1<!-- 2Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al. 3 4SPDX-License-Identifier: curl 5--> 6 7# bufq 8 9This is an internal module for managing I/O buffers. A `bufq` can be written 10to and read from. It manages read and write positions and has a maximum size. 11 12## read/write 13 14Its basic read/write functions have a similar signature and return code handling 15as many internal Curl read and write ones. 16 17 18``` 19ssize_t Curl_bufq_write(struct bufq *q, const unsigned char *buf, size_t len, CURLcode *err); 20 21- returns the length written into `q` or -1 on error. 22- writing to a full `q` returns -1 and set *err to CURLE_AGAIN 23 24ssize_t Curl_bufq_read(struct bufq *q, unsigned char *buf, size_t len, CURLcode *err); 25 26- returns the length read from `q` or -1 on error. 27- reading from an empty `q` returns -1 and set *err to CURLE_AGAIN 28 29``` 30 31To pass data into a `bufq` without an extra copy, read callbacks can be used. 32 33``` 34typedef ssize_t Curl_bufq_reader(void *reader_ctx, unsigned char *buf, size_t len, 35 CURLcode *err); 36 37ssize_t Curl_bufq_slurp(struct bufq *q, Curl_bufq_reader *reader, void *reader_ctx, 38 CURLcode *err); 39``` 40 41`Curl_bufq_slurp()` invokes the given `reader` callback, passing it its own 42internal buffer memory to write to. It may invoke the `reader` several times, 43as long as it has space and while the `reader` always returns the length that 44was requested. There are variations of `slurp` that call the `reader` at most 45once or only read in a maximum amount of bytes. 46 47The analog mechanism for write out buffer data is: 48 49``` 50typedef ssize_t Curl_bufq_writer(void *writer_ctx, const unsigned char *buf, size_t len, 51 CURLcode *err); 52 53ssize_t Curl_bufq_pass(struct bufq *q, Curl_bufq_writer *writer, void *writer_ctx, 54 CURLcode *err); 55``` 56 57`Curl_bufq_pass()` invokes the `writer`, passing its internal memory and 58remove the amount that `writer` reports. 59 60## peek and skip 61 62It is possible to get access to the memory of data stored in a `bufq` with: 63 64``` 65bool Curl_bufq_peek(const struct bufq *q, const unsigned char **pbuf, size_t *plen); 66``` 67 68On returning TRUE, `pbuf` points to internal memory with `plen` bytes that one 69may read. This is only valid until another operation on `bufq` is performed. 70 71Instead of reading `bufq` data, one may simply skip it: 72 73``` 74void Curl_bufq_skip(struct bufq *q, size_t amount); 75``` 76 77This removes `amount` number of bytes from the `bufq`. 78 79 80## lifetime 81 82`bufq` is initialized and freed similar to the `dynbuf` module. Code using 83`bufq` holds a `struct bufq` somewhere. Before it uses it, it invokes: 84 85``` 86void Curl_bufq_init(struct bufq *q, size_t chunk_size, size_t max_chunks); 87``` 88 89The `bufq` is told how many "chunks" of data it shall hold at maximum and how 90large those "chunks" should be. There are some variants of this, allowing for 91more options. How "chunks" are handled in a `bufq` is presented in the section 92about memory management. 93 94The user of the `bufq` has the responsibility to call: 95 96``` 97void Curl_bufq_free(struct bufq *q); 98``` 99to free all resources held by `q`. It is possible to reset a `bufq` to empty via: 100 101``` 102void Curl_bufq_reset(struct bufq *q); 103``` 104 105## memory management 106 107Internally, a `bufq` uses allocation of fixed size, e.g. the "chunk_size", up 108to a maximum number, e.g. "max_chunks". These chunks are allocated on demand, 109therefore writing to a `bufq` may return `CURLE_OUT_OF_MEMORY`. Once the max 110number of chunks are used, the `bufq` reports that it is "full". 111 112Each chunks has a `read` and `write` index. A `bufq` keeps its chunks in a 113list. Reading happens always at the head chunk, writing always goes to the 114tail chunk. When the head chunk becomes empty, it is removed. When the tail 115chunk becomes full, another chunk is added to the end of the list, becoming 116the new tail. 117 118Chunks that are no longer used are returned to a `spare` list by default. If 119the `bufq` is created with option `BUFQ_OPT_NO_SPARES` those chunks are freed 120right away. 121 122If a `bufq` is created with a `bufc_pool`, the no longer used chunks are 123returned to the pool. Also `bufq` asks the pool for a chunk when it needs one. 124More in section "pools". 125 126## empty, full and overflow 127 128One can ask about the state of a `bufq` with methods such as 129`Curl_bufq_is_empty(q)`, `Curl_bufq_is_full(q)`, etc. The amount of data held 130by a `bufq` is the sum of the data in all its chunks. This is what is reported 131by `Curl_bufq_len(q)`. 132 133Note that a `bufq` length and it being "full" are only loosely related. A 134simple example: 135 136* create a `bufq` with chunk_size=1000 and max_chunks=4. 137* write 4000 bytes to it, it reports "full" 138* read 1 bytes from it, it still reports "full" 139* read 999 more bytes from it, and it is no longer "full" 140 141The reason for this is that full really means: *bufq uses max_chunks and the 142last one cannot be written to*. 143 144When you read 1 byte from the head chunk in the example above, the head still 145hold 999 unread bytes. Only when those are also read, can the head chunk be 146removed and a new tail be added. 147 148There is another variation to this. If you initialized a `bufq` with option 149`BUFQ_OPT_SOFT_LIMIT`, it allows writes **beyond** the `max_chunks`. It 150reports **full**, but one can **still** write. This option is necessary, if 151partial writes need to be avoided. It means that you need other checks to keep 152the `bufq` from growing ever larger and larger. 153 154 155## pools 156 157A `struct bufc_pool` may be used to create chunks for a `bufq` and keep spare 158ones around. It is initialized and used via: 159 160``` 161void Curl_bufcp_init(struct bufc_pool *pool, size_t chunk_size, size_t spare_max); 162 163void Curl_bufq_initp(struct bufq *q, struct bufc_pool *pool, size_t max_chunks, int opts); 164``` 165 166The pool gets the size and the mount of spares to keep. The `bufq` gets the 167pool and the `max_chunks`. It no longer needs to know the chunk sizes, as 168those are managed by the pool. 169 170A pool can be shared between many `bufq`s, as long as all of them operate in 171the same thread. In curl that would be true for all transfers using the same 172multi handle. The advantages of a pool are: 173 174* when all `bufq`s are empty, only memory for `max_spare` chunks in the pool 175 is used. Empty `bufq`s holds no memory. 176* the latest spare chunk is the first to be handed out again, no matter which 177 `bufq` needs it. This keeps the footprint of "recently used" memory smaller. 178