similar to: Panic: file buffer.c: assertion failed: (used_size <= buf->alloc)

This is log from dovecot 2.2.10: 2014-01-11T09:55:14+01:00 server/a.b.c.d dovecot: imap(user at domain): Error: Corrupted transaction log file /pracmail/domain/user/mail/.sent-mail/dovecot.index.log seq 16777216: log file shrank (868 < 6144) (sync_offset=6144) 2014-01-11T09:55:14+01:00 server/a.b.c.d dovecot: imap(user at domain): Panic: file buffer.c: line 307 (buffer_set_used_size):

I haven't seen this before 1.1b16, it happened to two users today while they were searching using fts. Feb 18 16:41:36 hill dovecot: IMAP(username): file buffer.c: line 288 (buffer_set_used_size): assertion failed: (used_size <= buf->alloc) Feb 18 16:41:36 hill dovecot: child 53560 (imap) killed with signal 6 I can probably narrow it down to an example mail if needed. GNU gdb 6.1.1

Hello, I am trying to use Xen Enterprise (its free version). Until now I have no problem, but now I want to create a new virtual machine based in debian etch, not debian sarge (the debian version included in xen templates). So I am looking for a way to create my own template, but I haven''t found any documentation about that. Looking the files, I have found that templates are stored in

This patch series allows user-specified mount commands to be sent in via kernel command line ("kinit_mount=...") or via an embedded /etc/fstab file. The first patch is a cleanup of a patch sent last November by San Mehat (http://web.archiveorange.com/archive/v/EazJNBMORV2U7E0coh5h); the next two are small improvements or bug fixes.

It was noticed that the copy_user() friends that was used to access virtqueue metdata tends to be very expensive for dataplane implementation like vhost since it involves lots of software check, speculation barrier, hardware feature toggling (e.g SMAP). The extra cost will be more obvious when transferring small packets. This patch tries to eliminate those overhead by pin vq metadata pages and

Whitespace neatening... Multiline statement argument alignment. Argument wrapping. Use kmalloc_array instead of kmalloc. --- drivers/misc/mic/card/mic_virtio.c | 17 ++++--- drivers/misc/mic/card/mic_x100.c | 4 +- drivers/misc/mic/host/mic_debugfs.c | 91 ++++++++++++++++++------------------- drivers/misc/mic/host/mic_fops.c | 6 +-- drivers/misc/mic/host/mic_intr.c | 37

Dear All, I think I've hit a bug after migrating our mail server to the newest version of dovecot (version 2.2.15). When I try to list all imap folders for one user the imap process crashes. The listing to a specific folder (ex. INBOX) works fine, only the "*" crashes. Client side: --------------------- [root at mb log]# telnet 127.0.0.1 143 Trying 127.0.0.1... Connected to

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. Test shows about 24% improvement on TX PPS. It should benefit other cases as well. Please review Jason Wang (3): vhost: generalize adding used elem vhost: fine grain userspace memory

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. Test shows about 24% improvement on TX PPS. It should benefit other cases as well. Please review Jason Wang (3): vhost: generalize adding used elem vhost: fine grain userspace memory

From: Ashutosh Dixit <ashutosh.dixit at intel.com> This patch introduces the card "Virtio over PCIe" interface for Intel MIC. It allows virtio drivers on the card to communicate with their user space backends on the host via a device page. Ring 3 apps on the host can add, remove and configure virtio devices. A thin MIC specific virtio_config_ops is implemented which is borrowed

From: Ashutosh Dixit <ashutosh.dixit at intel.com> This patch introduces the card "Virtio over PCIe" interface for Intel MIC. It allows virtio drivers on the card to communicate with their user space backends on the host via a device page. Ring 3 apps on the host can add, remove and configure virtio devices. A thin MIC specific virtio_config_ops is implemented which is borrowed

On Thu, Dec 13, 2018 at 06:10:22PM +0800, Jason Wang wrote: > It was noticed that the copy_user() friends that was used to access > virtqueue metdata tends to be very expensive for dataplane > implementation like vhost since it involves lots of software check, > speculation barrier, hardware feature toggling (e.g SMAP). The > extra cost will be more obvious when transferring small

On Thu, Dec 13, 2018 at 06:10:22PM +0800, Jason Wang wrote: > It was noticed that the copy_user() friends that was used to access > virtqueue metdata tends to be very expensive for dataplane > implementation like vhost since it involves lots of software check, > speculation barrier, hardware feature toggling (e.g SMAP). The > extra cost will be more obvious when transferring small

Endianness issues are now consistent as per the documentation in host/mic_virtio.h. Note that the host can be both BE or LE whereas the card is always LE. Memory space sparse warnings are fixed for now by using __force. This is sufficient for now since the driver depends on x86 but will need to be revisited if we support other architectures which treat I/O memory differently from system memory.

Endianness issues are now consistent as per the documentation in host/mic_virtio.h. Note that the host can be both BE or LE whereas the card is always LE. Memory space sparse warnings are fixed for now by using __force. This is sufficient for now since the driver depends on x86 but will need to be revisited if we support other architectures which treat I/O memory differently from system memory.

http://dovecot.org/test/ - Fixed dying with SIGIO signal with busy Maildir + Linux + IDLE system - Fixed maildir syncing: "UIDs not ordered in file" errors could have occured sometimes wrongly - --with-ioloop=epoll enables epoll code with Linux 2.6. Untested after my changes. - Pretty big changes in internal index handling code (I started this change 2-3 months ago..) Keywords are

It was noticed that the copy_user() friends that was used to access virtqueue metdata tends to be very expensive for dataplane implementation like vhost since it involves lots of software checks, speculation barrier, hardware feature toggling (e.g SMAP). The extra cost will be more obvious when transferring small packets since the time spent on metadata accessing become significant.. This patch

We are working on migrating Dovecot 1.2.17 running on AIX 5.3 (believe it or not!) to Dovecot 2.0.13 running on Ubuntu. We have hundreds of users mboxes we will be migrating. My question is regarding the index files. Should we remove those after the migration, but before we open it up to users so Dovecot can create new ones? I did a test migration of a single user, and Dovecot detects the

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. Test shows about 24% improvement on TX PPS. It should benefit other cases as well. Changes from V1: - instead of pinning pages, use MMU notifier to invalidate vmaps and remap duing

These patches fix various issues which were reported or found with the MIC driver. Changelog ========= v2 => v3: * Reorder patch 5 in v2 to patch 4 in v3. * Split patch 4 in v2 into patches 5 and 6 in v3. v1 => v2: @ https://lkml.org/lkml/2013/11/26/376 * Address review comments @ https://lkml.org/lkml/2013/11/26/443 in patch 5. v1: Initial post @ https://lkml.org/lkml/2013/11/26/321