vgacon_build_attr, .con_invert_region = vgacon_invert_region, - .con_flush_scrollback = vgacon_flush_scrollback, }; EXPORT_SYMBOL(vga_con); -- 2.25.1

1 0

openEuler 任务打榜赛6月21日-6月28日周榜单来啦
by Marketing openEuler 29 Jun '21

29 Jun '21

openEuler 任务打榜赛6月21日-6月28日周榜单来啦，快来看看有没有你的名字吧。这里提醒已领取赛题但是还没有加群的小伙伴赶紧联系小助手加群呀。好啦~，下面让我们看一下有哪些开发者上榜吧！ gitee用户账号积分值解决issue/pr数量 yaozc701 20 4 dada5208 16 5 shuyao13 14 4 houlibao 12 3 liuerwuliu 12 3 hw999666 8 2 gc7500 8 2 weidongkl 6 1 shao1112 6 2 wss1235 2 1 恭喜以上开发者，请联系小助手领奖。没有上榜的小伙伴不要灰心，请再接再历。周榜、月榜前十名均有奖励，期待在榜单上早点见到你。添加小助手微信（SoftCorePro）回复“*任务打榜赛*”进入交流群 *了解更多* - 官网：https://openeuler.org/zh/interaction/hdc/2021-developer.html - 我要报名： https://competition.huaweicloud.com/information/1000041420/introduction - 我要做题： https://gitee.com/organizations/src-openeuler/issues?assignee_id=&author_id…

1 0

[PATCH kernel-4.19] vhost_net: avoid tx queue stuck when sendmsg fails
by Yang Yingliang 28 Jun '21

28 Jun '21

From: Yunjian Wang <wangyunjian(a)huawei.com> mainline inclusion from mainline-5.12-rc1 commit dc9c9e72ff3ba01ae63e6263ac26234ba1869cd7 category: bugfix bugzilla: NA CVE: NA ------------------------------------------------- Currently the driver doesn't drop a packet which can't be sent by tun (e.g bad packet). In this case, the driver will always process the same packet lead to the tx queue stuck. To fix this issue: 1. in the case of persistent failure (e.g bad packet), the driver can skip this descriptor by ignoring the error. 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM), the driver schedules the worker to try again. Signed-off-by: Yunjian Wang <wangyunjian(a)huawei.com> Acked-by: Jason Wang <jasowang(a)redhat.com> Acked-by: Willem de Bruijn <willemb(a)google.com> Acked-by: Michael S. Tsirkin <mst(a)redhat.com> Link: https://lore.kernel.org/r/1610685980-38608-1-git-send-email-wangyunjian@hua… Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> Reviewed-by: Yue Haibing <yuehaibing(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/vhost/net.c | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index 0c7bbc92b22a9..1d99f5c443eeb 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -580,14 +580,15 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock) else msg.msg_flags &= ~MSG_MORE; - /* TODO: Check specific error and bomb out unless ENOBUFS? */ err = sock->ops->sendmsg(sock, &msg, len); if (unlikely(err < 0)) { - vhost_discard_vq_desc(vq, 1); - vhost_net_enable_vq(net, vq); - break; - } - if (err != len) + if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) { + vhost_discard_vq_desc(vq, 1); + vhost_net_enable_vq(net, vq); + break; + } + pr_debug("Fail to send packet: err %d", err); + } else if (unlikely(err != len)) pr_debug("Truncated TX packet: len %d != %zd\n", err, len); if (++nvq->done_idx >= VHOST_NET_BATCH) @@ -670,7 +671,6 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock) msg.msg_flags &= ~MSG_MORE; } - /* TODO: Check specific error and bomb out unless ENOBUFS? */ err = sock->ops->sendmsg(sock, &msg, len); if (unlikely(err < 0)) { if (zcopy_used) { @@ -679,11 +679,13 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock) nvq->upend_idx = ((unsigned)nvq->upend_idx - 1) % UIO_MAXIOV; } - vhost_discard_vq_desc(vq, 1); - vhost_net_enable_vq(net, vq); - break; - } - if (err != len) + if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) { + vhost_discard_vq_desc(vq, 1); + vhost_net_enable_vq(net, vq); + break; + } + pr_debug("Fail to send packet: err %d", err); + } else if (unlikely(err != len)) pr_debug("Truncated TX packet: " " len %d != %zd\n", err, len); if (!zcopy_used) -- 2.25.1

1 0

[PATCH hulk-4.19-next] itrace: add missing itrace_hardirqs_ignore() calls
by Bixuan Cui 28 Jun '21

28 Jun '21

From: Bixuan Cui <c00283067(a)huawei.com> ascend inclusion category: feature bugzilla: NA DTS: #601 CVE: NA --------------------------------------------- There is an anomalous statistics in itrace irqsoff: max_time:3945(us) caller:__do_softirq+0x8c/0x310 max_time:7996(us) caller:__do_softirq+0x8c/0x310 Use function tracer to trace: stress-ng-cpu-3723 [011] d... 2817.510377: rcu_nmi_exit <-rcu_irq_exit stress-ng-cpu-3723 [011] d... 2817.510377: rcu_dynticks_curr_cpu_in_eqs <-rcu_nmi_exit stress-ng-cpu-3723 [011] d... 2817.514356: gic_handle_irq <-el0_irq_naked stress-ng-cpu-3723 [011] d... 2817.514356: __handle_domain_irq <-gic_handle_irq ... The per-cpu interrupt is enabled after irq_handler in el0_irq_naked() is executed. However, the itrace does not add the irqsoff monitoring to el0_irq_naked. The same omission appears in the following files: arch/arm64/include/asm/daifflags.h arch/arm64/kernel/entry.S arch/arm64/kernel/syscall.c arch/arm64/mm/fault.c Add missing itrace_hardirqs_ignore() calls into it. Signed-off-by: Bixuan Cui <cuibixuan(a)huawei.com> --- arch/arm64/include/asm/daifflags.h | 3 +++ arch/arm64/kernel/entry.S | 9 +++++++++ arch/arm64/kernel/syscall.c | 3 +++ arch/arm64/mm/fault.c | 6 +++++- 4 files changed, 20 insertions(+), 1 deletion(-) diff --git a/arch/arm64/include/asm/daifflags.h b/arch/arm64/include/asm/daifflags.h index 1230923b032d..1d832814eae8 100644 --- a/arch/arm64/include/asm/daifflags.h +++ b/arch/arm64/include/asm/daifflags.h @@ -72,6 +72,9 @@ static inline void local_daif_restore(unsigned long flags) if (!irq_disabled) { trace_hardirqs_on(); +#ifdef CONFIG_ITRACE_IRQSOFF + itrace_hardirqs_ignore(); +#endif if (system_uses_irq_prio_masking()) { gic_write_pmr(GIC_PRIO_IRQON); diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S index 2eb1b657de2f..2c9b8d94d367 100644 --- a/arch/arm64/kernel/entry.S +++ b/arch/arm64/kernel/entry.S @@ -718,6 +718,9 @@ alternative_else_nop_endif #endif bl trace_hardirqs_on 1: +#endif +#ifdef CONFIG_ITRACE_IRQSOFF + bl itrace_hardirqs_ignore #endif kernel_exit 1 @@ -954,6 +957,9 @@ el0_irq_naked: #ifdef CONFIG_TRACE_IRQFLAGS bl trace_hardirqs_on +#endif +#ifdef CONFIG_ITRACE_IRQSOFF + bl itrace_hardirqs_ignore #endif b ret_to_user ENDPROC(el0_irq) @@ -989,6 +995,9 @@ work_pending: bl do_notify_resume #ifdef CONFIG_TRACE_IRQFLAGS bl trace_hardirqs_on // enabled while in userspace +#endif +#ifdef CONFIG_ITRACE_IRQSOFF + bl itrace_hardirqs_ignore #endif ldr x1, [tsk, #TSK_TI_FLAGS] // re-check for single-step b finish_ret_to_user diff --git a/arch/arm64/kernel/syscall.c b/arch/arm64/kernel/syscall.c index f2d2dbbbfca2..964df44d7429 100644 --- a/arch/arm64/kernel/syscall.c +++ b/arch/arm64/kernel/syscall.c @@ -131,6 +131,9 @@ static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr, * the SPSR. */ trace_hardirqs_on(); +#ifdef CONFIG_ITRACE_IRQSOFF + itrace_hardirqs_ignore(); +#endif return; } local_daif_restore(DAIF_PROCCTX); diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c index 6cd448d9835c..9f30be41bb87 100644 --- a/arch/arm64/mm/fault.c +++ b/arch/arm64/mm/fault.c @@ -934,8 +934,12 @@ asmlinkage int __exception do_debug_exception(unsigned long addr_if_watchpoint, rv = 0; } - if (interrupts_enabled(regs)) + if (interrupts_enabled(regs)) { trace_hardirqs_on(); +#ifdef CONFIG_ITRACE_IRQSOFF + itrace_hardirqs_ignore(); +#endif + } return rv; } -- 2.17.1

1 1

[PATCH kernel-4.19 1/2] mm: vmalloc: prevent use after free in _vm_unmap_aliases
by Yang Yingliang 28 Jun '21

28 Jun '21

From: Vijayanand Jitta <vjitta(a)codeaurora.org> mainline inclusion from mainline-5.13-rc1 commit ad216c0316ad6391d90f4de0a7f59396b2925a06 category: feature bugzilla: NA CVE: NA --------------------------- A potential use after free can occur in _vm_unmap_aliases where an already freed vmap_area could be accessed, Consider the following scenario: Process 1 Process 2 __vm_unmap_aliases __vm_unmap_aliases purge_fragmented_blocks_allcpus rcu_read_lock() rcu_read_lock() list_del_rcu(&vb->free_list) list_for_each_entry_rcu(vb .. ) __purge_vmap_area_lazy kmem_cache_free(va) va_start = vb->va->va_start Here Process 1 is in purge path and it does list_del_rcu on vmap_block and later frees the vmap_area, since Process 2 was holding the rcu lock at this time vmap_block will still be present in and Process 2 accesse it and thereby it tries to access vmap_area of that vmap_block which was already freed by Process 1 and this results in use after free. Fix this by adding a check for vb->dirty before accessing vmap_area structure since vb->dirty will be set to VMAP_BBMAP_BITS in purge path checking for this will prevent the use after free. Link: https://lkml.kernel.org/r/1616062105-23263-1-git-send-email-vjitta@codeauro… Signed-off-by: Vijayanand Jitta <vjitta(a)codeaurora.org> Reviewed-by: Uladzislau Rezki (Sony) <urezki(a)gmail.com> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> Signed-off-by: Rui Xiang <rui.xiang(a)huawei.com> Reviewed-by: Ding Tianhong <dingtianhong(a)huawei.com> Reviewed-by: Weilong Chen <chenweilong(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- mm/vmalloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 8c70131e0b078..011a84ebec04d 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -1953,7 +1953,7 @@ void vm_unmap_aliases(void) rcu_read_lock(); list_for_each_entry_rcu(vb, &vbq->free, free_list) { spin_lock(&vb->lock); - if (vb->dirty) { + if (vb->dirty && vb->dirty != VMAP_BBMAP_BITS) { unsigned long va_start = vb->va->va_start; unsigned long s, e; -- 2.25.1

1 1

[PATCH kernel-4.19 1/4] iommu/vt-d: Add support for ACPI device use physical, node as pci device to establish identity mapping
by Yang Yingliang 28 Jun '21

28 Jun '21

From: LeoLiu-oc <LeoLiu-oc(a)zhaoxin.com> zhaoxin inclusion category: feature bugzilla: https://bugzilla.openeuler.org/show_bug.cgi?id=19 CVE: NA ---------------------------------------------------------------- When using ACPI device to create identity mapping, if the physical node of the ACPI device is a pci device, domain_1 will be created for these two devices. But if the pci device and other pci devices belong to another domain_2, this will happen conflict and destroy another domain. Such as PCI devices under the PCIE-to-PCI bridge. Therefore, when the physical node of the ACPI device is a PCI device, this patch uses the PCI device to create the domain. In this way, there is only one domain. Signed-off-by: LeoLiu-oc <LeoLiu-oc(a)zhaoxin.com> Reviewed-by: Xiongfeng Wang <wangxiongfeng2(a)huawei.com> Signed-off-by: Cheng Jian <cj.chengjian(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/iommu/intel-iommu.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c index 18e0be8e05a53..3e5e1791abbb3 100644 --- a/drivers/iommu/intel-iommu.c +++ b/drivers/iommu/intel-iommu.c @@ -2756,6 +2756,27 @@ static int domain_prepare_identity_map(struct device *dev, return iommu_domain_identity_map(domain, start, end); } +static struct device *acpi_dev_find_pci_dev(struct device *dev) +{ + struct acpi_device_physical_node *pn; + struct acpi_device *adev; + + if (dev->bus == &acpi_bus_type) { + adev = to_acpi_device(dev); + + mutex_lock(&adev->physical_node_lock); + list_for_each_entry(pn, &adev->physical_node_list, node) { + if (dev_is_pci(pn->dev)) { + mutex_unlock(&adev->physical_node_lock); + return pn->dev; + } + } + mutex_unlock(&adev->physical_node_lock); + } + + return dev; +} + static int iommu_prepare_identity_map(struct device *dev, unsigned long long start, unsigned long long end) @@ -2763,6 +2784,8 @@ static int iommu_prepare_identity_map(struct device *dev, struct dmar_domain *domain; int ret; + dev = acpi_dev_find_pci_dev(dev); + domain = get_domain_for_dev(dev, DEFAULT_DOMAIN_ADDRESS_WIDTH); if (!domain) return -ENOMEM; -- 2.25.1

1 3

Re: SMMU Cmdq Related Patch Unable to Reproduce Performance Improvement
by Leizhen (ThunderTown) 28 Jun '21

28 Jun '21

On 2021/6/26 22:33, 李泽斌 wrote: > > Hi all > > I am recently curious about a patch mentioned in OpenEuler Kernel Meeting about SMMU. The patch name is "iommu/arm-smmu-v3: Reduce contention during command-queue insertion", and it has been merged, the commit link is here(https://github.com/torvalds/linux/commit/587e6c10a7ce89a5924fdbeff2ec5…. > > But I find it difficult to reproduce the huge performance improvement, and I notice that there is an evaluation result figure in the meeting. So I turn to you guys for help. I wonder if you can provide more evaluation details to reproduce the improvement. TIA First，it only works in strict mode on ARM64, so please make sure: "iommu.passthrough=1" or "iommu.strict=0" is not set. Second, only some high-end peripherals have this performance bottleneck. For example: 1) 100Gb/s or higher NIC interconnection (direct connection), use command 'iperf' to test it 2) Attach more than 12 SSDs or NVMEs to the PCI expander, use command 'fio' to test it For most low-speed devices, their performance in strict mode is close to iommu.passthrough=1. There is no room for further optimization. > > > Best Regards > > Zebin Li > > . >

1 0

[PATCH kernel-4.19] block: dio: ensure the memory order between bi_private and bi_css
by Yang Yingliang 26 Jun '21

26 Jun '21

From: Hou Tao <houtao1(a)huawei.com> hulk inclusion category: bugfix bugzilla: 167067 CVE: NA -------------------------------- In __blkdev_direct_IO_simple(), when bi_private is NULL, it assumes bi_css must be NULL as showed below: CPU 1: CPU 2: __blkdev_direct_IO_simple submit_bio bio_endio bio_uninit(bio) css_put(bi_css) bi_css = NULL set_current_state(TASK_UNINTERRUPTIBLE) bio->bi_end_io blkdev_bio_end_io_simple bio->bi_private = NULL // bi_private is NULL READ_ONCE(bio->bi_private) wake_up_process smp_mb__after_spinlock bio_unint(bio) // read bi_css as no-NULL css_put(bi_css) Because there is no memory barrier between the reading and the writing of these two variables, the assumption is wrong under weak-memory model machine (e.g. arm64). bi_css will be put twice and leads to the following warning: percpu_ref_switch_to_atomic_rcu: percpu ref (css_release) <= 0 (-3) after switching to atomic There is a similar problem in __blkdev_direct_IO() which occurs between dio->waiter and bio.bi_status. Fixing it by adding a smp_rmb() between the reads of two variables, and a corresponding smp_wmb() between the writes. Signed-off-by: Hou Tao <houtao1(a)huawei.com> Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> Reviewed-by: Jason Yan <yanaijie(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- fs/block_dev.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index 786d105692e85..30dd7b19bd2e3 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -195,6 +195,11 @@ static void blkdev_bio_end_io_simple(struct bio *bio) { struct task_struct *waiter = bio->bi_private; + /* + * Paired with smp_rmb() after reading bio->bi_private + * in __blkdev_direct_IO_simple() + */ + smp_wmb(); WRITE_ONCE(bio->bi_private, NULL); wake_up_process(waiter); } @@ -251,8 +256,14 @@ __blkdev_direct_IO_simple(struct kiocb *iocb, struct iov_iter *iter, qc = submit_bio(&bio); for (;;) { set_current_state(TASK_UNINTERRUPTIBLE); - if (!READ_ONCE(bio.bi_private)) + if (!READ_ONCE(bio.bi_private)) { + /* + * Paired with smp_wmb() in + * blkdev_bio_end_io_simple() + */ + smp_rmb(); break; + } if (!(iocb->ki_flags & IOCB_HIPRI) || !blk_poll(bdev_get_queue(bdev), qc)) io_schedule(); @@ -317,6 +328,12 @@ static void blkdev_bio_end_io(struct bio *bio) } else { struct task_struct *waiter = dio->waiter; + if (!dio->multi_bio) + /* + * Paired with smp_rmb() after reading + * dio->waiter in __blkdev_direct_IO() + */ + smp_wmb(); WRITE_ONCE(dio->waiter, NULL); wake_up_process(waiter); } @@ -417,8 +434,11 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages) for (;;) { set_current_state(TASK_UNINTERRUPTIBLE); - if (!READ_ONCE(dio->waiter)) + if (!READ_ONCE(dio->waiter)) { + /* Paired with smp_wmb() in blkdev_bio_end_io() */ + smp_rmb(); break; + } if (!(iocb->ki_flags & IOCB_HIPRI) || !blk_poll(bdev_get_queue(bdev), qc)) -- 2.25.1

1 0

[PATCH openEuler-1.0-LTS] btrfs: allow btrfs_truncate_block() to fallback to nocow for data space reservation
by Cheng Jian 25 Jun '21

25 Jun '21

From: Qu Wenruo <wqu(a)suse.com> mainline inclusion from mainline-5.13.0-rc5 commit 6d4572a9d71d5fc2affee0258d8582d39859188c category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I39MZM CVE: NA ------------------------------------------------------ From: Gou Hao <gouhao(a)uniontech.com> btrfs: make btrfs_truncate_block check NOCOW attribute For the modification of the main line, there are two places not merge. The mainline code adds the parameter nowait in btrfs_check_can_nocow, during the round, it only changes the name of check_can_nocow. The code of the main line calls the btrfs_drew_write_unlock function at the end of btrfs_truncate_block, and 4.19 kernel of the code does not have this function. In the Reference address, the two codes were not merge too. ------------------------------------------------------ btrfs: allow btrfs_truncate_block() to fallback to nocow for data space reservation [BUG] When the data space is exhausted, even if the inode has NOCOW attribute, we will still refuse to truncate unaligned range due to ENOSPC. The following script can reproduce it pretty easily: #!/bin/bash dev=/dev/test/test mnt=/mnt/btrfs umount $dev &> /dev/null umount $mnt &> /dev/null mkfs.btrfs -f $dev -b 1G mount -o nospace_cache $dev $mnt touch $mnt/foobar chattr +C $mnt/foobar xfs_io -f -c "pwrite -b 4k 0 4k" $mnt/foobar > /dev/null xfs_io -f -c "pwrite -b 4k 0 1G" $mnt/padding &> /dev/null sync xfs_io -c "fpunch 0 2k" $mnt/foobar umount $mnt Currently this will fail at the fpunch part. [CAUSE] Because btrfs_truncate_block() always reserves space without checking the NOCOW attribute. Since the writeback path follows NOCOW bit, we only need to bother the space reservation code in btrfs_truncate_block(). [FIX] Make btrfs_truncate_block() follow btrfs_buffered_write() to try to reserve data space first, and fall back to NOCOW check only when we don't have enough space. Such always-try-reserve is an optimization introduced in btrfs_buffered_write(), to avoid expensive btrfs_check_can_nocow() call. This patch will export check_can_nocow() as btrfs_check_can_nocow(), and use it in btrfs_truncate_block() to fix the problem. Reference: https://patchwork.kernel.org/project/linux-btrfs/patch/20200130052822.11765… Reported-by: Martin Doucha <martin.doucha(a)suse.com> Reviewed-by: Filipe Manana <fdmanana(a)suse.com> Reviewed-by: Anand Jain <anand.jain(a)oracle.com> Signed-off-by: Qu Wenruo <wqu(a)suse.com> Reviewed-by: David Sterba <dsterba(a)suse.com> Signed-off-by: David Sterba <dsterba(a)suse.com> Signed-off-by: Gou Hao <gouhao(a)uniontech.com> Signed-off-by: Cheng Jian <cj.chengjian(a)huawei.com> --- fs/btrfs/ctree.h | 3 ++- fs/btrfs/file.c | 8 ++++---- fs/btrfs/inode.c | 39 +++++++++++++++++++++++++++++++++------ 3 files changed, 39 insertions(+), 11 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 49ef2e48a8c6..3d8c699e44ea 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3273,7 +3273,8 @@ int btrfs_dirty_pages(struct inode *inode, struct page **pages, int btrfs_fdatawrite_range(struct inode *inode, loff_t start, loff_t end); int btrfs_clone_file_range(struct file *file_in, loff_t pos_in, struct file *file_out, loff_t pos_out, u64 len); - +int btrfs_check_can_nocow(struct btrfs_inode *inode, loff_t pos, + size_t *write_bytes); /* tree-defrag.c */ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, struct btrfs_root *root); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index d9d90f0b66d2..f0600b1c6d90 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1536,8 +1536,8 @@ lock_and_cleanup_extent_if_need(struct btrfs_inode *inode, struct page **pages, return ret; } -static noinline int check_can_nocow(struct btrfs_inode *inode, loff_t pos, - size_t *write_bytes) +int btrfs_check_can_nocow(struct btrfs_inode *inode, loff_t pos, + size_t *write_bytes) { struct btrfs_fs_info *fs_info = inode->root->fs_info; struct btrfs_root *root = inode->root; @@ -1647,7 +1647,7 @@ static noinline ssize_t btrfs_buffered_write(struct kiocb *iocb, if (ret < 0) { if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW | BTRFS_INODE_PREALLOC)) && - check_can_nocow(BTRFS_I(inode), pos, + btrfs_check_can_nocow(BTRFS_I(inode), pos, &write_bytes) > 0) { /* * For nodata cow case, no need to reserve @@ -1923,7 +1923,7 @@ static ssize_t btrfs_file_write_iter(struct kiocb *iocb, */ if (!(BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW | BTRFS_INODE_PREALLOC)) || - check_can_nocow(BTRFS_I(inode), pos, &count) <= 0) { + btrfs_check_can_nocow(BTRFS_I(inode), pos, &count) <= 0) { inode_unlock(inode); return -EAGAIN; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 69e376d27bcc..52da573741ed 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4901,11 +4901,13 @@ int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len, struct extent_state *cached_state = NULL; struct extent_changeset *data_reserved = NULL; char *kaddr; + bool only_release_metadata = false; u32 blocksize = fs_info->sectorsize; pgoff_t index = from >> PAGE_SHIFT; unsigned offset = from & (blocksize - 1); struct page *page; gfp_t mask = btrfs_alloc_write_mask(mapping); + size_t write_bytes = blocksize; int ret = 0; u64 block_start; u64 block_end; @@ -4917,10 +4919,26 @@ int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len, block_start = round_down(from, blocksize); block_end = block_start + blocksize - 1; - ret = btrfs_delalloc_reserve_space(inode, &data_reserved, - block_start, blocksize); - if (ret) + ret = btrfs_check_data_free_space(inode, &data_reserved, + block_start, blocksize); + if (ret < 0) { + if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW | + BTRFS_INODE_PREALLOC)) && + btrfs_check_can_nocow(BTRFS_I(inode), block_start, + &write_bytes) > 0) { + /* For nocow case, no need to reserve data space */ + only_release_metadata = true; + } else { + goto out; + } + } + ret = btrfs_delalloc_reserve_metadata(BTRFS_I(inode), blocksize); + if (ret < 0) { + if (!only_release_metadata) + btrfs_free_reserved_data_space(inode, data_reserved, + block_start, blocksize); goto out; + } again: page = find_or_create_page(mapping, index, mask); @@ -4991,10 +5009,19 @@ int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len, set_page_dirty(page); unlock_extent_cached(io_tree, block_start, block_end, &cached_state); + if (only_release_metadata) + set_extent_bit(&BTRFS_I(inode)->io_tree, block_start, + block_end, EXTENT_NORESERVE, NULL, NULL, + GFP_NOFS); out_unlock: - if (ret) - btrfs_delalloc_release_space(inode, data_reserved, block_start, - blocksize, true); + if (ret) { + if (only_release_metadata) + btrfs_delalloc_release_metadata(BTRFS_I(inode), + blocksize, true); + else + btrfs_delalloc_release_space(inode, data_reserved, + block_start, blocksize, true); + } btrfs_delalloc_release_extents(BTRFS_I(inode), blocksize); unlock_page(page); put_page(page); -- 2.25.1

1 0

2024

2023

2022

2021

2020

2019

Kernel June 2021