View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000381 | 1003.1(2008)/Issue 7 | Rationale | public | 2011-02-17 00:18 | 2013-04-16 13:06 |
| Reporter | RossPJohnson | Assigned To | ajosey | ||
| Priority | normal | Severity | Editorial | Type | Clarification Requested |
| Status | Closed | Resolution | Accepted As Marked | ||
| Name | Ross Johnson | ||||
| Organization | |||||
| User Reference | |||||
| Section | System Interfaces - Threads - Thread-Specific Data | ||||
| Page Number | 3567 (http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09) | ||||
| Line Number | 120744 | ||||
| Interp Status | --- | ||||
| Final Accepted Text | See 0000381:0000694 | ||||
| Summary | 0000381: Claimed hashability of pthread_t is misleading | ||||
| Description | The text under model option 1: <quote> No Explicit Support A standard thread-specific data interface is not strictly necessary to support applications that require per-thread context. One could, for example, provide a hash function that converted a pthread_t into an integer value that could then be used to index into a global array of per-thread data pointers. This hash function, in conjunction with pthread_self(), would be all the interface required to support a mechanism of this sort. Unfortunately, this technique is cumbersome. It can lead to duplicated code as each set of cooperating modules implements their own per-thread data management schemes. </quote> The description claims that pthread_t is hashable. According to authoritative comments elsewhere, opague pthread_t instances may contain values that are variable between copies and instances. This may be as a result of alignment padding or by design where additional non-thread-identifying values are included to support the implementation. An application has no knowledge of pthread_t internals and therefore would need to hash over its entire content. As a result, for example, a hash of pthread_t objects returned by successive calls to pthread_self() may return a different result each time. | ||||
| Desired Action | If the description is historically based on a previous hashable definition of pthread_t then provide an indication that this model is no longer valid, or remove the entire model option 1 from future editions. | ||||
| Tags | tc1-2008 | ||||
|
|
You miss the point of the rationale. It documents what was considered during the writing of the standard. Not everything written in the rationale has any meaning wrt the current standard. In this specific case option 3 was selected for the standard and the other options are just there as an explanation. |
|
|
The rationale documents accurately the models that were considered. Add a new sentence to the end of line 120771: "Therefore, the third model was chosen." Also add to 120753: "This technique would also require that pthread_t not be an opaque type." |
|
|
This bug deals only with text in the rationale, which will be adjusted in TC1. Meanwhile, it was noted in the 10 Mar 2011 teleconference that there was a lot of list traffic debating whether it would also be worth creating an interface to allow the hashing of pthread_t values; such an interface proposal should occur as a separate bug report targeting the next revision of the standard, and should be accompanied with proposed wording for the requirements on the new interface(s). |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2011-02-17 00:18 | RossPJohnson | New Issue | |
| 2011-02-17 00:18 | RossPJohnson | Status | New => Under Review |
| 2011-02-17 00:18 | RossPJohnson | Assigned To | => ajosey |
| 2011-02-17 00:18 | RossPJohnson | Name | => Ross Johnson |
| 2011-02-17 00:18 | RossPJohnson | Section | => System Interfaces - Threads - Thread-Specific Data |
| 2011-02-17 00:18 | RossPJohnson | Page Number | => http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09 |
| 2011-02-17 00:18 | RossPJohnson | Line Number | => http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09 |
| 2011-02-17 10:46 | drepper | Note Added: 0000674 | |
| 2011-03-10 17:16 | nick | Page Number | http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09 => 3567 (http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09) |
| 2011-03-10 17:16 | nick | Line Number | http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xsh_chap02.html#tag_22_02_09 => 120744 |
| 2011-03-10 17:16 | nick | Interp Status | => --- |
| 2011-03-10 17:23 | nick | Note Added: 0000694 | |
| 2011-03-10 17:23 | nick | Status | Under Review => Resolved |
| 2011-03-10 17:23 | nick | Resolution | Open => Accepted As Marked |
| 2011-03-10 17:23 | nick | Final Accepted Text | => See 0000381:0000694 |
| 2011-03-10 17:24 | nick | Note Edited: 0000694 | |
| 2011-03-10 17:26 | nick | Tag Attached: tc1-2008 | |
| 2011-03-10 17:38 | eblake | Note Added: 0000695 | |
| 2013-04-16 13:06 | ajosey | Status | Resolved => Closed |
